1097: Mailbag - Tyler Shough, 12 Personnel, and More

00:00:02
Speaker 1: This is the Late Podcast with your host, JJ zachera JJ Zachers.

00:00:11
Speaker 2: What's up everyone.

00:00:12
Speaker 1: It's JJ Zachary sin in this episode ten ninety seven of the Late Round Fantasy Football Podcast.

00:00:20
Speaker 2: Thanks for tuning in.

00:00:22
Speaker 1: We've got some great questions this week, and we'll get to them in just a second. First, remember to check out the Late Round Prospect Guide. It's all finished up, and it's been finished since the draft ended. There's well over one hundred and seventy pages of content. Updated player write ups with their draft capital and landing spot in mind, updated ZAP scores, rookie rankings, and more. You can get it over on lateround dot com. Right now, now, let's get to these questions. The first one's from Matt on Patreon. It says JJ you touched on a couple specific tight end situations in yesterday's episode, sadekon Stours and on past episodes, you've talked about the uptick of twelve percent and that we should expect to continue this year as offenses respond to smaller defenses. For example, you've talked about Terrence Ferguson on the RAMS as McVeigh likely being a year ahead of the trend. But then The Rams took Max Clare in the second round, with ten tight ends taken in the first three rounds, which do you take as a bad sign for the incumbents on those teams versus teams just wanting more tight end depth to be able to run more twelve Curious to unpack the situations that these guys got drafted to. So I think this is a really important question. It's one where each individual team is going to have a different answer. But what I'm seeing in a high level bird's eye view is teams trying to move heavier James Gladstone the Jags GM. He even mentioned how heavier tight end sets are a trend in the NFL and that led to their rationale for some of their selections. So clearly, more and more teams are hopping onto this bandwagon. The Jags last year, according to Fantasy Life data, they ran out of eleven personnel at the six highest rate in the league.

00:02:01
Speaker 2: Even though I think.

00:02:01
Speaker 1: Their current wide receiver room dictates that should continue, their draft says that it won't continue at quite the same rate, and we know league wide we're seeing more twelve and thirteen personnel. Now, I do think this is one of those places where reality doesn't have to equate to fantasy football. Teams are often doing things that are in the interest of their team as a whole, not scoring fantasy points. Some tight ends they're gonna be better at blocking, and they might not be pure receivers like an Eli Stowers. So just because a tight end was picked with decent draft capital, it doesn't mean that that tight end has more fantasy football upside than one of the tight ends already on the team. Like in Jacksonville actually think their fifth round tight end has more fantasy football upside than their second round one, and the ZAT model actually agrees. So yes, the amount of tight ends being drafted early in how teams spent their capital in the position definitely signals an uptick and heavier personnel use. Gladstone even said that, however, we shouldn't treat all of these tight ends the same. I know, we all know that, but tight end can be difficult to evaluate for this very reason, because some of these guys are made for blocking while others are made for pass catching. And in fantasy football, sure we won a well rounded tight end because we won them on the field. We want them on the field. A lot, but we want the pass catcher. And then on top of that, anytime a team makes a draft selection, you have to look at their future situation. The Rams taking Max Claire accomplishes two things. Number One, they get more tight end depth because they are running out of a lot of heavier personnel. But number two, smart teams think ahead. The Rams historically have not really used their rookies on offense under Sean McVay. And look at what's happening in twenty twenty seven next year. Colby Parkinson, free agent Davis Allen, free agent Tyler Higbee is signed through twenty twenty seven, but he's thirty three years old and he's not making a ton of money. Max Claire is a good prosper who can play a receiving role, but he's also someone who profiles more as a twenty twenty seven player than a twenty twenty six to one. Now, look, if I'm being totally honest here and I'm giving my more amateur take on this tight end renaissance, I also would imagine that some teams just aren't thinking about this properly. The beauty of what the Rams are doing is that they have flexibility, having a player like Terrence Ferguson can be hypothetically beneficial for them. And I say hypothetically because we just don't know right now, but I definitely see the vision. Ferguson is more of that tweener type. You can get them out there in these heavier packages, but you can shift them to the outside and let them play X receiver at times. You can get really really creative with a tight end like Terrence Ferguson. Now, some teams may not have the same offensive genius that Sean McVay does to really pull something like this off, like he can run tight ends out of eleven personnel effectively. He did that last year, which is why it's extra important to remember to still focus on the things we no matter at tight end for fantasy purposes, which is athleticism and receiving production. Some of the tight ends selected earlier in this year's draft, they didn't really have those two things. Some of them did, like an Eli Rarenan, who I still think is really undervalued in rookie drafts right now, but it's gonna be extremely case by case, and we have to keep in mind through all of this, it's difficult for tight ends to make serious fantasy football impacts. So if there's a player that you believe is just him, that he's got the tools to be a crazy good pass catcher, I wouldn't let a crowded tight end room get in the way of drafting that player, unless you're looking at something at the extremes like Sam Rouch. In Chicago, things are extremely crowded. It's really a unique situation. I'm more thinking about like the Rams or the Jags when I say something like that, where there's no Colston Loveland, chances are the best receiving tight end will still emerge because he's gonna have opportunity to run routes and he can differentiate himself from there. So I don't see it as an automatic bad thing for the incumbents. It just depends on the player in the situation, as it usually does. This A's questions from Twitter's from at Jeff Blaylock. It says, out of many head scratching Day two picks to choose from, which for you was the most head scratchiest. So people are probably expecting me to go off in the forty nine ers picks, but design stribbling. He at least was gaining steam during the draft process, and Kyle Shanahan has done the round three running back reach many times as forty nine Ers head coach. The one for me, the craziest pick in my opinion was Caleb Douglas. At least was Zavian Thomas, who was another big reach. At least you can point to him being a gadgety and special teams type player, and it's on a team where there's a ton of established talent at pass catcher. With Caleb Douglas, you have a death chart that needs serious help at wide receiver. And Douglas is just not a good prospect analytically, and he's one who is projected to go past pick two hundred. Douglas had a twenty fifth percentile breakout scorer, a twenty eighth percentile career yards perrout run rate, a twenty third percentile first downs per rout run rate, and he never saw higher than a twenty percent target share in a single collegiate season. He has size and he's not too old, but man, it's a really, really rough profile to be drafted in round three. I expected Stribbling to at least go in round three given pre draft draft capital and with the running back class being as weak as it was, I'm not overly surprised that we saw a random player get Day two Draft capital.

00:07:34
Speaker 2: Now.

00:07:34
Speaker 1: Did I think it was gonna be Kaylon Black? No, but Caleb Douglas. I hope he proves me wrong. It's just a tough profile to get behind, especially when they almost definitely could have gotten him with one of their fourth round picks. This next question is also from Twitter, is from at d Harwit. It says, how would you compare Tyler Schuck's success last year with Bryce Young's flashes in the prior year? Are there any signals that would point him I'm building on that in year two versus it being a mirage. I don't even think that you have to look at Bryce Young from a couple of years ago. You can literally look at Bryce Young last year, his third year in the league, which was also his best by most measures. And maybe this is controversial. I don't know, because I don't think these quarterbacks are compared very often, but you could argue very strongly that Schuck was the better quarterback last year than Bryce Young as a rookie. I would argue that he was completion percentage over expected according to next gen stats. It was plus point six percent for Shuck. It was minus point three percent for Bryce Young. Yards per attempt, seven point three for Shuck, six point three for Young, epa per dropback, minus point zero five for Shuck, minus point zero eight for Bryce Young. PFF passing grade if you're into that kind of thing. Schuck was at seventy three point one for passing grade, Young sixty eight point two. I said it on the show earlier this week, but I love what the Saints are doing, or what they did in the draft. You surround your quarterback with as much talent as you possibly can find. That way, there are no question marks about how good that quarterback is. But Shuck actually showed some solid signs last year, and he's got more rushing juice than Bryce Young has two. I don't think this means Shuck is a sure thing, but I do love him as a late round quarterback this year. We should always look for year two quarterbacks with mobility, and he has that all while getting serious upgrades of pass catcher and at running back. There's some promise to his overall profile. This best question is from at FPL Mahomes. It says, there are so many people creating models in the fantasy space today, each think that theirs is great. It's hard for people to know what to listen to. Would love to hear more in depth about what makes a good model in your eyes, the concept of overfitting, or how we can evaluate them better, etc. I had someone else who said that they echoed this question, So sure, I'll get to this question. Let's talk about it first and foremost. I'm not a data scientist. I didn't go to school for mathematics or statistics. I was always really good at math. I was a couple of years ahead than the average student in high school humblebrag. And I was able to take a few statistics courses across high school and college. But like, I've just gotten better at this through repetition and learning. When I started my fantasy football analysis journey back in twenty eleven twenty twelve, that timeframe, my statistical analysis was absolutely trash.

00:10:28
Speaker 2: It was horrible.

00:10:29
Speaker 1: I look back and I cringe. There were so many flaws, and I even find flaws to what I did a year or two ago. That's gonna happen. In my opinion, though, this is about learning and getting better with more data and more information. I wanted to preface with that though, because my perspective might be different than someone else's perspective on a topic like this. With that being said, we've seen more and more prospect models pop up over the last couple of years, and I think that's going to keep exponentially growing with A and that's a really good thing from the perspective of getting people to critically think and formulate these things. Where it's bad is some of the things that you mentioned in this question, like overfitting and not really understanding what these models are saying and doing. In math, overfitting is when a formula or a model fits too closely to its training data and it results in a model that just can't actually forecast properly. So, for example, we know that pookin Akua was a really big Day three hit, some model will use his data. It'll use that information when testing it, and it might say, oh, man, we've got to find as many pookin akuas as we possibly can find. And what that does is focused too heavily on the wrong variables without realizing that Nakua is probably just a one of one. You can find some things that this profile had, like an elite yards pro out run rate. But it's very easy for models to focus too heavily on those variables, especially when those models have smaller sample sizes, like in a hypothetical imagine a model had just ten wide receivers that it was analyzing. With the result of a Nakua, that model is gonna think that every hit needs to look just like him.

00:12:12
Speaker 2: He's playing a huge role.

00:12:14
Speaker 1: In how that model is going to forecast, and it's probably going to downplay the role that draft capital has when forecasting wide receiver production. That's overfitting. So how do you know if a model's overfitting? You don't for sure you're not the one who created that model, and you probably don't have the formula staring right back at you, and you don't have the test set, you don't have the data. I'd say one very quick thing to do to look at though, is the combination of knowing what the model is measuring against versus how far it strays away from draft capital. Now, look, every model has a different goal. For the ZAP model, I measure against B two s or how well a player performed across his first three seasons averaging his best two PPR point per game seasons some models look at top twelve, top twenty four seasons. There's no single way to do it. It's just important for a model to have some goal in mind. If someone's just eyeballing things, that's a huge problem. But generally speaking, you can take that goal and you can combine it with how far these players are shifting against draft capital. Because we know, for instance, that Day three wide receivers there haven't been a lot of hits historically, so if a model is training off of history, it's more than likely not gonna see a ton.

00:13:27
Speaker 2: Of Day three hits.

00:13:29
Speaker 1: Now, some models out there, they're gonna just look at bus let's say, and it might say to just fade a player, even if that player has really good draft capital because all of his metrics are just totally egregious. That's not the way that I would do things, because I want to keep more of an open mind about being wrong. But okay, whatever, Some people might fade an early second round pick for a Day three wide receiver, even though we know that Day three wide receivers hit at a way worse rate.

00:13:54
Speaker 2: Than early Day two picks.

00:13:56
Speaker 1: What I'm really talking about here is a bigger problem when you see a Day three wideide receiver with maybe slightly better numbers overall compared to like a Round one wide receiver, and that day three wide out is ranked above a Round one wide out, especially when people are looking at things in hindsight, not forecasting. But when you look at someone sharing the results of their model and you see all these day three wide receivers ranked ahead of what are admittedly Round one bus at the very least, I would message that person and want to learn more, because that's a really interesting result. And when you get really really interesting results that's stray that far away from what the NFL is doing.

00:14:35
Speaker 2: I want to learn more.

00:14:37
Speaker 1: It's not necessarily that they're wrong, it's just that you should investigate more. And that leads me to this next part. I think whoever is publishing the model, whoever's talking about it, if they're talking about it without humility, that's a big red flag. Anyone with experience with this stuff knows that they're going to be humbled really quickly. This stuff is really hard to forecast. Perfection is impossible. And if someone out there constantly touting how good their model is, how accurate it is, and they're only showing partial results and they're not giving you any glimpse at all as to how it's calculated or what the inputs are. That is a red flag. My overall sentiment, though, is that it's awesome to see more models in the fantasy space. I've learned a lot from other people who are also building prospect models. I mean, just listen to last week's Perspectives episode with Addison Hayes. He's got a totally different methodology than how I approach things, but it's a thoughtful methodology and that's what's most important. So do your best to sniff out the snake oil salesman, because unfortunately those people do exist in this space. The last question this week is from Patreon. It's from Norm. It says, when you're doing your rookie rankings, what does your methodology look like When a player you like slides quite a bit, Bell from Miami as an example, Everyone calls it a great value, but I can't shake the voice in my head that said every single team passed on him three, four, and sometimes five times before he was selected. So to reiterate some of the stuff that I talked about this week, just so we're all on the same page. In the ZAP Model database. I have players projected draft capital, which is from NFL mock Draft database, and then I have his actual draft capital. I then use Chase Stewart's NFL Draft pick value chart, and I say, okay, at pick fifty, let's say this is the value of that pick. At pick sixty, this is the value. That way, I can subtract actual real value from one pick to another when comparing projected draft capital to actual draft capital, and that'll tell me how big of a reach that player was for that team. And I do it this way because looking strictly at raw picks doesn't tell us that much. If a player's drafted tenth overall and he was expected to go third overall, that's a seven pick difference. If he's drafted one hundred and fiftieth overall but he was expected to go one hundred and fortieth overall, that's a ten pick diff difference. But clearly the player going tenth overall, that's the bigger value, just given how much those picks mean from a value perspective, how much they're worth. So as we've learned the biggest reaches by using this methodology, they've not been great for fantasy Football. I talked about that earlier this week. I'm measuring this via B two s, which, like I said earlier, it's a player's best two seasons and PPR points per game across his first three years. That's what the ZAP model is trying to project from twenty sixteen, which is when this data dates back to. But from twenty sixteen through twenty twenty three, we've had fifteen wide receivers who saw a loss of four points or more in draft capital value. That's according to the Chase Stewart chart. It's just projected draft capital minus actual draft capital. In this case, these are the fifteen biggest reaches of the position. Only three of those players provided a point or more and B two s over expected. I know this can be kind of confusing, but essentially every draft slot has some sort of expense B two s. If a player's fiftieth overall, they'd be expected to score X, and B two s if they're seventieth overall, that's X minus y. It's less because they're drafted later. Only three of the fifteen biggest reaches since twenty sixteen have truly outperformed B two s.

00:18:17
Speaker 2: That's a horrible hit rate.

00:18:19
Speaker 1: Now, when you look at the opposite direction, the fifteen biggest values, you see six out of the fifteen, giving you a point or more over expected, and B two s expectation. The median outcome for this group is one point one to three PPR points per game over expected. The median outcome for the reaches is like minus a point or so. This doesn't mean that we should really really draft all of the values that we can get, though. We need to keep in mind that we're looking at this versus expectation. When a player is a value, that player is more likely to be a Day three pick, and from that perspective, it's easier to outperform expectation. It doesn't take that much. So instead of looking at all the wide receivers in the model, let's just look at top one hundred picks. If we look at just top one hundred picks, players with top one hundred draft capital, the correlation between the value gained by a team with a pick versus how will those picks outperform expectation in fantasy football, there's very very minor correlation. I'm talking like in our value of point one five, not even our squared it's not super predictive, but it does tell us that in general, at a really small rate, values are better than reaches. But it's minuscule, and what's really doing the heavy lifting here is the reaches, not the huge values. I put a thread together on Thursday on Twitter about this. But there's another way to look at all this. I'm looking at one hundred and six top one hundred wide receivers in this sample. Let's split those one hundred and six wide receivers into four groups. Group one is the biggest values and it includes twenty six wide receivers. Group two that's the second best values. It includes the twenty seven wide receivers. Group three is next that has twenty seven wide receivers, and then Group four is last the biggest reaches, which also includes twenty six wide receivers. We can call these buckets huge values, minor values, minor reaches, and huge reaches. Let's look at the percentage of wide receivers who played above expected versus their expected B two s and let's start with a huge values group. Fifty four percent of them outperform their expectation. The next two groups, though, were at sixty three percent, which means when players were either minor values or minor reaches, they did better than the players who were huge values, and then with the huge reaches the final group, just forty two percent of them played above expected by far the worst of the subset. You can look at this from the perspective of these groups outperforming expectation by one or two points and B two s as well. Overall, it all stays the same. The huge values are better than the huge reaches, but the players in the middle they're the ones who are winning out most. The big values give us a median value of plus point five to one and B two s, so they on average outperform B two S expectation by about half a point. The next two groups, though, the middle ones were at two point nine and three point zero. They were awesome against expectation, and then the huge reach group they were at minus one point. Now, those numbers don't add up to zero for a lot of reasons, but one of them is because the formula is looking at the entire draft, not just top one hundred picks. But this is all still relative. What this data tells us is that reaches are the worst. The guys who are reached for at an extreme rate but huge values, they're actually second worst. The best players versus expectation have been the ones who were minor values are minor reaches. And this makes so much sense. Obviously, when a player's a reach, a team is going against the wisdom of the crowds. That player is probably less talented, so he's not great at performing to expectation. But when a player is a huge value, like you asked in this question, that means the player is passed on by team after team after team, over and over again. That's not necessarily a good thing. Something must be up, and generally speaking, those players have technically been worse than the guys who are just slight reaches and slight values.

00:22:28
Speaker 2: Chris Bell falls in that huge value bucket.

00:22:31
Speaker 1: I don't think it means that you avoid him, but I don't think that him being a huge NFL value should be considered a good thing. From my perspective with Chris Bell, I don't mind him. I see the ceiling vision, but I think people need to recognize that seriously a highly volatile profile. I'll take that on in rookie drafts, especially this year. Rookie drafts are weak, and I'll chase ceiling, but Bell had a completely neutral draft capital delta. His breakout score wasn't terrible, and honestly, that's not super easy to find in this year's class, and I would imagine his draft capital would have been higher had he not torn his ACL. But it's a high variance profile, even if it's a good profile compared to a lot of the wide receivers in this class. So I think your fear is very valid. When a team decides to not pick a wide receiver as high as consensus has him, it doesn't mean he's bound to perform well. It does not mean that he's a great pick. It means there might be something wrong to his profile that the NFL knows and that we don't. That's why those players, while they're not as bad as the huge reaches, there's still scarier bets than the players who are just minor values and minor reaches. That's it for today's show, though, thanks to all of you for listening. If you had subscribed to the Late Round Fantasy Football podcast, make sure you are by starching for it pretty much anywhere podcast can be found, and to follow me on Twitter and on Blue Sky at Late Round QB. Thanks for listening, this week everyone, I greatly appreciate it, and I greatly appreciate all the support with the Late Round Prospect Guide. It seriously means so much. I hope you crush your rookie drafts. I hope you enjoy that prospect Guide.

00:23:59
Speaker 2: Have a great weekend. I'll catch you next week.