Late-Round Perspectives: Subvertadown

00:00:00
Speaker 1: Subverted Down is a fantasy analyst who's known for a statistical analysis around streaming various positions in fantasy football. He owns and operates subvertedown dot com. He's a frequent Reddit contributor. He's got more models, and I have bad dad jokes. And this is actually the first time he's ever hopped on a podcast. These are his late round perspectives. I just saw that this is that that you're not feeling one hundred percent that uh oh yeah, this is like.

00:00:33
Speaker 2: Well, I don't know that flu or covid or covid flu whatever going around. Yeah, I don't know, but but I trust you're going to bring me through this, all right. You know, I.

00:00:43
Speaker 3: Appreciate that, man, I appreciate that. I do.

00:00:45
Speaker 2: I gotta what's that The only prescription is more JJ Yeah, there you go, There you go.

00:00:52
Speaker 1: I did want to, uh to kick off today's show by sort of talking like, I feel like your your presence online is relatively elusive. It's it's you're you're sort of an enigma at this point where you know you've done so much incredible work. You have an incredible website, you know you've you've built this following via Reddit, shout out Reddit. But let's first talk about like, is this is this your first appearance basically like with Fantasy football.

00:01:21
Speaker 2: Yeah, yeah, I you brought me out of my basement, I'd say, except I actually deliberately chose the basement because I thought redditors would be disappointed if I didn't go with the proverbial rhetor in the basement cliche. So that's right. But but yeah, no, I thought it was a good chance. You know, it's uh and you know, you sort of sold me on it saying that basically, hey, we're basically going to nerd out, and you know for me, that's that's the safe space, so you know, pull me out of my shell.

00:01:49
Speaker 3: Yeah, there you go. Yeah, it's it's it's awesome.

00:01:51
Speaker 1: And I also want so so you go by Subverted down. How did that all come about? Like even just the naming convention and you coming up with that name to be as your alias essentially?

00:02:02
Speaker 2: Yeah, I wish it was more exciting in a way, but you know, because when you're choosing your first log in name, you know, you've got that moment like okay, I got to get going. So it was a little bit fast, but I you know, sub was you know, a subreddit was a new term to me, and I thought, oh love, what with that? And I thought subverb, Okay, I'm trying to sort of disrupt what's going on, and you know, think different and all that. And then the you know down, I guess, in a way, had two things. One was obviously a down in football, like so the idea of like you would be trying to you know, trick your way into getting that one extra you know, not not an extra touchdown, but you know, just a down, you know, to get that marginal advantage.

00:02:40
Speaker 3: Yeah.

00:02:40
Speaker 2: And I'd had a hobby also ten years ago, following a lot of you know, rock music history as it turned out, So I was followed you know, from way back Duke Ellington and everything, and and I was up to a system of a down at the time I was listening to. So I thought, okay, I don't know, whatever makes sense, has nothing to do with it at all, but basically, and then I kept it with you know, there was this debate about the website like stay like that, and then we decided, you know, it kind of fits still, it doesn't roll off the tongue. So that's the I think the bigger problem.

00:03:09
Speaker 3: I like it. I like it.

00:03:10
Speaker 1: You know, it's funny because I got my start and some people don't like. I mean, I would assume a lot of people don't even realize this, but like my handle on Twitter is Late Round QB, and you know, my Reddit handle is late Round QB.

00:03:23
Speaker 3: And the brand.

00:03:24
Speaker 1: That I now run in the business I now run is Late Round Fantasy Football. And that all started because in twenty twelve, I wrote an ebook called The Late Round Quarterback. Because this is back when like people were drafted quarterbacks in the first round and second round and stuff like that. And again, I like, similar to you, I wanted to disrupt write and come up with something that was different than what was out there, and I wrote an e book. And this is when ebooks were not like super ever. I mean now every like there's e books just everywhere, and you know, it's just a different time. But I just used the Late Round moniker essentially and just a lot carried it throughout my career and so I can appreciate, you know, a little, a little alias. I mean, like Matthew Berry has talented mister like TMR, Like people might not even realize that that's from something else that's just not some random letters that are out there.

00:04:11
Speaker 2: I think it's a pretty much you know, feel free to go with subvert that's fine, or you know, sub whatever. Do you want to go old school so b D or something. People don't do that, but whatever, whatever it goes. People use a lot of different things on Reddit.

00:04:25
Speaker 1: Well let's let's talk about Reddit, and let's talk about your sort of journey and doing what you're doing. You know, again, this is a different platform for you, you know, versus what you've typically been doing with Reddit. So there's gonna be people that naturally don't know your backstory, and you know, not to downplay what other people have done who have come on perspectives, but I think yours is the most unique or one of the more unique of someone who has been on this show, just because you know, a lot of people are to be fair, like Andrew Cooper, for instance, who's very active on Reddit. He did start you know, his his career. Coop did VA read it? But I just think your story is really really interesting, so talk through it and how you got into this thing in the first place.

00:05:06
Speaker 2: No, it wasn't. It was definitely not straightforward as I as I see as some others go, I mean, I come from other industries. My background is really diverse across different stuff. So I was just starting to basically and involve myself with the forum and read what's going on, and uh, and I guess I really started to think about, you know, trying my hand at this, because I think everyone can associate with this. Things just weren't going good. I was, I was streaming, I was following another noticerespect there was another you know, great guy before me who was doing some DST stuff, was well known, but for whatever reason, it just wasn't working for me. Every week seemed and I know that's you know, by the way, there's a fraction, there's a percentage of people out there who think the same exactly about me right now, Like they're like I picked the bust every single week, Like statistically that has to happen, right, So, like there's someone out there, so I but I think my thought was that there were a few things I thought were missing and wanted to bring and create. One of them was being a bit transparent about the methodology, like trying to find out what are the assumptions going into this, you know, everything I do, I would sum up as trying to bring a like rational I think the people who follow my stuff would say they're looking for a rational view of making decisions based on data and stuff. So I think that would you know, mesh a lot with your point of view as well. So and then so I wanted to bring that very heavy data based, you know, fatual effort. And also I wanted to see, like do these people don't care about the accuracy. The accuracy is a big thing for me. I spend a lot of time following my own accuracy, comparing it against others and sharing it and saying, you know, and I try to put an effort saying like people, this is an exceptionally good year or an exceptionally bad year, like to you know, sort of gaye like don't feel bad if things are not going your way, but you know, or things up or down. So I think it was, oh yeah, the other element I really wanted to bring with strategy. So besides, you know, just like having a very good statistically sound model. I can talk more about models, oh well, and looking at accuracy, but the whole idea of using this first strategy, I wanted to plan future weeks as well. I hated when I would like drop a DST or something and come back next week, and that one was recommended at the top, you know, like let's involve this, And I think that was one of the eventually big things that I added, was trying to give that that look ahead to help people's strategy. And I I feel like it's caught on mores and I see a lot more people doing the look ahead thing on Reddit as well, So those were sort of the yeah, the how it initiated.

00:07:50
Speaker 3: Yeah, you know, it's it's funny because obviously.

00:07:53
Speaker 1: Like you've made i mean, your your bread and butter has been streaming right and it's been it's been you know, with Defense with Kicker, et cetera. And it's funny because one of like, I have another podcast Living the Stream, and that podcast started in like twenty thirteen ish.

00:08:09
Speaker 3: It's it's sort of.

00:08:10
Speaker 1: Become its own, like you know, niche kind of show where we're only fifteen minutes of the actual hours talked about fantasy football and the other you know, X number min you know, forty five minutes is me and my my co host Denny Carter talking about just like just that stuff and just you know, random things in life. But but like I too, you know, like there's a few things that you just said that I think is just incredibly important. One of the things is transparency, which I think is one of the most important pieces of being a content creator in this space. I understand this urge for especially like newcomers in fantasy football as and analysts to like not want to be overly transparent and feel like they should only celebrate their victories and not necessarily, you know, talk bad about themselves because they may. But the reality of the situation is that the majority of like you know, coherent and kind and normal people out there who are consuming this content understand that you're going to get stuff wrong and that you're not going to be flawless because there's so much variance in this game, right, Like, there's just it's naturally you're not going to get everything right, and you got to hope that, Like I remember last year, I don't know if you remember this or if you had this defense high last year, but last year in the championship, the Colts were a great streaming against the Giants, right, like a really good streaming play, and on paper, everything came together. If that had happened, because the Colts you know, they craped the bed, But if that had happened in Week six, no one would really care that much because that's just a blip, right, But it happened in the most important week of the fantasy season, and so it was a bigger deal than it otherwise would be.

00:09:45
Speaker 3: And so for me at least, And I think you're kind of alluding to this too.

00:09:50
Speaker 1: As an analyst, you're trying to just be as right as you can and you're trying to guide people in the right direction as frequently as possible.

00:09:57
Speaker 3: And then the other thing that you mentioned is.

00:09:59
Speaker 1: That there's going to be someone out there who thinks that you suck because they use your advice and naturally it was the wrong advice, right. And it's the same thing that I feel about rankings too, Like going into the season, someone could use the exact same set of rankings and literally just draft off of the rankings in the tiers, right, and it just so happens that the way their draft goes down and where they're positioned in their draft, they end up with Brian Thomas and they end up with these guys who end up busting even though the rankings were really great for somebody else, and you're just hoping again as an analyst, that your rankings are helping far more people than they're hurting people right in the end. And I'm assuming that that's kind of like what your general philosophy is as well. And so you started on Reddit, you started doing all this you know, streaming streaming work, and you were writing, I mean, you were doing write ups and such, and then you've moved to other platforms. What has that been like and what have you done there?

00:10:57
Speaker 2: Yeah? Right, well that only came after a bit of evil. I mean, to your point, my very first year was not a good one. I honestly just assumed this was going to be quite easy and that I would be able to deliver a democratized, very simple equation that everyone could just use and hey, guys, look here's the new way. And that didn't happen. So there was an evolution of a lot of work in revising the model. It was after about a year and a half that I think that things really started to click, and it was because I had introduced the correct, proper way of doing actual predictive modeling, which I was hoping it wouldn't have to go that far initially, but it was also a fun journey. So anyway, then it just yeah, the journey. I guess what happened was the reason I eventually switched. The next thing would have been Patreon actually that I went to after Reddit, And that was actually because after a few years things got really intense, like there were a lot of demands on the model. It had not only evolved, I think people really enjoyed the journey. I mean, that's what I like to share. It's like, hey, I made these are the improvements I made. You know, when I do the whole thing about accuracy improvements is I focus on here's what I think I got wrong, and next time I'm going to improve that. So it had been through all that, but there was so much I was so busy because and read it is not the best platform for all the sharing. Like I would make a post on Tuesday and for DSTs, and I would end up like editing it three times a day or something, you know, because the veggus lines would change or somebody would say, hey, now looks like there's snow or something, and oh, what are you doing now about this injury? All this quarterbacks out and so it's like it became like this really exhausting. It consumed a lot of effort to like, it's not I think that's one thing you could also identify with. It's not just like a model, like this simple thing that you put out in the world and it just like runs itself. Yes, this was what I used to think of it, like it's constant maintenance. And I think people will also, So I'm going on a tangent here, So I think people didn't realize on the in the background that like, as the season was going on, I was revising the model, like you know, and it wasn't you know, not because it was bad. I was adding new things to it. I was looking at like I was studying like in the past, like hey, maybe the mid season has different drivers predictive factors, and let's look you know. So I was kind of doing all these like extra studies to sort of increase the accuracy. But anyway, this sharing of it, and they're like constant updating became quite a lot. So I eventually started putting things out on Actually I was was at some point sharing a Google sheets or some like PDF output or something so that people and that and that. Yeah, that that eventually led over to Patreon as well as you alluded to. I've done a lot of studies like and I can just mention the thing I really enjoy doing actually is not the rankings itself. That's that's like a headache. That kind of invites a lot of scrutiny. Yes, yeah, and it's so much effort, you know, it's it's not just for me. I really liked these studies, Like I just did one yesterday that was about should you pair your DST against another player you have if they're opposing teams, so you know, diving into all that like super interesting stuff. But people wanted the rankings. But anyway, I needed a place to sort of catalog all these studies. I felt like Reddit was quite terrible for like being an archive, Like there was no place to like, so I tried to make posts that like summarized where they were. So anyway, Patreon became a thing I originally also, Yeah, it got a bit overburdened and I started, you know, some people think like, oh, he was going there to try to just to earn money or something like that, and that wasn't the story about it. Really. I kind of came to a point where it was so much work, and the you know, the feedback I was getting was not always positive. So I said, okay, either I can you know, stop, I've like come as far as I can. The model's good, it's great, it's things are working, it's got predicted how it's or I should find a way to continue and make it sustainable. Yeah. And I had this idea maybe of a bit of a robin hood you know thing where like my biggest supporters would probably say, yeah, I'm really into that stuff and the studies and the like the week's look ahead and all that, I'll subscribe or something, And that was just I Also it was also I guess mixed with the idea of I like to measure, you know, like as by nature. Yeah, this is this something that's actually people like appreciate. So so eventually it moved a lot of the stuff to Patreon for that reason. Patreon itself was not ideal. And that's going.

00:15:41
Speaker 3: Through that right now too. I feel that yeah, yeah.

00:15:43
Speaker 2: Yeah, yeah, I mean I had, first of all, with the subscribers, there was this thing where like you can pause the payments each month, but like so in the off season, I wanted to make sure people weren't paying and I was, and you know, I was just charging like what was it one dollar a month to the lowest It was like, you know, it's like nothing, right, or be three dollars, it was a seven dollars whatever that I had to like remember each month to pose it. And also just the whole thing about the functionality of it, Like if you contrast it now, anyone who used the website see that, Like you can just sort by any column, you can hover over here, you can see the opponent by hovering your mouse. You can see all these like extra barographs and distributions, ranges and all that like, and that basically could never happen the Patreon. But the way it happened, I was lucky enough that I was reached out to by a subscriber actually or a fan who said, hey, I see some potential in this, Like I think that you could really elevate the experience people have from following this stuff by putting it on a website. And of course you never know who you're going to meet on the internet. You can go any which way. But a fantastic guy. I'm so proud of the website right now, the work he's done on it to make it very you know, clean, simple and delivering all the stuff that's to the point. Yeah, So eventually it was basically closing down Patreon for that was one of the reasons.

00:17:08
Speaker 1: Yeah, I mean, look, the website's amazing, it's great, and you have a lot of tools on that site. Can you talk through some of the I mean, that's obviously something that you can't you know, do on a stately Patreon where you're just basically like uploading assets. That's that's essentially what I'm doing on Patreon, you know, with rankings and stuff like that. But what are some of the tools that you're offering to help fantasy managers out?

00:17:31
Speaker 2: Yeah, I think so. I guess the first one, of course, is the look ahead. I mean that I mentioned before, the strategic idea of viewing weeks ahead, which is one of the trickier parts of the things that I model actually because it means that I need to, okay, in simple terms, I basically need to model future game scores as well. I need to sort of be as good as Bookies and Vegas at predicting what that next game will be. So that's definitely the major aspect. Then there's you know, my the most common question that I always got, and this is what I really wanted to impart to people, was why is that ranking high? Why are the Seahawks low this week or something or whatever it is? And to be able to explain that, Like I could just give away the model and say, oh, yeah, it's two point seven times the past per passing yards per attemptance, like you know, but that's not useful. And by the way, it was changing every year, like you know, the actual with updates. So I think one of the really useful things I wanted to bring to people isn't interpret the model and give the like key reasons what's making this player higher or lower? I also include a big thing for me, like you mentioned the variants, and to me, that is a I've spent a lot of time and by the way, I think I have like twenty articles on my website about there's a category called expectations, like just to tell people like what's normal, what to expect, what you know things are like if you six points less than predicted, is that like reasonable out of the ordinary. So I have you know, risk, I actually model the risk the upside downside, so people can get a feeling if it's a safe bed or not. And I think one of the ones that I hope people are using because it really gets to that point, is this one I created that does similar historical matches. So we make these histograms. You know. It's a separately developed model, actually different from the normal one where I went in and tried to search through history. So it's constant. It's constantly updating each time there's any little change, and it searches through the last fifteen years of games and tries to pick out which one is close. It's based on I mean, I didn't create what is similar, you know, the computer decides what makes it a similar match. Anyway, the point of that is all It actually will display the distribution in the past of how this game has gone, so you can easily see that you know, hey, this game might be leaning, you know, they might be skewed towards lower value, but with some upside where you can see that actually there was like a to remindus eight point range for that particular scenario. And I hope that is adding value for people. I think it really is because a lot of people are looking for that intuition of what what what is normal and what can I really expect from these.

00:20:18
Speaker 1: Yeah, and look you you alluded to this earlier in saying that you know, a lot of people just want rankings, right, they just want to have that answer and they want to use that. But what I've found and discovered, and I think that, like if you were to go and you would see what analysts are talking about on like Twitter or even like from an analyst perspective, a lot of analysts would say that some of the stuff that some of the deep dives that that other analysts do are not necessary because people just want at the end of the day, they just want these rankings. And I have a very strong disagreement with that because I think that and not only that, I think my my audience in general would would just be the perfect exempt in your audience on Reddit and such would be a perfect example of this. But this game that we're playing right now, people are not into this hobby simply because they only want to win, right, This is an outlet for people to learn different things, to really dive in and analyze something that is separate from their day job and separate from their everyday life.

00:21:23
Speaker 3: And this is you.

00:21:24
Speaker 1: This is an ability for them for anyone, right like you and I started the same way, where we're just like, we like to dive deep into these topics and dig into it and analyze it, because, yes, the end goal is that we want to be better in fantasy football.

00:21:36
Speaker 3: We want to.

00:21:36
Speaker 1: Win, But there is such an appetite out there that people have to really dig in and try to understand this game on a deeper level. And that's why, like, I think the hardest part is finding that balance of simply providing rankings without any sort of backbone and just saying, like, here have them.

00:21:55
Speaker 3: You know, like, once you've established yourself as someone.

00:21:58
Speaker 1: Who you know has a problem and can do this, you can say, yeah, here's my rankings, just trust me.

00:22:03
Speaker 3: But that just doesn't do it for me, you know.

00:22:05
Speaker 1: Like I think that we need to always and constantly be providing that backbone because that's not only fair to customers and to people who are consumer or content.

00:22:14
Speaker 3: But I think people care.

00:22:16
Speaker 1: And what you're alluding and what you're saying here is that you're showing these histograms and you're showing these charts and such to allow people to dig in themselves and to still understand, to not just take what you're saying at face value.

00:22:27
Speaker 3: All the time and to dig into it themselves too.

00:22:31
Speaker 2: That's exactly right, and I think I should have mentioned earlier. I mean the whole point for wanting like this rational basis it increases people's enjoyment of the game. They like feel like they're actually understand and know what space they're operating in, right, And that can go two ways. That can be like I want to trust the model so I know that this player higher is the one I should choose, which is a little too simple, or like I want to know that I don't need to worry quite so much. Yes, that's a big point for me. Like I've got questions before, like you know, why do you have I'm just gonna say Boswell, because he got the most attention this year. Why do you have him like number nine or something? And it's like, well, okay, he's number nine, but like there's a zero point two difference between him and like the number four or number three, Like it's one tier, you could say, but you know, you need that projection, You need that ability to know like okay, what what actually I do? I need to sweat it? Can? I think? To me, it's about letting people have fun, Like like you have a lot of freedom. If there is one one and a half points different, then more than that, then yeah, okay, maybe start to think like you could make a better choice more. You've got a lot of flexibility to make it your game, to make you know your choices the ones you enjoy. I try to stress that quite frequently. People say like, oh, I've just got this feeling about make Feerson this week, and you know, but I see he's like in your ak and I'm just like, the difference is like two points. Let's you know, it's the difference between let's say I don't know, nine points and seven in the projection. Like you if you're gonna feel way worse if you listen to this guy here, like just tell you and it didn't you know, go right for you right then if it was a choice you owned, and then you know, okay, so what maybe you know next time I could do different, but at least it was your own choice. So I think a lot of the people that follow my stuff and your stuff are looking for that kind of experience.

00:24:27
Speaker 1: Yeah, and you mentioned like your own choice, like it's like, yes, it's gonna you might, you know, be really pissed off whenever things don't go your way after you made that choice, and maybe some people out there just want to blame the fantasy analyst, and okay, like we've we've we've heard that plenty. But also celebrating the wins is so important whenever you're the one who made Like when people come to me at the end of the season, they're like, thank you so much. You got me a championship. I didn't get you a championship. You literally you manage your roster, you did all that. I hopefully guided you in the right direction. But you know, I think there's no better example of what you were referring to with like projections and like just in tiers than then wide receiver this year and in fantasy, well there's there's like eight good wide receivers and then there's just a bunch of them, you know, like there and there's so much volatility this year in particular at the position, and you know you don't feel great about them. And then you know, you publish rankings or all publish rankings in the given week, and you know someone might say, why is this player twelve spots ahead of this player? But realistically there's really not much of a difference in terms of projection between those guys.

00:25:28
Speaker 3: It's just you see, this is my and I'm curious if you agree with this.

00:25:33
Speaker 1: This is where I have always had issues with rankings, is that they're so linear naturally, right, it's just one to two to three to four to five. And so our brains when we look at rankings were like, oh, this one is one away from two, that two is one away from three.

00:25:50
Speaker 3: But and this is why I try to tier all of my rankings.

00:25:53
Speaker 1: But realistically, one to four could be light years ahead of five to eight, you know, and in the way that the rankings are set.

00:26:01
Speaker 3: And I think just naturally, like the way people consume things.

00:26:04
Speaker 1: And this is my biggest issue with like just anything in the fantasy space and why I've never been one to submit anything to fantasy pros and such, because number one, I don't want my content to just be there.

00:26:14
Speaker 3: I want to own my content.

00:26:15
Speaker 1: But number two, like, I don't think that's necessarily like the way that people need to consume this content, you know, Like I think it's just it's too it's too linear in nature.

00:26:27
Speaker 2: Yeah, right, it is a problem. That's and tiers are a problem too sometimes because you don't know the drop off because like the different ones, and it's it's a simplification. So that's why I mean everything I do defaults to point projections somehow so that you can actually see and you know, I got the color gradient as in my classic thing since eight years, so you can also get a bit of a visualization of that. So no, that's absolutely right. And I used to study like how to get people to understand in terms of rankings as well. I've heard you talk about this before on the show, like just because you predict the number one spot for the week, you definitely don't expect that person to be the number one.

00:27:06
Speaker 3: Yeah, exactly.

00:27:06
Speaker 2: And in my case it I remember finding that, like for DST, whenever I have a number one, the most likely outcome is number ten. Yeah, like, which sounds pretty bad, but you know you're sort of ignoring that someone's going to have a surprise boom, and you know, maybe it doesn't go as well, like there's this huge and that's not because the rankings are bad, but like everyone else I look at is at least that or it's not a little worse when I make compared, So it's like it's just this like constant, and that's what people need to understand. It's like, you couldn't have really done any better, you know, just the way randomness the scatter works out.

00:27:41
Speaker 3: Yeah, I mean it's similar.

00:27:42
Speaker 1: It's similar when people will talk about if a player busted or not in a given season, and they're like, well, I drafted this guy at RB seven and he finishes RB nine, and I'm like, that's a great season for the RB seven, you know, like you if you're getting that kind of production from the RB seven, that's actually that's why I always base everything off of ADP expect right, So just based on how many points we'd expect a player to score at their particular ADP versus how players historically have scored at that ADP. And so if the RB seven typically scores I don't know, fourteen and a half PPR points per game or whatever that number would be, and then that player who you drafted their scored fifteen PPR points per game, then that player actually out even if he's the RB nine or the RB ten, he actually outperformed that ADP.

00:28:26
Speaker 3: Expectation and he was actually a good pick, not a bad pick.

00:28:29
Speaker 1: And so that to me, again going more based on the numerical value as opposed to just the ranking value can go a really long way, not only in the way that we analyze this game, but in the way that people react, Like you're saying to this game, you know, like this guy was a horrible pick because he finished as you know, like he was three spots below where we drafted him.

00:28:50
Speaker 3: It's like, okay, but look at what's going.

00:28:52
Speaker 1: On around him, like everyone else also didn't perform, you know, And that's just the way that fantasy football goes. But but I do want to talk about these predictive models. Let's get let's get a little bit nerdier with this. Okay, yeah, yeah, even even more news say so, but it wasn't my decision.

00:29:09
Speaker 2: Go for it.

00:29:10
Speaker 1: So so you you have these models with like streaming defenses, and honestly, my the way I'll talk about my process really quick just because I think it'll help at least like show the discrepancy between how I approach things to how you approach things. My weekly process with ranking is that I have a very baseline projection that I'll run for every player. But as you noted, like it is a lot of work, Like it's not a button that you click and then all of a sudden, everything is just good for you for the rest of the week. Right, Like when it comes to projections and stuff, there's an insane amount of maintenance.

00:29:46
Speaker 2: The best two days are Tuesdays are crazy, They're nuts, and.

00:29:52
Speaker 1: The best projectors in the space, Like if you go to and establish the run for instance, you know, some of the sharpest people in DFS and best Ball in season long, and they have people that are working on projections, that are constantly fixing projections, and they're having literal discussions about one single player's target share to that, to that, you know, level of detail. And I'm one dude, who's who's doing this? And I'm obviously creating all this other content. So what I found as my best practice for being able to get as much done as I can get done is to have a baseline projection that I'll run each week at the at the beginning of the week. I don't really touch the projections again the rest of the week. It's more so a subjective like look at like then I'll dig in more into individual matchups and stuff for wide receivers, you know, like slot roles versus out wide roles, and see you know, which players just seem to have better matchups based on some of the more qualitative things, you know, just like knowing ball and like seeing what I'm seeing, you know, while watching games and stuff like that. So I use the projections as a baseline. But the difference between what I'm doing what you're doing is that your projections are really guiding how you're ranking these guys. And so I'm very curious as to what's going into that modeling for all these different positions. So let's start with defensive streaming. What are some of the big inputs that you're looking at when you are with your model and when you are ranking these these each week?

00:31:14
Speaker 2: I mean theirs. I'll actually open something up here because why not. But I mean, first of all, the way I run it is that you know, I don't choose I lit the computer. I basically program my own AI. I was doing AI classes at the end of the nineties and it was something, as you know, sort of pulled back wanted to get back into. So it's I made algorithms that sort of train based on and find which inputs are actually most valid. So the maybe a first thing to clarify, I don't know if you deal with this actually, but when one question is always like what is an average, like in the average number of sacks that a team has or interceptions or whatever. And I don't talk about this so much because it's so boring for people, But I spent an enormous amount of time every off season, you know, I don't. This goes all year for me, right like January hits. I'm going to be like revising this module. And what it does is first of all, figure out for each parameter, like what is the best average? Should use the last five weeks of interceptions or rushing yards or the last ten weeks? And should you look at trends or is it you know? And so it's sort of trained to learn on that data to predict like or to optimize what correlates with a future result as well. So so I that's like the first thing that goes into it. Then I have I don't know if anyone puts that much effort into like creating an average, I'd actually be really curious how others do it. But that's that's my way of doing it. It's very involved. And then when it comes to the models, So first of all, maybe I should point out. So I have three DST models, right, because I was always surprised there hasn't been another real like ESPN DST rankings of it, because the decisions are very different from the normal Yahoo sleeper, the one that Fantasy pros measures against with the yards allowed, so I have got it different models for that and also this one that became popular. You deserve better defensive scoring, so the models will be a bit different. Oh yeah, okay, I'm looking at it now. But it's not that I actually switched last year to actually first predicting sacks and interceptions and yards allowed like all of the individual components, and then combining them. That was just something I learned worked better. It doesn't mean that it has to be that way, but it gave better predictions. So then I need to look at all of those. But it's things like, okay, so here we go. Interceptions is based on it's actually more heavily weighted to defensive interceptions, a little bit less towards the opponent the quarterback thrown interceptions just slightly, uh, and how many pass first downs they have and completions. It's also based it's tied to Vegas. So in this case interceptions is related to the spread, Yeah, the spread vegus. I don't think that's a big surprise. And then there's this small little nudge from Monday Games having slightly more. But you know, so it's stuff like that, and you know whether it would be included as well.

00:34:25
Speaker 3: Yeah, indoor outdoor.

00:34:26
Speaker 2: I'm assuming there's yeah, all of that. So yeah, I don't know if I can give anything to like elucidating, because it's really it's interesting to hear about like what is, but I think it is.

00:34:39
Speaker 1: I think it's good for people to hear the amount of detail that goes into that kind of thing because like, I don't know if you like how you feel about this, But I do a lot of prospect modeling, right, so one of like my big if you were to say, what is like the big model that you run and do, it would be the ZAP model, the Zach reason adjust, a prospect model that I do that's just based on incoming running back wide receiver tight end model is not bad, but it's not amazing. I haven't put as much effort into the tight end one as I have. It's just harder because the data set's not as large right as the as the running back and wide receiver one. But I've found through and like you said, like you you started and then you got into it, and then like the first year you were you were manipulating even more than you do now just because you saw you know, it's the same thing with me, Like you're learning constantly as you're in, you're adding and you're and you're altering and changing things.

00:35:29
Speaker 3: And you know, I've gotten.

00:35:30
Speaker 1: Feedback before with the ZAP model where every year, like every year in January, most of my January is spent or a lot of my January is spent manipulating the model and just kind of like testing other metrics that I haven't tested before and just seeing if there's anything to make it.

00:35:44
Speaker 3: More predictive than it's been.

00:35:46
Speaker 1: And sometimes it turns people off because they've gotten results in the past where hopefully it's been good, but now they're saying, oh, well, this result in the past. You know, this player's prospect score was a ninety five and now it's a ninety seven, and because you changed some inputs and it changed.

00:36:02
Speaker 3: It's like, yeah, but don't you want me to be more accurate? You know, Like that's that's.

00:36:06
Speaker 2: My That was a common thing on Reddit for a while, people, why did your rankings change? It's so frustrating that well, you know, suddenly there came snow in the weather report, Like what do you want bad writings?

00:36:16
Speaker 3: You know, right, exactly exactly.

00:36:18
Speaker 2: And I think these inputs actually are extremely interesting. And I tried to share, like I did a couple posts earlier this year on like myth kicks for example, Yeah, and to try to show that like predicting future miskicks isn't just by looking at kickers accuracy. It's based on all these other things spread maybe fumble recoveries and the weather if there's wind like that. There are a lot of reasons that contribute to a miskick besides just previous myth kicks. Yeah, looking at fumble recoveries here, and it's got stuff like playoff chances as a part of it, or the pace of game and stuff like that. And I wanted to mention there was one other thing. I'm just looking at one out of the corner of my eye here. But the way I have it set up to work to give a little bit of extra insight to the model. I actually have differentiated models for high vegas implied scores and low vegas implied scores and also mid or whole range, so I actually have and they're totally different, prim like what predicts a good DST score if their opponent is expected to score high has different inputs like I forget which one and one of them interceptions isn't important and the other one sacks isn't as important like it's It actually makes like a difference that makes sense.

00:37:31
Speaker 3: Particularly that's intuitive.

00:37:32
Speaker 1: Right, like just because if you're in a low vegas game, like it's it's likely going to be a little bit more ground heavy or it's not, you know whereas and I do think that from a defense, And I want to hear your thoughts on this definitely, because like last week, for instance, I remember I had the RAMS ranked relatively high against Arizona, and you know, and I had some feedback like, well, Arizona, you know, is playing well offensively, and I don't know how much I trust, whatever the case may be. And but but I think that they're is like this misunderstanding a little bit by the everyday fantasy manager that you don't just want to chase low points point totals for the opposing team because you want because where we score points defensively is via dropbacks, right like we want because you get a sack, you can need an interception. There's there's gonna be obviously sack fumbles and that's a huge you know, it can be a huge play, and most of those plays are going to be you know, you're not there's really not that big of a differential, you know, unless you're getting like a shutout or something. In most platforms with like allowing fourteen points versus twenty seven points, you know, it's not going to be as dramatic as some of the as a turnover stuff would be. Is that one of the key reasons that you would then have the two separated models of like high totals and low totals.

00:38:46
Speaker 2: Exactly. Yeah, And it's also by the way, for different phases of this season, so different inputs for the last Now we're in playoffs, so it's it's heavily weighted towards even a different set depending on other things off chances and the like as well. So yeah, exactly, it's just letting the I mean really honestly, it's letting the computer explore that space and testing theories. I really liked what you said about the way you update your models in the off season up model, because that's how I think of it. This isn't just like I kind of don't even like the word model sometimes because it just sort of implies somebody like puts some like equations together or something. In a way it is, but it's it's every update is like a small study. It's like it's it's a small bit of research to test whether something is impactful or not. And so I wish I could. I mean, I have so much more to say if people had the patience and time to listen about what interesting factors actually drive this one or that one. But that's how it is, a bunch of micro studies.

00:39:49
Speaker 3: Yeah, so you also do kicker models.

00:39:51
Speaker 1: And the funny thing about the kicker stuff is number one, you know, kickers are obviously there's they're a hot topic every year in fantasy where some and you know a lot of managers just don't want them in the game, and some managers do. I find it just funny because I last summer or two summers ago, I told people on the Late Round discord that I was going to build a kicker model just because I want to keep getting kicker questions. I don't do Kicker rankings at all, Like, I just don't, like it's never been part of like what I offer. And I semi joke that I was going to build a Kicker model, and then I started modeling, Like, I started messing around with it a little. I finally had some data that I got and stuff with Kickers, and I.

00:40:28
Speaker 3: Started looking into it, and I'm like, I'm not I'm not going to do this.

00:40:32
Speaker 1: So I got to ask you with the Kicker model, like, what are some of the big inputs that you definitely, you know, feel like are important? But then number two, is it a harder thing to model than defensive streaming?

00:40:43
Speaker 2: And right, No, I was jeered for even suggesting I would try to make a Kicker model. Originally, back if you go eight years ago, it was like, why would you subject yourself to that? But it's been It's the most frustrating one to do, but it's also brings so much more learning and insights I think into what actually drives it. So I think for that reason, it's always the you know, the curiosity factor is what sort of compels more investigation. But it is definitely very difficult. Like, so first of all, like my DST models. I'm very happy with. I keep like I think they have reached like I don't have when I do my accuracy reviews, I would always have a year where I would say it looks like I'm like third this year or what you know, I like admitting the faults. At this point, it's just been like week after week, it's just you know, and unfortunately it's not like you say, celebrating the winds tangent now. But my problem is that sometimes the real benefit of the model is downgrading the ones that actually are busts. Yeah, people don't see that, you know, because like there were some weeks where I, like, I don't know, everyone else had the Lions high and my momor Broncos or something I might have them a lot lower, but nobody was like, oh, thank you for not telling me to do anyway. So DST model I feel great about. Quarterback model is just like on Fire. I'm so happy with it. The Kicker model is in constant development, and I think it's going to probably be the only one that I really do some other updates on. First of all, like last year it did really well. There was a guy on Fantasy submitted to Fantasy pros my kicker rankings, and it came number two last year in weeks. Some weeks they weren't submitted, so there were some discrepancies, but I felt very good about it. But the game is changing a lot, and that has disturbed my kicker models. So for example, like I don't know, if you go back six years ago, we had the emergence of some of the air raid strategies having impact in the NFL, and that had like you know, that was hard a little harder to predict. There was a bit less field goals from that, but the big one, like I actually the way I started here, a lot of the you act about it. Before, a lot of my earlier kicker rankings were going fantastic, I mean twenty nineteen twenty they were like it was picking out kickers that you wouldn't normally expect and they were delivering. But the times change, and one of the things that changed, especially in like twenty two twenty one, twenty two, was this prevalence of fourth downs, yes, and the increased you know, I guess some clever analysts came up with the idea, you know, so much for them, but to encourage that they should go for more fourth downs than they did, and that was disturbing a lot of the kickerings, so I had to update my model for that. And the more recent one I would say is actually this about the k ball. Yeah, I guess you're not be familiar, but just the fact that now kickers can sort of choose which ball they will use for their kicks, so they're familiar with it and they're going for longer kicks. It's exciting also because we're seeing many more long kicks, making the position, you know, more exciting. But I had a tough time, so my mom has been doing great this year, except for two specific weeks, week three and five, where like it was like you could just turn that ranking upside down. Yeah, you know, it was just awful. This is actually a good example of something where I, you know, I see a bad result, I totally admit it and then dig into the reasons why. You can correct me if I'm wrong. But it looks like from what I can see that a lot more teams were trying out this idea like hey, maybe my kicker is just two. And there were all these teams, like a quarter of the league that in Week one, two, three, those kicks from fifty yards and longer were like more than double the historical amount. It slowed down after week seven, so now we're like back to normal. But it was like the first month was just full of really long kicks, and the model had been trained to pick up on that and say, oh this, you know, the way it had learned from the past. In the past, you could trust that if there were some long kicks, that would make a good kicker choice for the next few weeks. But it turned out like the Bills, Eagles, Giants, Ravens, Rams, Cardinals, Chiefs, like all these teams like had a lot of long kicks early and then none for like like still maybe they've had, like you maybe two or three of those since then, but if they just stopped, and that basically caused these week three and five rankings to say, hey, we want those kickers, but actually they weren't doing that anymore. And I think we're seeing an evolution of the game now where like they need to decide at that you know, midfield range, like am I going to go for a fourth down? Or am I going to take the long kick? And like they're working through that strategy and that you know, so that's one of the challenges with recently with the kicker modeling.

00:45:56
Speaker 1: I was gonna say, I mean, NFL trends. It has to be so difficult right now for so many reasons. Number One, you have the fourth down stuff, and I do want to say, like, have you when.

00:46:05
Speaker 3: It comes up? And I'll get to the other thing that I want to talk about here in a second.

00:46:08
Speaker 2: Two.

00:46:08
Speaker 1: But when it comes to the fourth down stuff, have you gotten as granular as looking at like coaching data and seeing how frequently the individual coach might be going for fourth down or is that something where you're just kind of looking at general team for that season and how things are trending with like their fourth down aggressiveness.

00:46:24
Speaker 2: Yes, it's the second one. I would like to get more granularity, and actually dig in, I haven't integrated like where on the field position are the fourth down and I think that's something I would really like to add into it as well. But I think the bigger issue has become these longer kicks and it works in a little bit like if they go So there's actually, if I can mention one other kind of interesting to me. It's a really interesting fact there because if you see all these long kicks. You think, oh, isn't aren't NFL scores increasing? Then we have like if we're getting more field goals, aren't game scored? And I dug into it, and you know, really it's been so low. Even if it doubled, it would add basically half a point per game, you know, on average, with the amount of increased long kicks. And we've seen that fluctuation before, like the average score can go you know, up one or two points historically. I mean, it's that's reasonable. But I was also looking into what has anything decreased, and actually it has. It turns out that the near term kicks are decreased or have decreased the last couple of years at the same time. So we used to be able to sort of trust that like there would be this sort of like same ratio of short kicks and long kicks. And by the way, this has been This is the exciting part to me, because there's been this narrative for a while that people say, well, you want to pick a kicker who has a team that can get down the field and then stalls in the rit zone. That's been like the historical narrative, and actually now it's going away from that, I mean, and you know that was true for some teams. It was allays a little dodgy, like if other people found this too, But basically, you know, there wasn't a huge difference between red zone success and you know, drive success rate looking up upfield as well. But I think what we're seeing is more of the fourth downs closer to the red zone as well, So you're seeing fewer of those short, short kicks.

00:48:21
Speaker 1: Yeah, And I think to not to interrupt, but I think that that was the other thing that I was going to bring up was that I think a lot of that has to do with the kickoff rules where field position is so different now right, starting to like why would you you know, punt at this point or kick the google at this point and have because because the expected points on a given drive is going to just look different based on where that drive is actually starting. And so you know, to your to what you're saying though about like the shorter kicks not happening as frequently, because teams are then going to be more aggressive and say, no, we're just going to go for an on fourth and three because we have good We'll give them bad bad field position if we miss this I think there's a lot of things that are merging here. It's like, on one end, it's definitely the kickoff rules and you know all that, but then on the other end, it's also what you noted. You know, there's analysts and people who have done analytics that have said this is when you you should be going.

00:49:12
Speaker 3: Forward here more frequently.

00:49:13
Speaker 1: And so it's like this merging of like, you know, just progressiveness in the way that coaches are going to call those plays just based on data, but also the rules in the NFL. But you know, you said, you know, we used to very like shorthand wise with a kicker. Okay, we would say we want kickers in high implied team total teams or on those types of teams that they're probably favorites, you know, like in a good spot or at least like a you don't want a negative game script because then they they'd be going for it a lot, right, and then you know, you want them to be like maybe not as great in the red zone. Is there like a shorthand if you were to just dumb it down like a couple of bullet points or no, the answer could be no, I'm not trying to I know.

00:49:53
Speaker 3: I'm putting you on the spot a little bit.

00:49:54
Speaker 1: Yeah, yeah, But is there like a shorthand of like a couple of things that you know this year rest of season, we got three weeks left. I mean, maybe I'm selfishly saying this too because I have a couple of kicker leagues that are still going on. But is there like a shorthand of like what we should be looking for in a kicker and making that decision.

00:50:11
Speaker 2: I'm going to say no, I don't think you can do that. But that's also part of what keeps it really interesting. I mean, there are a lot of reasons why a kicker could end up with more opportunities, and it doesn't just depend It depends on is does the team you know, pass more than run or not? And I mean the whole dynamics of the team. And this was my initial goal. The reason I was interested in kickers was really to bring in this idea of like the whole team interaction and dynamics into dependencies, not just and maybe there was a time where you could simplify it a bit. I have shared, you know. I called it actually the simple equation where just like you said, you can look at the implied total and the spread the spread matters as well, and make an estimate for what the kicker rules score and that worksably well. But one of the tools I didn't mention that I've added is this stalling curve that I have for kickers as well, where I simulate the likely end of a you know, field position at the end of each drive. And that's something else that I model. So this actually has created This year, I was releasing the first decimal scoring for kicker, So there's a model for that, and it's based on this idea of looking at where the drive stops at the thirty yard kicks, forty yard kicks and so on, and each one of those has a different driver, like, which is really I've posted about this before, so I'm looking through here just I feel like I can pull out a good example here. Okay, thirty yard line depends a lot on those thirty plus yard kicks, depends on red zone efficiency, and if the opponent is punting the twenty yard plus depends on the spread, you know, so as you actually get more field, it's less dependent on different vegas inputs and the like in divisional games, missed points, pace of game. Yeah, So I think there. It's just it's too complicated than there. And that's the only way actually to get an edge one care modeling, and that's why it works to sort of accept like this could go a lot of different ways, but the best guess we have is to actually look at these sort of discrete micro thingk segments, yeah, yea, and then combine them together.

00:52:33
Speaker 3: You've done a lot of interesting studies.

00:52:35
Speaker 1: I'm going to touch on a couple of them, a few of them that you can maybe talk through. Let's just start with a very high level one based on the fact that we're talking about streaming. I've done a podcast about streaming for thirteen years now. Clearly we think streaming can be effective, but is it is it officially better to stream At those three we'll say, oh, defense, quarterback, and kicker versus you know, going with one of those positions maybe when you're drafting.

00:52:59
Speaker 2: Right, you know, these are some of the ones I like to do because that was a question sometimes and you know, you talk about the late round when it was an early round quarterback, and that was a philosophy get a good defense that got you know, basically ten fifteen years ago, overturn the idea and say no, you actually can can stream. So that's something that I definitely like to simulate, model and figure out also at which point it doesn't make sense anymore. So the DST is clearly like you should stream kickers is a difficult one. I mean it changes again with the different years, but in general the conclusion is that kickers streaming can work better. However, I've recently been working on a new analysis I haven't published this one yet, that is reconsidering the idea that you know, maybe at some point you can hold your kicker maybe, like you know, it takes halfway through the season maybe to figure out which teams have a propensity for kicking and they're gonna get And it changes, right because like you know, Boswell, people have loved him, especially last year, and a bit like you it's been around in a decade and it wasn't like he was always you know, So so teams change, they're taking opportunities change, and halfway through the season maybe you can learn. And so I've been looking at does it at some point make sense to look ahead and hold on to the kicker, so you know, simulating that your league mates could pick them up for example, Yeah, and protect against that, and that seems to have some validity I think. So I've I've given you know, posts about holding recommendations quarterback is really interesting.

00:54:31
Speaker 3: It's very Yeah, I'm very curious to hear.

00:54:33
Speaker 2: What the way I look at this one is I do the same thing, and so I simulate, like if your league is sort of holding their quarterbacks and you're the one trying to stream and they're always just grabbing like the best scoring quarterback, Like maybe they give it a first month, that's normally what seems to be the best strategy, Like you try out four weeks and then I think this fits in real life too. Then you see your league mates kind of pick up, Hey, that guy's score a lot, and it'll just pick up the top scoring quarterback. So anyway, I simulated this, and what I look for is sort of what's the cut, Like how many quarterbacks does your league need to own before it becomes a bad idea? Right? And it's it's kind of a toss up here. I think it depends how big your league is. Yeah, I mean if you're if you're in a twelve plus team league and you know your league everyone's holding two quarterbacks, then you know you're scraping the bottle of the barrel, right, So it's it gets tough. But for smaller leagues, if everyone holds two quarterbacks, there's still some good wiggle room that you can you know, still optimize and stream your quarterback. So that's the way I try to plot it, and you know, communicate that. It depends on your league settings. Also, if you have a league where everyone's holding three quarterbacks or whatever, you know, that's a different scenario.

00:55:51
Speaker 1: Yeah, I also think that the quarterback stuff is very interesting because you know, unlike I mean, I would assume it's not like I've dug into the kicker stuff, but I would assume that there are very few kickers where over a five year period that maybe Brandon Aubrey's going to become that guy, but very rarely are there guys who are just like, yeah, we're going to get kicker one production from this guy this season, and we know that from a predictability standpoint going into the year. Whereas a quarterback what we've seen since about basically since Lamar Jackson's first MVP season. I always kind of reference that point because I think the market, the fantasy football market switched in the way that they view mobility at the quarterback position, in the importance of rushing at the quarterback position. And so since then, we've seen the ADPs the average rapt position of these high end quarterbacks that are mobile and these unicorns who are dual threat the Josh Allen's, Lamar Jackson's, you know, the Jalen Hurts of the world who are able to do and I know, I say Jalen Hurts after he's coming off the game that he came off of, but you know, those guys are now being pushed up in ADP, which makes it very difficult for those players to then be available on the waiver wire.

00:57:00
Speaker 3: For us to find or for us to get late.

00:57:02
Speaker 1: And there's always gonna be like, you know, Drake May was a pretty easy, low hanging fruit, late round quarterback target this year. You know, and this is not even me saying in hindsight. You guys can listen to all the you know, shows the in the off season where I talked about him, but you know Jayden Daniels last year where it's like, Okay, yeah, he's probably gonna have some mobility, we can throw that dart. But the thing that I found because we track our quarterback picks on living the stream and when I look at what we've done, what we did from twenty thirteen when we started tracking, or maybe we started tracking in like twenty fifteen through about twenty nineteen twenty twenty, we were consistently QB five, QB six and points per game with our streaming picks. And and for the record, our streaming picks are based on you know, low percentage rostered you know, quarterbacks, and we don't have the benefit of finding a guy and just using him for the rest of the season if he's a great quarterback starter. Right, so we're we're only looking at free agency, so we're constantly and consistently mid QB one.

00:57:59
Speaker 3: And then all of a sudden, when people started to really the market started to.

00:58:04
Speaker 1: Realize that these josh Allen Lamar Jackson players are unicorns, and more of those unicorns came to be within the fantasy space.

00:58:11
Speaker 3: Now we're streaming like QB eleven in points per game.

00:58:14
Speaker 1: You know, it's just not as easy to find those guys off the waiver wire because the market itself has become a lot more in tune with what works at the quarterback position.

00:58:25
Speaker 2: Yeah, that's right, No, So I in my analysis, I did a quarterback streaming post. Was it like around week six this year? Looking at this and you know, I also look at the trends of there are some years where it's just impossible, yeah, and then it's changing. It's yeah as always. So the best you can do is try to find out what is typical and is there a trend is it actually in this direction?

00:58:49
Speaker 3: Yeah.

00:58:50
Speaker 1: So another question, another study that you've done, and I get this question probably a dozen times a year, is should you pair your quarterback with your with his kicker in the same time.

00:59:03
Speaker 2: Yeah. That was when I was really proud to sort of elucidate because there were a lot of there were there was not a lot, but some articles saying that yes, you should do that because you decrease your risk. And the idea there was that you would you know, if your quarterback didn't throw a touchdown, then your kicker would be in position to get a field goal, so you kind of give this offsetting effect. And my finding was, yeah, that's not true. That generally is not true because you know, if you sort of consider the situations your quarterback scores high quarterback scores, low kickerscores high kick. It's actually more likely you're in one of the other quadrants, Like if the game is going well, then your quarterback is going to score high and your kicker will get opportunities and and vice versa. If they're doing worse than so they tend to actually move move together. That was a surprise as well.

00:59:53
Speaker 1: That's definitely a surprise because you feel like there's a there's a finite number of drives in a game, right and scoring opportunities. But you're saying that there is a favorable correlation between quarterbacks scoring well and kicker scoring well.

01:00:08
Speaker 2: Yeah, if you want to sort of it's not stacking, so to speak. But if you want to pair them, if you want to sort of get more variants than you should pair them. Very interesting, Yeah, very very Or if you want to decrease variance, choose the kicker that you're opposing your opponents League Mats quarterback.

01:00:25
Speaker 3: Oh yeah, that's another Yeah, that.

01:00:27
Speaker 2: Was the other way to look at it.

01:00:29
Speaker 3: Yeah, wow, that's very interesting.

01:00:31
Speaker 1: All right, let's talk about the fantasy playoffs real quick, because obviously that's that's the time of year and that's where we're at right now. You have these tools, is there any defense out there, defenses out there that you think could crush over the next few weeks.

01:00:45
Speaker 2: Yeah, I mean some. I I'm glad I pointed out what was it, like my week ten post or something was called Texans hold them and I think that that still holds true. But yeah, you alluded to the tool that. I mean, this is one of the things that was so glad that this was like a year. I don't know if people realize of just releasing a lot of stuff. For me, I was able to have a lot more time in this Like pairing tool for DSTs was one of them. I hope people got good use out of it who are subscribers, but it lets you actually more easily see the ones that are nice to pair. So yeah, for sure, Texans is obvious. I'm looking at it here, So you know, the best ones for the playoffs, I would say, is them that There's been some debate about the Chiefs. I think the Chiefs are a good pick as well. People are worried if they I can hear that if they don't make playoffs, then like will they give up? I don't see that. What do you think?

01:01:39
Speaker 1: Yeah, I mean I think that those narratives are generally a little bit overrated. Like like I think it's sort of a selection bias where like we see it happen maybe once a year with like a certain player or something like that. My general stance is that these guys are NFL athletes.

01:01:53
Speaker 3: They want to play. This is their literal job, you know.

01:01:57
Speaker 1: And maybe where I get nervous is when and there's a player on a team that's out of it and an individual player gets hurt, like I've been nervous about like Garrett.

01:02:06
Speaker 3: Wilson for instance.

01:02:07
Speaker 1: You know, he he gets injured and he's out, and you know he could he could maybe hypothetically come back, but he probably won't come back until he's totally one hundred percent. Whereas you know, if there's a team that's that's in the player that's that's competing to get in the playoffs, if he's eighty percent, he might go out there and play, and that would still be more valuable for your fantasy team than you know, someone else that you've been replacing him with. So that's kind of my stance is that, like I don't I don't know if teams like really just give up and and you have to you know, there's eleven players on the field on a team, and like, like you have to collectively give up, you know, like this isn't It's like if one dude, like you know, isn't playing at as one hundred percent and his full strength, like that means that other guys have to And yeah, that might impact things a little bit, especially if it's like a quarterback. But I think quarterbacks in particular have too big of egos to do that too.

01:02:55
Speaker 3: So I'm just not I don't know.

01:02:57
Speaker 1: I try not to let that over because I think it's very easy to, like in those instances where we have these narratives, we have these thoughts, to let it overshadow and almost take over what's rational. Right, you can think about it, and I think it's fine to say, like this might happen, but oh, when you start to do that, you.

01:03:13
Speaker 3: Start to then let that overtake any sort.

01:03:16
Speaker 1: Of rational thought that you have about that particular team and that in this case, that defense.

01:03:21
Speaker 2: Right, it's again this idea of questioning what you believed is that my vile on it and then going with that. So chiefs and Saints and giants both have great schedules. They're great ones if you can only hold one dst of course, at this stage of year, the you know, the stashing idea is for sure strategy, so you know, hold the you know, if you looked ahead on you would have the Eagles this week probably maybe you get the Bills next week for example, and then after that Steelers and Lions and and you know week seventeen, so you could stash all of those and make a big dst PI or whatever.

01:04:02
Speaker 3: But yeah, it does.

01:04:04
Speaker 1: I've found that because people just kind of want a golden key for the fantasy playoffs and they're like, hey, give me a defense that I can just play.

01:04:11
Speaker 3: And I'm like, but but why limit yourself?

01:04:13
Speaker 1: You know, like, why why do that when there's probably gonna be some defense available that can be better than that one defense for that week.

01:04:19
Speaker 2: You can probably get them, or at least you might want to block your opponent from getting right. That's that's really the Yeah.

01:04:25
Speaker 1: On are there any kickers in the fantasy playoffs that people should be interested in?

01:04:29
Speaker 2: Yeah, so I've done a whole playoffs thing as well for them. I mean, yeah, it's something obviously, you know, probably their own. But I've been talking Myers and a fair baron if he's healthy, and obviously those are like the obvious ones if you find yourself streaming. You know, I think it's interesting with Baits. Actually, people are pretty down on Baits because of the whole fourth down thing there was, you know, I know, like two three years ago or whatever the Lions were especially, you know, their kickers were suffering because of the whole fourth down thing. But that hasn't played out since. Since their bye week, They've averaged like nine and a half points I think, or something like. They've been a like Basis had like some six point weeks, but a lot of twelve to fifteen point weeks as well. So I still you know, people say, why do the rankers keep putting them that high when they go for fourth downs? But I think it's a good to pick up otherwise. Yeah, Little and McPherson. Yeah now then too, Yeah, because I've been down on him for some weeks and to the ire of some people, but.

01:05:28
Speaker 1: Yeah, yeah, I mean he's now now that Burrow's back too, you know, you got Yeah, I'll be totally honest with you. I'm to the point with kickers where I don't even know starting kickers when the season starts, you know, like I, I there are a guy where I literally am learning who the kickers are because I it's just been but but I you know what, this conversation and other conversations, I'll make it my twenty twenty six resolution to get better with kickers.

01:05:53
Speaker 2: Well, I don't even look at them on I mean a lot of people do go with the name, like right, you'll draft an Aubrey, and that's great, sure do that. But like people were really surprised when I said pick up Strader at week one and he turned out, you know, to be an amazing pick. That same thing happened a few years ago when people were down on Nick Folk. I can't remember exactly which year, but you know, but but you can still. That's why I don't always need to look at the specific kicker and his name and everything. What is the situation around him, what is the you know, is there a new quarterback in town? Like uncertainty at the quarterback position. That's been a really so I've been really big recently. I'm looking at other kicker drivers and that that turns out to be an interesting one as well.

01:06:36
Speaker 1: Kind of is that due to the fact that the team not maybe having confidence in that quarterback and then just go or.

01:06:41
Speaker 2: Just not knowing like the what I found. I tested this because I realized that I rarely had a good accuracy week with kicker and quarterback at the same time, like I for whatever, and I thought, could could they actually be related? Like and then I looked at and I actually model the likely variants of quarterback comes by and I actually test, and that seems to be actually a significant input very in the wider range of outcomes that a quarterback can have. That that that helps the kicker apparently interesting. Yeah, these are some of the directions I keep trying to poke and push and yeah.

01:07:18
Speaker 3: It's very cool.

01:07:18
Speaker 1: All right, Before I let you go, I want to I want to hear a bold prediction that you have. I do this with every perspective's guest, So give me a bold prediction for week fifteen.

01:07:27
Speaker 2: Yeah, I think I got to go with one that's kicker or dst right, So I one thing we didn't get to mention, just as a quick segue here a side note is an interesting study that I did was looking and I when do kickers score these like twenty point below game? Right? Like when does that happen? Because that's really frustrating as a user, as you know, like why did the rankers miss that? I mean it's like twenty two points? How did they not see that coming? Yeah, And one kind of satisfying conclusion I got to I think I'm still working on this one, is like they happen to usually be those like big point games and when the team is expected to lose. So normally, in your right mind as a ranker, you wouldn't in your right mind put that kicker like in the top half. I mean maybe you would, but but and then you know, you can still get of course, fifteen plus point games from if they're expected to win, but like those big games often seem to come from the lower ones. So for this week, I'm gonna go on Limb because you asked me to and say, right, Card, maybe I think he's the one that might have the blow up game. I mean, it depends so much on you know, McCarthy, but yeah, you know, with the uncertainty there, maybe he's the one that blows up this week.

01:08:43
Speaker 3: I like it.

01:08:44
Speaker 2: And then at DST, I think I'll you know, there's a narrative here with the Ravens and what's going on with them. I'll say that they have a chance to get some defensive touchdowns again, they get a rematch with the Bengals, which has been giving up, by the way, some defense of touchdowns, and maybe they've maybe they know their number at this point, and you know, with only a what twenty something percent chance for the playoffs, they've definitely got something to play for. So maybe maybe a defensive touchdown for the Ravens.

01:09:15
Speaker 1: I like that call, and I remember there being in the DFS world, like a decade ago, there was a lot of talk about I haven't studied this in research this in detail. Maybe you have, maybe not, but this is something that we both can look at if we haven't. The second time a divisional rival plays each other in the in season, it's typically a lower scoring just not as fantasy friendly type of performance.

01:09:37
Speaker 2: Yeah, that's really interesting and I haven't looked into it, but it always comes up that people say, oh, look at what happened last time that right, you know, Macpherson scored twenty four points, and it's like, well, but that's not going to happen again. I mean, they've got to see each other and it's not something that repeats. There is something's got to change. So I think it's super interesting. I have not analyzed that yet.

01:10:00
Speaker 1: Yeah, I want to dig into it too, because I don't know why I haven't, because it's been in my head for years now. I just haven't, you know, sat down to it because it's it's one of those like very micro things that you know, may or may not even like yield much. But at this point in the data availability and the ease and kind of analyzing that kind of thing, I do think that it could be.

01:10:17
Speaker 2: It could be interesting. I spend my time on all that kind of stuff. Are they going to kick more in an international game? Dude? Tight ends? Score lower after Monday game, like whatever, Like little micro advantages you can get like that.

01:10:28
Speaker 3: That's awesome. That's what makes you you man, That's that's why you've been crushing it. This.

01:10:32
Speaker 1: This has been such an awesome conversation. I think people are going to really dig it. Shout out Reddit, Shout out all the people over there. They get a bad rap over there, you know. There there's a lot of people in the fantasy space that don't like to get in the Reddit streets.

01:10:46
Speaker 3: But I'll say I'll.

01:10:47
Speaker 1: Say this, It's no different on any other social media platform. It's just the way that that's just the way that it is. I think that if you come with good energy, you'll get good energy back. Is my my general vibes. Sometimes I can be unnecessarily you know, rude, and I don't mean to. But if I see that I am, then someone responds back, and I probably deserved it.

01:11:05
Speaker 2: It's such a mix. I am so thankful to the Reddit community. I mean, there have been so many people who really dig the stuff and they want to get these like rational answers and want to know the analysis and everything, and they have supported over the years. I mean, I would, like I said, I would have stopped if it wasn't didn't sort of trigger that that interest continually. There was, of course the other side as well, the toxicity and stuff that that was one of the things where I was like, okay, you know, but you know, the mods have done I think an excellent job, have just toned down a lot. They keep it much better place to be. It's yeah, no, it's I continually look forward to feedback. You know, every bit of criticism that I've got, I try to well first to ask if it makes any sense at all, and then you know, take it to heart and see if I can improve things. And I think that's that's what it's about.

01:11:52
Speaker 3: I've always speaking of the criticism.

01:11:53
Speaker 1: Thing I've always said with Reddit, the thing that I love about it is that when people maybe say something negative at me, which is going to be frequent because you know, I'm publishing content and I'm putting myself out there, and you know people are going to get things wrong based on the things that I say.

01:12:06
Speaker 3: That's just part of the game, you know, that's just part of of what this is.

01:12:09
Speaker 1: And you know, the goal is that we're transparent and like I said, everything at the top, just to bring it all back together, you know.

01:12:14
Speaker 3: The hope is that you know, people are understanding.

01:12:17
Speaker 1: But the thing that I've always said about Reddit that I love is that when I'm on Twitter, for instance, the other big sort of like big fantasy football platform if you will. Obviously there's now TikTok and stuff, but like the place where people use words to speak to one another, uh in writing. The difference is that on Twitter, when people are going to criticize me, they'll do it where they're actually mentioning me and adding me right, so that I see it in my mentions and they're trying to to troll in a way. The difference is that very rarely am I actually getting tagged on Reddit, and so I can see what people are saying about me, and I use like I'll go through, you know, maybe once a week just to see kind of like what the temperature is about the content that I'm doing and stuff and seeing if people are talking about it. Not because like I need an ego booth. I'm trying to get better at what I'm doing, and if someone is saying, hey, I don't like the way that this is being presented or I don't but they're doing it without tagging me, right, and so they're not doing it to get under my skin.

01:13:12
Speaker 3: They're doing it genuinely, genuinely.

01:13:14
Speaker 1: Believe that thing, you know, And I think that's that's more powerful to get better as an analyst and to get better as a as a presenter of information on podcasts and stuff than it is when someone's trying to get under your skin, you know, by by saying that you're this or that and and tagging you in it. And I just want to say, you know, like like it's great that Reddit has now connected us, and uh, you know that you were able to get your voice out there via Reddit because it's brilliant and you're doing amazing, amazing stuff, and I got to give Reddit and then the people have read it in the community.

01:13:47
Speaker 3: Just a lot of props for you know, like giving multiple people.

01:13:51
Speaker 1: Platforms, like you know, like if there's good work, they're recognizing that good work. And it's great to see someone like you or someone like Andrew Cooper or whomever really come out of that and be able to spread gospel about what they're creating in the fantasy space to more and more people.

01:14:05
Speaker 2: That's right, No, I think it's it's it's been wonderful. I would like I do have to combine my rankings with some of the OC you know, the content that I put out there sometimes because it does tend to draw people that just want So that was a trick I had to learn a while ago, Like if I just put out this study like I did yesterday, I should do pair your DST against the other players, Like I think the visibility would have been extremely low on that. Reddit has changed a lot in recent year. People use it almost like Twitter, like so they just read the headline and yeah, yeah, like they just want like and that's been the most frustrating part I think recently for me is that you know, you'll get like Buckey Irving was seen at practice on Wednesday and it's like is that you know, and it's be like a thousand up folks thing, such a low effort, and then I got this, like you know, ten hours of analysis and it's like, yeah, so that's.

01:14:56
Speaker 3: So don't stop. Don't stop.

01:14:57
Speaker 1: It's important to not stop doing that based on that reaction, you know. Like it's it's like I say all the time, like I do this, you know, I'm very thankful that I can now do this and support my family and all that kind of stuff, but like this has been completely drawn from passion and it's never gonna go away. And if you hear me get excited on the podcast and my voice gets high, and I get, it's because I.

01:15:21
Speaker 3: Love this stuff.

01:15:22
Speaker 1: And you wouldn't want me to not love this stuff, you know, like because then you're not getting, you know, the fantasy content that I think you deserve.

01:15:29
Speaker 3: You know, that to me is what's really really important.

01:15:32
Speaker 2: I too have a tendency to get carried away. I don't know if my voice goes high or whatever, but I definitely write too much sometime, and I think, actually, the thing I need to and I'm starting to get better at, is simplifying things, making it contextualized, to sort of really simplify and get the message, because that's what's important to me, is really getting people to get that feeling like, ah, I get it, I got that insight now, And I think that's what you're really into too, and you know that that's rewarding, and it's it's this like empowerment, like Okay, I know how to play the game now.

01:16:01
Speaker 3: Yeah. Yeah.

01:16:02
Speaker 1: And and honestly from from my perspective, and I know that we're kind of just rambling talking about random stuff right now, but hopefully people are digging it. But like, to me, it's like it's so cool that we get to kind of go in this lab and and and and figure stuff out that has not been studied or researched or figured out. And then you're so vulnerable and putting it out there. And I'm not saying okay when I say so vulnerable, I know it's just fantasy football, guys.

01:16:28
Speaker 3: I know that this is not like life or death.

01:16:30
Speaker 1: But it's one of the like you're open for criticism because you're putting yourself out there and you're you've researched all this stuff and you think you have this great idea and sometimes it goes really south, and you know, sometimes it goes really really well.

01:16:42
Speaker 3: But that's so cool.

01:16:43
Speaker 1: It's just such a cool thing to be able to do and to be able to have that platform, and it's important. And again I'm going to stress this, like, it's very easy to be defeated in seeing a threadbro on Twitter, you know, putting some some garbage stuff together, or you know, seeing the up votes for like a really uh easy low hanging fruit post on Reddit. But I do think that it is important to have voices in the space that are doing really really good work and that's exactly what you're doing, absolutely, and it.

01:17:12
Speaker 2: Really helps to have their input as well, like when they I mean a lot of the studies are just inspired by somebody like saying they know something or they have this. Yeah, yeah, the one I got the biggest you mentioned the backlash. I think one of the most controversial ones I put was if a kicker has multiple double digit weeks they scored ten to twelve, fifteen, whatever, Yeah, then they're going to score double digit the following week as well. So I just pulled out data and said, look it actually it basically the attrition rated half, so like, you know, there's like a fifty to fifty chance at any point along a streak that they're going to drop out, and that that's somehow that was probably one of the most like angry sounding responses. I think, what was it myers at the time, someone that they really wanted to believe it was going to continue. But yeah, anyway, but it's just about you know, I don't regret that. I think that's actually it was a wonderful take and I think you know a lot of people really dug what was behind that, and yeah.

01:18:09
Speaker 1: We should be That's what this is about, man, We're having like it's important to have conversations with people.

01:18:15
Speaker 3: And that's why when like.

01:18:16
Speaker 1: People like just just ignore people who are criticizing your work or whatever, it's like, no, I want to just converse with people. I just want to have conversations and get better at what I'm doing.

01:18:24
Speaker 3: If someone's saying something negative.

01:18:26
Speaker 1: About what's going on, I want to know why that's the case, and then I want to dig into the why and see, like, oh, is there a blind spot in my analysis that I can get better? Because I want to be good at this, you know, I want to be able to give good advice and have people listen and keep coming back because the conversations and these relationships are just so important.

01:18:45
Speaker 2: But yeah, like strategy, it feels so much better to play when you know what's going on. Yes, Like, yes, I sorry if I'm just dragging this out, but I did you know another one I just completed recently Again I repeated, was you know, do quarterbacks throw more in garbage time? Like if their defense is bad? Right? And that has been an assumption for a long time, like, oh, you pick them, they're des bad, they got to keep up with the game, And it turns out no, there's actually no basis when you've divide it up, And I think that's just like super interesting.

01:19:13
Speaker 1: Yes, yes, especially because like that's the way and people.

01:19:17
Speaker 3: The other thing to that, again we are rambling.

01:19:20
Speaker 1: But the other thing to that is that when when we have these narratives, we then attribute those narratives to why things go down the way they go down without actually attributing it to the actual quarterback performance. Right, So like Dak Prescott playing on Dallas this year. You know, everyone's like, oh, this is the correctly, this is the the Bengals of.

01:19:41
Speaker 3: Twenty twenty five, right, because the Bengals last year had that horrible defense.

01:19:44
Speaker 1: But the like, realistically, you need a really good quarterback to be in that position more than likely to be able to actually like take advantage of that. It can't just be any old guy that's going to be in these negative game scripts.

01:19:57
Speaker 3: To be able to perform in that way.

01:19:59
Speaker 1: They don't need to be good and efficient, you know, and and that, and so like you're you're saying, oh, Joe Burrow was was good because he had that bad defense, So like you still have to be a good quarterback to be able to do what he's doing.

01:20:09
Speaker 3: From a numbers perspective, you know.

01:20:10
Speaker 2: Yeah, exactly, And it's not correlated necessarily with the defense. It is, however, interestingly, and because you can do that with the data actually divide it up, it is correlated with the opposing offensive power.

01:20:21
Speaker 3: So yes, actually right, actually high high game totals and all that.

01:20:24
Speaker 2: So the idea of a shootout or whatever, that's valid.

01:20:27
Speaker 3: Yeah, man, this was great.

01:20:29
Speaker 1: Let everyone know where they can find you, and and the things that you're working on and what they can look forward to.

01:20:33
Speaker 2: Yeah, I'm going to keep up fighting a good fight. And yeah, there's like you said, no stopping it. It's just too interesting stuff and there's a lot more to come. So I'm really excited also for for the next year. This has been one of the best years. I think it's going to be hard to yeah to the top, but keep trying.

01:20:50
Speaker 3: Yeah, you got to stay ahead, got to stay ahead.

01:20:52
Speaker 2: So subvert it down dot subverted down dot com.

01:20:55
Speaker 3: Yes, yeah, awesome, man, I really appreciate it.

01:20:58
Speaker 1: As always, you guys can check out my stuff lateround dot com. Make sure you subscribe to the YouTube channel YouTube dot com slash at late round f F but otherwise everyone, thanks for tuning in.