373: Revisiting the Process of Building Projections

00:00:02
Speaker 1: This is the Late Podcast with your host JJ zacher Retha j J Zacherson.

00:00:11
Speaker 2: What's up everyone, It's JJ Zachares and the editor in chief at fandel and at number fire dot com in this episode three hundred and seventy three of the Late Round podcast. Well and many shows that are part of the FanDuel podcast network. Thanks for tuning in. One of the very first Late Round podcast episodes was with ESPN's Mike Clay and it was on the topic of building projections. Since that episode a few years ago, I've gotten loads and loads of questions on this topic. They've ranged from how do you project efficiency? To how do I start building projections of my own? So I thought it would be good to do another show on the process of building projections, and this time I wanted to bring on my buddy Rich Reebar from Sharp Footballnalysis dot com to talk about it. Rich is my sounding board for fantasy football thoughts. If I have a weird idea about something that I want to investigate, or if I found something interesting when looking at data, I go to Rich. I run it by him and see what his thoughts are on the topic. Because he's one of the smartest fantasy football analysts in the business. He just gets it, and that goes for building projections as well. I sat down with Rich to talk about his projection building process, how he thinks people misuse and misunderstand fantasy football projections, and we talk about how he specifically projects things like volume, efficiency and touchdowns. We try to cover it all on this show so that you, if you're interested, you can start building projections on your own and taking your fantasy football game to the next level. Now, before get into the interview, I want to let you guys know that Fandel's Sportsbook is giving you a chance to get in on the action. Risk free. Create a Fandel's Sportsbook account today and place your first bet. If you lose, you'll receive a refund of up to five hundred dollars in cite credit.

00:01:50
Speaker 1: It's that easy.

00:01:52
Speaker 2: For more details, VI was It Sportsbook dot fandle dot com offer valid for new users only. Must be twenty one years old. Gambling problem call one in gambler. Now sit back and get your mind ready to listen to Reebs dig deep into his projection building process. All right, Rich, let's start with sort of more of a philosophical question about projections. Do you think that projections in the fantasy football world are misused and maybe overused? Or maybe to put this another way, what are some of the flaws of drafting your fantasy football team via projections.

00:02:26
Speaker 3: First of all, let me say that I'm flattered that you had me on this show about projections. You know, I'm probably not someone in the industry thinks it was like a pantheon on this topic. But when you think of projections that you think guys like mke clay. But projections themselves have taken such a root in fantasy now. You know every site has a lineup optimizer, you know which is projection based. You know, the big sites like Roto Grinders, you know, Osmo, Fantasy Cruncher are all like people that are just on the champions of this topic. I'm also pretty sure that you're big enough now you probably could get Nate Silver to come on this show and break down projections with how you've gone. But you and I frequently behind closed doors and if there ever was a humble brag, you know, talk about fantasy and share a lot of projection stuff.

00:03:10
Speaker 1: So you know everyone everyone gets me here.

00:03:12
Speaker 3: So let me get back on track here, before we need a Denzel and Chris Pine to forget us back on the on the right track for sure, though we could carry that, this could carry the whole show, this topic, I mean, but definitely the way you know, rankings and projections are presented and consumed in the fantasy industry are along running kind of you know, pet Peeve, I'm guilty of also playing a role in there, despite my best efforts to bridge that gap. So if you want to call me a hypocrite, I mean, I could take that. But you know, sometimes you got to play the game the way it is played. But today's fantasy landscape is more statistically saturated than ever, something I'm definitely grateful for and I'm sure you are as well.

00:03:48
Speaker 1: But that has pushed more.

00:03:49
Speaker 3: Work towards you know, models and a linear consumption and presentation of fantasy football in general, and we're losing a lot of big picture focused now on game theory, understanding and applying ranges of outcomes some other ticks that are.

00:04:00
Speaker 1: Part of you know, being consistently good.

00:04:02
Speaker 3: You know, Fantasy Player and Dynasty season long more devest no matter what you're playing, and you know, projections and rankings if you look at them, they typically just resemble what just happened, which we know is already you know, kind of a fallacy. But the strongest correlation coefficiency metrics and fantasy football still carry and maintain, you know, a mountain of variants. We're trying to be in a strike zone of projections to help us get there. But they and the projections and subsequent ranks that come from the projections are often presented and consumed as absoluteses, you know, but really projections, like I said, are just trying to get in that strike zone, you know, probability instead of you know, trying to split three arrows down in bull's eyes.

00:04:37
Speaker 1: You know.

00:04:37
Speaker 3: So even when we nail projections and we're dead on accurate from a seasonal stance, which I believe is easier than to be accurate in a weekly stance, those numbers actually mean per player, you know, what those numbers actually mean for a fantasy player is also inherently variant because you know what they how they contribute to season outcomes. Since the majority of players basically thrive on having pockets of spice spike the production and valls, you know, over abundance of guys that are just weakly wired to wire true fantasy assets carrying your team. So that's why I've always kind of done my tiers. You know, how players score versus a linear list because arbitrage is driven strongest by the manner in which production is accrued, and then the order of those players the rankings is driven by opportunity you know, on on that comes from a player and team level, which can fluctuate. So I just want to know how player is being used firsthand, and you know, then the opportunity level, you know, gets gets weighed in you know, those rankings. But if you're drafting solely on projections as an absolute, you know, we can lump rankings in even though it's in a ranking show you're missing a lot of the whole picture and limiting yourself to one outcome that needs to be accurate, which you know we've kind of shown we're not great at.

00:05:41
Speaker 1: Yeah.

00:05:41
Speaker 2: Well, speaking of the act of building projections, you know, people who don't think projections are helpful, they might say that it's because there's subjectivity involved in building them.

00:05:51
Speaker 1: So do you think that there's.

00:05:52
Speaker 2: Any substance to that sort of claim, and what do you think is helpful about the act of building projections and having them.

00:05:58
Speaker 3: It comes down to who you take you know, your information from, and their process and building their projections. You know, there are threads of subjectivity and just about everything fantasy oriented, and I do believe a lot some retrofit their models in a sense for past hit rates over true predictability, and some people will build you know, their model or projections to you know, one tell them themselves the story that they want to hear. When like we said, that's what rankings are for, you know, it's the rank That's what the rankings are for.

00:06:24
Speaker 1: The difference between rankings.

00:06:25
Speaker 3: But I would say for me personally and how I build and use projections that in fact the opposite is true. More often than not, my projections are causing me to go back and re evaluate players I'm subjective subjectively higher or lower on than what my projections are suggesting. You know, they're kind of a speed governor on subjective influence the way I use them, and that's how they should be used, you know, from a drafting, from a seasonal stance, you know as sobering, sobering checks and balances, you know, process of understanding how top down team output is distributed. And opportunity among the players on those teams. This especially applicable in this current moment when we're fresh off the NFL Draft. In every NFL draft pick is viewed through the tunnel vision of having only ceiling outcome not only for their careers, but immediately in twenty twenty. So, I mean, you know, they also should be used the highlight things like you or the market is overlooking. You know a guy we talked about frequently, you know, last year with Alan Robinson in that light last year, you know, that was something that definitely paid off for both of us. But there is no doubt that, you know, the projections process itself has policed my subjectivity more than vice versa.

00:07:28
Speaker 2: All Right, so let's talk a little bit about the projection building process, the actual act of building, because we know, as we just talked about that building projections can at least be helpful and showing us different paths and where a ceiling and floor is for a particular player. And people listening to this right now are probably most interested in how they themselves can make projections, but they have no idea where to start. So where do you think they should start? How do you think that they should be thinking about building projections. Should they just be looking at players individually, or should they be looking at things sort of more holistically, you know, from a top down, team level first approach.

00:08:05
Speaker 3: I mean there's a lot that goes into building projections, especially for my and I'm sure people like I said that that I mentioned earlier show them a more simplistic way. But I use you know, historical league trends and then scale down for team production and trends you know, which are coaching staff and player influenced. Then scale down to you know, into the actual you know, players getting the opportunities. So from a league perspective stance, you know, league average, you know, yards for play pass rate, scoring rates, yards per attempt are all significantly higher than they were just ten years ago, and where a lot of these coaches and coordinators were actually still part of and people still apply to the current era. But you know, that's where I start, you know, to get a surface area. Then I go team by team by play volume first, you know, and I use a lot of expected wind totals to generate you know, potential game script for these teams. And then once they're it's all about you know, diving up the opportunities and then the players giving those opportunities.

00:08:54
Speaker 1: I mean, but game.

00:08:55
Speaker 3: Script literally drives everything on a play calling stance in the NFL. I mean, you just look at the the San Francisco forty nine ers who just went to the Super Bowl. Kyle Shanahan and his system. We know what a system is, but their run rates since Kyle Shannon's in there have been thirty nine percent and then forty nine percent. Last year they get up to fifty percent because of game script. You know, that's how you get those big spikes. We know inherently what Kyle Shannon wants to do, but you know what they can do afforded by where the team's positioned on the scoreboard is a different story. So we've got a few of these teams. I can give a couple examples of. We got one new a new head coach, new system in a new place, which is always interesting to kind of you know, involve in projections, Browns hiring Kevin Stefanski, you know, and the expectation of limited value. You know, people see the Vikings only ran nine hundred and seventy plays, near the bottom of the league. He has only got twenty one games as a sample as a play caller, you know, the Vikings only through four and sixty six passes a team last year. That a lot of people see that and they say, well, this is what the Browns are going to be. But you know, Minnesota was a ten and sixteam. They were twenty seventh the NFL. Like I said, in offensive snaps, mind you, Cleveland only ran three more plays than the Vikings last year. But the Vikings were fifteenth and overall on offensive pace, but they led for forty seven percent of their plays, which is seventh in the NFL. But if you look under the hood, Minnesota actually played really fast in neutral negative game script. They were twenty fifth in offensive seconds per play in the second half of games, but they were leading in the second But when leading, they were twenty seventh in pace. But if you look at from an overall perspective, they were eighth in the NFL. And first half offensive pace period whether they're winning or losing, and then at any point in the game trailing, they were sixth in offensive pace. Now, given the limitations of the Minnesota roster, you know, am Dealing only played you know, forty three percent of snaps, and what they had offensively outside is Defind Diggs and Dalvin Cook, who also missed two games. I mean, you bring that pace and that fast pace that they had, and that's encouraging when you take into the account of if you're not expecting Cleveland to be one as good as Minnesota and in terms of controlling games, and then they've got better, you know, offensive players, I would say, objectively than what Minnesota had last year from a talent perspective. And then another one wrote real quick is you know, the Packers are a really good team to look at misregard, especially since we're fresh off the draft and the assumption based on their signal from the draft and not even that actually what they've told us publicly, you know, the desire to run. But we should objectively be expecting the Packers they actually potentially have to throw more in twenty twenty than they did in twenty nineteen, which is kind of another ridiculous component to their approach of that draft.

00:11:21
Speaker 1: But you know, they were.

00:11:21
Speaker 3: Thirteen and three despite their you know, expected record should have been ten and six. They went nine to one on one score games. Early lines have them at eight and a half over underwins. They go from having our at sharp football on alysis our thirteenth easiest schedule in twenty nineteen to the eighth hardest this year. They're another team they led for forty eight percent of their snaps, a sixth in the NFL. They led for fifty seven percent of their fourth quarter plays. But I can take that projected win total now and I can drop it, you know, up and raise it to you know, teams that historically have gone from eight wins to ten wins and say if they under undershoot or overshoot, and look at the potential you know, game script that that team might have, and it shows to me that they're they're probably going to have to roll more and being more passing game scripts than they were a year ago, regardless of their initial you know ram Rod philosophy that they're entering this offseason with.

00:12:08
Speaker 2: Yeah, something that really sheds light to this too, is that if you look at previous season past to rush attempt ratio, just overall ratio. So you mentioned the Vikings for instance, I think they were twenty ninth in the league in overall ratio, and then their neutral script ratio is like twenty third or something. So obviously there's a difference, meaning you know, when games are close, they're a little bit more pass heavy. But if you look at the predictiveness of overall ratio year over year versus neutral script ratio, there's a much stronger correlation with neutral script ratio from one year and then the next year overall ratio versus overall ratio from one year to the next. And that's what that's basically telling us is that game script will fluctuate a lot year over year because the NFL that just has a lot of inherent variance. There's just a lot that changes year over year, which is why we see, you know, teams like the forty nine ers come out of nowhere and they're very, very competitive. But that's really what's key is that you have to focus and you have to recognize game script. And one way of doing that, as you alluded to, Rich, is looking at projected total win totals, which you can get, you know in places like FanDuel Sportsbook. There's there's my plug, there's my.

00:13:14
Speaker 1: Plug for the show. Yeah professional, I'm.

00:13:16
Speaker 2: A professional here. So let's look at player projections then, because they really come down to how you project volume and how you project efficiency. You know, those are the two big things when we're projecting players. So we already talked about that, you're starting at the team level. Once you get those team level numbers down, you know, plays per game, et cetera. Now you're looking at the player level. Now with volume, we're looking at market share statistics like rushing share and target share. So how do you properly assign those numbers, those volume numbers to players. What's sort of your process for that?

00:13:47
Speaker 1: This is interesting.

00:13:48
Speaker 3: So when I first started out, it used to be like we hit on the open, you know, more coordinator base and looking at this historical how players were used.

00:13:57
Speaker 1: In specific systems.

00:13:59
Speaker 3: But one thing I've scaled act on significantly over the course of time is overfitting those previous coordinator coach usage and distributions per position.

00:14:06
Speaker 1: On a blanket level.

00:14:07
Speaker 3: I found that, you know, over the time of doing this, that that personnel dictates more of what a team actually does in the distribution department, over coaches setting out to get specific positions the ball more that's generally driven by the talent level. At those positions, there's still a framework of their coaching scheme and philosophy applied. What goes into things, you know, such as the actual personnel usage distribution, like snaps played, that can be applied at the foundation level, like we talked about Kevin Spansky, Like the wide receiver three for the Browns is probably not going to play a ton of snaps this year versus other teams, whereas the tight end two will. But the ball goes to the best players in the offense more often than not, regardless of the scheme and the coach's prior history, which I said is largely you know, just like that, personnel driven in the.

00:14:49
Speaker 1: First place, So I tend to say in the framework of.

00:14:51
Speaker 3: The scheme, and then distribution to talent and work from there. Also I'm looking at, you know, per game opportunities on a player level. That is, how often players that were injured the year prior or just disappointed on opportunities slip through the cracks. I'll that Alan Robinson, you know, a year ago, you know, you know. But the great, the great part of getting to the usage stage of projections is when this is now where you can start to scope out your ranges of outcomes for players. Do you want to see what Clyde Edwards Hilaire looks like on sixty percent of the team carries and fourteen percent of the team targets the average of what Damian Williams had over his final seven full games last season. To scope what a ceiling, you know, outcome of usage could be. What about a floor You can adjust for any type of split when you get to this level, and this is how you get your range with outcomes. But that's how you get your floor, your media and your ceiling for players to gauge. And this is how then you can apply risk and upside, you know, to draft those players and into their cost and NFL you know, draft the fantasy drafts, the ADP especially ones you know that have usage like I use Clyde or Hilaire that will very likely have projectable variants on a weekly level compared to an overall season one.

00:15:57
Speaker 1: So that's where you can kind of play with that. But I definitely I'm.

00:16:00
Speaker 3: Looking at more on you know, a talent and player level, you know, over the coaching level when I get to that point.

00:16:07
Speaker 2: Yeah, And you can also use sort of league wide trends and what we've seen historically from players to know, you know, a wide receiver three is likely not going to hit a twenty percent target shair in an offense. We just know that's likely not going in most offenses. Right, Even if you don't know exactly how those offenses look, if you're not digging into the offensive scheme, et cetera, you at least know where certain certain thresholds are hit. Like, for instance, elite wide receivers are almost always going to see at least a twenty five percent target shair, right, And so you can look at DeAndre Hopkins moving over to Arizona. You know that they have a different type of offense where we're going to see four plus wide receivers on the field often. Maybe that saturates things a little bit, but he's still DeAndre Hopkins. He's still going to see a really high target chair. And the same was true for Odell Beckham last season. Where if you have an elite talent, and you know that it's an elite talent, at least have some baseline that you can work with and you shouldn't be projecting then players for you know, thirty five percent target shairs and forty percent target shairs even though those things can happen. You know, the Brandon Marshall back in the day getting that forty percent er Michael Thomas and what he's been able to do. It can't happen. But we're also looking at the most likely outcome whenever we're projecting these things. And then, like you said, Rich, you can sort of go through and look at the range of outcomes for these different players if you want to bump them up, and just see how that affects their projection in general. That's the beauty of building projections. So I'm gonna use that same question, but instead of looking at volume, let's look at efficiency. We know that efficiency metrics aren't very sticky year over year, volume metrics are stickier than than efficiency metrics are. So how do you project things like yards per carry or yards per target?

00:17:48
Speaker 3: So using stats for you know, individual players have a good amount of you know, the usage stats have a good amount of correlation. You know, in context for running back snaps, touches, you know, touches per game, those are some of the stickiest year over year metrics for running backs, Targets per game, target share per game, and you know, average depth of target are sticky. Wide receiver stats, so we can use a lot of that. But then when you get to the yards per carry, yards per catch guards, for target touchdown rates, all those are like a mountain of variants. You know, definitely a lot of a lot of variants in her year. So for efficiency, I stay within the parameters of like the league base rates based on opportunity when and you know where the touches are coming from, you know, carries on the field. I said, average depth of target is very sticky. That's very useful in figuring out you know, catch rates and you know yards perception, and then within that I will also will then you incorporate like the base rate of the individual players career output. You know, some players are better with the targets that they're just our better players and worst players in the NFL.

00:18:49
Speaker 1: So you know, then I'll work from there.

00:18:51
Speaker 3: So a guy can have spikes and like how he's used, you know, like Stefan Diggs was a guy at a huge average depth of target spike just kind of out of left field, and how he was used last year compared to previous But for the most part, that stuff stays sticky and we can kind of build off of their understanding. You know, where guys are getting the ball on the field, and you know depth, photographer, ride receivers, and then running back carries is just like I said, I'm worstly hanging on to the league base rate and then you know, incorporating some player base rate stuff to get in that strike zone, you know, in player context. That's really basically it. I mean, I'm really just operating on like a really league wide level and like player level and operating in that strike zone instead of going, you know, in an absorbent way of saying, I'm going to predict a guy to have twenty yards to catch when he's only averaged fourteen yards per catch, you know, over over his career and stuff like that. So I'm really just operating within the parameters of like the base rate for the league and the player.

00:19:40
Speaker 1: It's really just as simple as that. I wish I had a lot more to get into.

00:19:44
Speaker 2: No, No, I mean that's exactly that's the approach that I take as well. And I think that though this is a good time to talk about regression a little bit and how people will ask questions like if I say, AJ Brown is my wide receiver fourteen this year, I don't even know what he is, but let's just say that that's the example, and someone might come to me and they say, well, isn't AJ Brown supposed to regress? And I'm saying yes, that's based on regression. We're looking at someone like AJ Brown's yards per target last year. We're seeing that players who have that high vards per target rate, even though we only have a one year sample on AJ Brown, the chances that he's able to maintain that efficiency year over year is very small. Basically, it's not going to happen. And so you can look at history and look at historical trends with not just rookie wide receivers, You're looking at all wide receivers and how they just generally regress, and then you can put that into your projection and realize that AJ Brown is not going to maintain that yards per target rate, but even still, you know this is what his projection is going to be, and then you can talk about AJ Brown a little bit more intelligently. And I think the other thing too, that this goes to touchdown rate. And this is a nice segue as well, because touchdowns do drive fantasy football. But with touchdown rate, for instance, with quarterbacks, we know that quarterbacks are going to have a wide range from the from the worst starting quarterback and touchdown rate to the best, there's going to be a fairly wide range. Let's say it's going to be seven percent, six to seven percent each year. But we know that those numbers start to regress and kind of meet in the middle year over year, even though we're going to see that same spread in one particular season. But I'm curious, Rich, because touchdowns do drive fantasy football. Do you have a specific way that you project touchdowns?

00:21:27
Speaker 3: Yes, I mean I keep eating the same drum here. But you know, we're going back to the team level. You know, we're looking at offensive production. Over the past decade, offensive yardage has an R squared of zero point seven to zero two five in relation offensive touchdowns, meaning roughly seventy percent of all offensive touchdowns can be explained by yardage gained, which we can get through the measures that we just talked about the question before in terms of performance on opportunities.

00:21:51
Speaker 1: And then I want to.

00:21:52
Speaker 3: Look at, you know, team touchdown distribution. I do an article on this every year too, which you could check. You know, league rate of touchdowns scored has a sixty five to thirty five passing to rushing distribution. Some play callers run a little bit Hotter. You know, if you look at Bruce Arians, you know his offense is typically run around seventy percent you know, passing touchdowns and rushing touchdowns. But you know, in the article that I just did last year that was about last year's you know, I was projecting the Bills and Ravens they have more passing touchdowns last year base on what they did the year prior, because they were an outlier. And you know, not did I have Lamar Jackson projected for thirty six passing touchdowns? No? Did I have Josh Allen projected, you know forty. But those guys are on my radar to have more passing upside than they were given credit for their first initial years then there than their rates were. You know, Lamar Jackson at three and a half percent touchdown rate. You know Josh Allens all his passing stats were disaster his rookie season and he took it significant step forward. That that put me on to those guys, you know. You know last year, you know, Carolina had a league low forty six percent of their touchdowns come via passing. I already told you league a were just sixty five percent. And you look at so historically in some context there of the thirty six teams to be sub fifty percent. Since two thousand and nine, thirty two of those thirty six teams have seen a spy and passing touchdowns with an average of plus seven point three passing scores the following season. Some other teams that ran cold there and expected play calling Arizona, Green Bay, the Rams, all those teams should be expected to have more passing touchdowns this year than they had a year ago, at least in terms of split. So I mean get back to getting getting back to that baseline split and on the flip side of team like Jacksonville, if you and Leonard for an out last year, you already know this pain twenty four of their twenty seven offensive touchdowns came via passing. Pretty unheard of, you know, teams like Detroit, the Steelers, the Jets, there's some other teams that ran overly reliant on passing to rushing touchdowns. So then we're scaling the team and split down from a player and talent level opportunity. Again, once we have like you know, we're using the team yards you get a touchdown total. We're using the base race splits to kind of get a split from passing and rushing touchdowns, and then we're using the individuals that.

00:23:49
Speaker 1: Are going to occupy and absorb some of those touchdowns.

00:23:52
Speaker 3: I do something similar to that, Mike Clay, does you know to OTD and you know similar to you know depth of target, you know where players are getting their targets from. You know, a guy like Tyler Bloyd, He's had six end zone targets the past two years, two and four. He's just not a guy that's gonna score have a large share of the team touchdowns. Couldn't flip and very and have variantce once he's in.

00:24:12
Speaker 1: Yeah.

00:24:12
Speaker 3: Absolutely, you know we as opposed to a guy like DK Metcalf, You know we have one you're seeing me eighteen end zone targets, you know, Kenny Galla and another guy like those guys are gonna, you know, based on the opportunities you're getting, just have more chances to score touchdowns. So I'm using something similar to you know, Mike Clay and is OTD. And why would I not borrow from from Mike Clay. You know, he is the kind of the godfather of the projections. And when I see someone do smart things, of course I'm going to incorporate that into what I do.

00:24:37
Speaker 1: But that's kind of where we.

00:24:38
Speaker 3: Get to, you know, again, team level, scale down down to the player level, work your way down from their same story.

00:24:45
Speaker 2: Yeah, and here's the other thing too, for anyone who's listening to this and you're motivated to make some projections to just get a better idea of the fantasy landscape, you know, to Rich's point, and I've obviously talked about yards to touchdowns plenty on this show, to the point where people tilt their faces whenever I bring it up, because it's not there to be this exact measure and exact way of doing things. But if we start at the team level and we look at how many touchdowns the team's supposed to score based on how many yards they're compiling, you can do it at the player level if you really want to, and then manually adjust based on what you think or what you foresee how teams are going to utilize certain personnel. I'm not saying that this is an exact way of doing it. I'm just saying that it's a more approachable way for a lot of people who might be trying to build projections on their own. So a good example of this might be let's just let's just take the Chargers for instance, with Austin Eckler and Josh Kelly. Austin Eckler sees nine goal line touches last year, scores twice on those goal line touches. We know he's not necessarily a bigger bodied between the tackles running back as opposed to Josh Kelly, who they just drafted, who's sort of the Melvin Gordon replacement, or they're viewing him in that way where he's a bigger bodied guy. You know, he's a downhill runner and he was used a lot in that area of the field of u se La. We can at least think that Josh Kelly is going to see a decent chunk of that goal line work. So if you're doing a yards to touchdown conversion, for instance, you can then get that number, what that yardage total tells you and then bump up Josh Kelly a little bit and bumped and that will naturally, if everything's tied together the way it should be, it'll just naturally bump down Austin Eckler a little bit. So even though we're talking about things like OTD and average depth of target which we incorporate into our models and into our projections. That doesn't mean that someone listening to the show can't take that on sort of a more baseline level, just to get a general idea of how these guys are going to score. All right, Rich, one last question for you. If you could give one piece of advice to anyone who's wanting to build projections, what would that one piece of advice be.

00:26:42
Speaker 3: I mean, don't overly retrofit coaching data that doesn't match personnel. Don't treat projections as rankings, as tiers, you know, use them to highlight potential discrepancies in the draft market. Don't use projections to tell yourself the story you want to hear. Just do the objective job that you're set out to do. That we talk about starting that team level, work your way down to that, get to that strike zone a player level.

00:27:05
Speaker 1: But that's, you know pretty much it.

00:27:07
Speaker 3: So, I mean, listen, I wasn't Nate Silver, you know, definitely you probably wish you did ask Nate Silver at this point.

00:27:13
Speaker 1: No, No, you were.

00:27:14
Speaker 2: You were a perfect guest, perfect guest, Rich, I really appreciate you hopping on for any of you not following Rich in his work. You can check him out on Twitter at Lord Reeves and go read his work over on Sharp footballanalysis dot com. I'm also on Twitter at Late Round QB and if you're not subscribed to this podcast yet, make sure you're searching for the Late Round podcast pretty much anywhere podcasts can be found. Thanks for listening everyone, I'll be back in your ears later this week.