Rewind: Superflex Strategy

00:00:00
Speaker 1: It's Late Round podcast Rewind Day, and I've got one of my favorites for you this week. You know, it's funny Monday's episode this week was a long one for the show. It was twenty seven minutes long. And then I looked at my rewind schedule for this week, and on that schedule was this episode that you're about to listen to, this Superflex strategy episode, and this one when I published it a couple of years ago, this one was over twenty five minutes long. So you guys are getting a whole lot of JJ this week. But this episode, like I said, is one of my favorites. It's fast paced and there's tons and tons of information, so I understand if it can sometimes be hard to follow. But a couple of years ago, I did this really big superflex league study. I wanted to show what the true value of quarterbacks was in superflex formats and then provide a simple strategy on how to attack those leagues. So that's what I did, and then I podcasted about it. Now, keep in mind this is more a focus on redraft than dynasty leagues. I'm in a ton of superflex dynasty leagues and the dynamic changes a bit because you're looking at longevity and quarterbacks in those leagues are also currency, so they're more value than how I describe things in this episode. Now, with that being said, the concepts that I talk through can still very much be applied to Dynasty formats as well. I use the ideas from the show all the time, and I hope that you're able to as well. So enjoy this long one. Hopefully your brain doesn't hurt after you're done listening to it, and I'll catch you guys on Friday with the Weekly Mailbag show. Back in twenty twelve, five quarterbacks were being selected in the first round and a half of fantasy football drafts. That seems crazy, and it was crazy. That's why I ended up writing a book before that season about drafting quarterbacks late. It was because the popular thing to do was to get them early, but from a game theory perspective, that made absolutely zero sense. Since I published The Late Round Quarterback, a lot has changed, but my Twitter handle hasn't, and since it's still at late Round QB, I get crushed with questions all the time about quarterback strategy and fantasy football. But what about my league where quarterbacks get six points per touchdown pass and lose three points per inner. They'll ask, or what about my league where everyone hoards quarterbacks? If you have a question about quarterbacks and fantasy football, chances are I've heard it before. The one question that's popped up more and more over the last two years, though, relates to superflex and two quarterback leagues. People just don't know how to exactly approach the quarterback position in these formats, and it's easy to see why. In a superflex league you can flex a quarterback instead of the typical running back, wide receiver, or tight end, and since quarterbacks do score the most points in fantasy football, it makes sense to play a passer in that slot. They are, for all intents and purposes, two quarterback leagues, and when you're starting multiple quarterbacks on your fantasy football roster instead of just one, things change your approach shifts. Finally, after years of getting these super flex questions, I decided to dig in into a big super flex league study. If you follow me on Twitter or read content on number fire dot com, you may have already read it. It's been out there on the internet for about a week and a half for all of you to see, but I wanted to use the pop podcasting platform to talk through the study and add some additional context to it. Sometimes the written word just isn't enough, and maybe you read it and you didn't fully understand it. So let's go on a journey today. Everyone. Let's look at as many angles as possible to try and find the optimal way to draft in a superflex league. I think the easiest way to understand this shift that happens in draft philosophy when moving from a single quarterback league to a superflex one is by looking at quarterback value in the first place. Why is it that we can draft quarterbacks late in our traditional fantasy football leagues. Clearly I've done a lot of thinking on the subject, and I've broken the reason you should draft your quarterbacks late down into four main buckets. Supply and demand, opportunity, cost, predictability, and scoring variance. Let's start by looking at supply and demand. We've run through so many supplying demand examples on this podcast in the past, so I won't bore you with any more of them. But in general, in a competitive market, when the supply of something increases, the demand decreases, and the opposite of that is true too. If you've ever wondered why so many viable quarterbacks are drafted late, or why quarterback average draft positions are the way they are, this is the main reason. Supply and demand is the main reason. There are lots of quarterbacks to choose from and each team only needs one to fulfill their lineup requirements. Since twenty eleven, according to Myfantasy league dot COM's average draft position data, the average quarterback twelve has been selected at pick ninety in most drafts. That may seem higher than your typical one this year, but remember guys quarterbacks five and six years ago were drafted a lot earlier than they are today. Meanwhile, the RB twenty four has been selected around pick seventy, while the wide receiver twenty four in PPR formats has dropped off of boards at pick fifty eight. You may wonder why I'm comparing the QB twelve to these positions. It's because they're the worst hypothetical starter. In a twelve team league. At least two running backs and wide receivers are being started in most formats, meaning the demand is greater than what we're seeing at quarterback. Because that demand is greater. The last theoretical star in a twelve team league has a higher average cost. I mean this is no different than anything else that you see throughout your life. If everyone is looking to buy a new gadget, then the company who makes that gadget has the ability to raise the price in order to maximize revenue until people stop buying it due to the price. They're just playing the supply and demand game now. Because of supply and demand, and because you can get these starters later in your draft, the opportunity cost in grabbing a quarterback becomes larger. For those of you who are unfamiliar, opportunity cost is the loss of potential gain from other alternatives when one alternative is chosen. So when you make a decision whenever you didn't decide to do, that's essentially your opportunity cost. If you go to the movies and you get popcorn, your opportunity cost isn't the popcorn, it's anything else that you could have used that money on, like candy or something. When you select a player, like let's say that you pick leaveon Bell first, overall, you're getting Lavon Bell, but you're also foregoing the chance to get other players at his cost. Getting Leaveon Bell in a snake draft, it means that you're not getting Todd Gurley, Ezekiel Elliott, David Johnson, and so on. That's your opportunity cost. Those players are your opportunity cost. So imagine this. Imagine having the first overall pick in your fantasy football draft, and in this draft, pretend that the difference between the QB one to the QB twelve is the exact same as the RB one to the RB twenty four and the wide receiver one to the wide receiver twenty four. And not only that, but the point differential from one player to the next is exactly the same. Basic value over replacement level principles would tell us that it doesn't really matter who we select the advantage we're getting from one player or position to the next, it's basically the same. The problem in drafting like this is that market value isn't being factored in. These quarterback points are being stretched across more rounds than that running back and wide receiver, Meaning each time you decide to not get a quarterback, you're not losing as many points as if you were to forego drafting a running back or a wide receiver. If the difference between the QB one to the QB twelve is one hundred points and it's the same at run back in wide receiver, but you can get the QB twelve and round ten, while the RB twenty four and the wide receiver twenty four will drop off the board in round five, then why would you get the quarterback? You're just stretching these one hundred points across ten rounds versus five. Your cost advantage just isn't there, and when you get a quarterback early. This is a super simplistic way of looking at this. We know that things don't line up perfectly. We know our projections won't be as convenient as the example that I just walk through, but we also know that value over replacement player or VORP is also flawed. I think the concept is right and that we should be looking to see the advantage that we can get within positions rather than a cross positions in fantasy football. But as I talked through in episode twenty two of the Late Round podcast, there are flaws in looking at projected points versus a baseline player in order to make your decisions. For instance, that baseline player is subjective. I've been comparing to the worst hypothetical starter in twelve team leagues, but that's not a completely objective way to do this. VORP also relies heavily on projections, and we know that projections are going to be wrong. There's variance in fantasy football. That's especially true as you get Layton drafts and you're just looking for upside. You don't want a projection for a late round player. You want the upside for that late round player. And not only that, but fantasy points scored. Isn't the only thing that we should be worried about when we're analyzing players. Yes, that's what we're concerned with most. We're concerned with players who score the most fantasy points. But we've got to think about what goes into our decision making each week. We've got to think about the third late round quarterback concept, and that's predictability. You can think of predictability and fantasy football at both the season long level and at the weekly level. We'll start by looking at things at the season long level. If you talk to an early round quarterback Truther, one of the main reasons they get passers early is because they feel as though they bust at a lower rate than running backs and wide receivers do. And that goes for tight ends as well. But I'm not looking through the tight end position throughout today's podcast because that makes things more complicated. In the end, your choices in the early rounds are coming down to wide receivers and run The reason people want these quarterbacks early is because they feel safer. We know Aaron Rodgers is going to perform, We know Tom Brady is going to do well. But that's not really true when you look at the correlation between average draft position and points scored at each of the three positions that we're analyzing today, quarterbacks don't have a stronger correlation. We're not any better at predicting their success from a season long perspective. In the study, I plotted the average draft position since twenty eleven on one axis versus the number of points scored by players at those eightyp spots on the other axis. Top forty eight running backs and wide receivers showed stronger are squared values when doing this versus top twenty four quarterbacks. Average draft position to actual fantasy points scored actually correlates better among top forty eight running backs and wide receivers than it does for top twenty four quarterbacks, and actually top twenty four running backs and wide receivers versus top twelve quarterbacks do too. In other words, there's a stronger correlation between average draft position and fantasy points scored at running back and wide receiver across the first twenty four that are selected than there is at quarterback across the first twelve that are selected at that position. And then when looking at the top six quarterbacks in the top twelve running backs, there's no correlation. Top twelve wide receivers show a little bit of correlation, but there's not much. This may be kind of shocking, but when you look at the data, it's actually not that surprising. Just to give you a quick example, the top quarterbacks selected on Myfantasy league dot Com over the last seven years have actually averaged fewer points scored than QB two, QB three, QB four, and QB fives. Some of this has to do with injury, yes, but you can say the same for any of the other positions. From a season long perspective, we're not better at predicting quarterback performance than we are at running back and wide receiver, and that's especially true when you spread the data out and look at top sixty wide receivers and running backs, just because those later round picks rarely pan out, so the correlation becomes stronger. From a weekly perspective, though, predictability is often so overlooked, but it's so so important when games are about to begin during a random NFL week where met with a bunch of starter sit decisions for our fantasy football rosters. When we go to set our lineups, our consideration set is pretty small. We only consider starting guys on our roster, or maybe two or three players who are on the waiver wire. But last season, one hundred and ten different wide receivers finished with at least one top twenty four performance. To put this another way, one hundred and ten different wide outs were usable in at least one given week. Not many of them were predictably usable though. Look, guys, it's tough to know. When a random number five wide receiver on a team catches a sixty yard bomb for a touchdown, that makes him worthwhile in fantasy football, but it doesn't mean that we were actually using him as fantasy football owners. This is where predictability comes into play. Many of us know what standard deviation is. It's a measure used to quantify the amounts of variation in a set of data. A data set with a high standard deviation means the data points within that set are located further away from the set's average, and then the opposite is true as well. The problem with standard deviation is that it's hard to compare across very data sets. For instance, quarterbacks score the most points in fantasy football, so that means there's a larger chance for a wide range of scores and thus a higher standard deviation. When we relate standard deviation to a data sets mean, though, then we get the coefficient of variation or CV for short. It's just the standard deviation divided by the sets average. The higher the CV, the more variability in your data set, or to put this another way, the more unpredictable the data set becomes. In the Superflex study, I use CV to examine predictability within groups at quarterback, running back, and wide receiver, and this was all done with last year's data. The average coefficient of variation for the QB one to QB six group last year was thirty eight point seven percent. That means absolutely nothing to you. Right now. So this is the average CV of Russell Wilson, Cam Newton, Tom Brady, Alex Smith, Carson Wentz, and Kirk Cousins. It's looking at their standard deviations and weekly scoring and dividing that by their average. This thirty eight point seven percent number was smaller than the forty seven point one percent one that we saw from all the RB ones or RB one to RB twelve last year. And the same can be said for wide receiver ones and high end tight end ones. In other words, of the top players at each position, quarterbacks provided the least variability. When we move down positional groups, we see that running backs get wildly less predictable. Wide receivers do a little bit as well, but the quarterback position remains fairly predictable compared to other positions. Just to give you guys some context as to what the data says, quarterbacks ranked nineteenth to twenty fourth had less variability last season than the typical RB two. A player like Jacoby Brissett was more predictable than Jordan Howard. One of the key insights from the CV analysis is the dropoff in CV from one positional group to the next, a big reason high end running backs are so valuable in fantasy football is because not only are they providing high point totals each week relative to their position, but they're doing so at a very consistent rate. The dropoff in CV is pretty drastic at the position. At quarterback, you don't get that dropoff, And to be honest with you, this idea of lack of variability is actually kind of intuitive. Almost any quarterback starter is dropping back to pass thirty or more times in a given game. Running Backs are lucky to see fifteen or more touches, and wide receivers seeing double digit targets is not all that common. That means are sample size in a game for quarterbacks as largest they control their own destinies most and when that's the case, they simply become more predictable. That's why the numbers show what they do. Quarterbacks in fantasy football have a relatively nice floor, and we know going into a given week which quarterbacks are starting. We don't always know which RB three is going to get touches or which slot receiver is going to see the most targets. This, along with the number of quarterbacks available on your waiver wire each week, is why you can stream the position. But that brings me to the last bucket for late round quarterback drafting, and that's scoring variants. A player being predictable doesn't mean that he's valuable. A quarterback could predictably score six points per week, but that doesn't make him someone that you want on your fantasy squad. But these low end passers aren't scoring just six points per week, you guys. Combined with predictability, we should realistically be looking at the type of scoring edge players are giving us at each position. Like I said earlier, fantasy points are not everything, but fantasy points still matter more than anything else. If we see just a small number of quarterbacks performing far better than the rest of their position week in and week out, and year in and year out, then that would tell us that maybe having an elite quarterback is worth it. They give you a massive edge over your competition. So to look at this type of data simplistically, I analyzed value through the lens of what I call usable weeks or weeks where a quarterback finish as a top twelve passer and where a running back and receiver finish in the top twenty four at their respective positions, and for the record, I've excluded Week seventeen from all of this because it's worthless to the fantasy football season. It's hard to describe this data without some sort of visualization, but when doing this exercise, the average number of running backs with fifteen or more usable weeks in a given season over the last five years is point two per year, and that's because David Johnson is skewing that day. But we've generally seen one point six of them per season with fourteen or more usable weeks, two point four with thirteen or more, and then we don't hit the twenty four mark until the six or seven or more range. This is basically saying the elite running backs in your league are going to have roughly seven more usable weeks than a low end starter. That spread and scoring is not the same at wide receiver, but it's definitely not the same at quarterback. On average, just one quarterback is providing eleven or more usable weeks, while twelve are providing seven or more in a given season. Think about what that means for a second. The best quarterback starter in your league is only providing four more usable weeks than the low end starter is That difference is larger at both running back and wide receiver. And then when you look at what I call elite performances or top six ones at quarterback and top twelve ones at wide receiver and running back, the same trends exist. Like I said, it's a little hard to describe this data, so it might be helpful to go to the study on number five and look for yourself. The scoring variance numbers are pretty interesting, but in some what they show us that high end quarterbacks aren't giving us this crazy significant edge over lower end starters. If you want that edge, look at running backs and then wide receivers. So those four reasons are why you can draft your quarterback late. You've got supplying demand, opportunity, costs, predictability, and scoring variants. But what happens when you add a quarterback to the mix. Well, naturally, your supplying demand formula has changed. The supply of players is the exact same, but the demand has increased. Instead of twelve quarterbacks being started in a given league, we're now seeing closer to twenty four. My buddy sal who's on Twitter at two qb FFB, provided some two quarterback average draft position data for me, and I utilize that data to note the changes that we see from single quarterback to two quarterback leagues over the last five years. In a one quarterback league, the top quarterback since twenty thirteen has been drafted on average at pick seventeen point five. That jumps to four point seven. In a two quarterback league. The QB six goes from pick fifty nine to pick twenty nine, the QB twelve from ninety two to five eight. The QB twenty four goes from almost an afterthought at pick one to seventy in a single quarterback league to pick one oh seven in a superflex format. The market value for these passers has changed. If you're selling a product that everyone wants in a really small town, but you decide to take that same product to a large town where everyone still wants it, your demand is going to grow because more people want it. More people want quarterbacks because they need more quarterbacks, So quarterback pricing increases in superflex formats, and that means our opportunity cost equation shifts as well. Before the main reason from an opportunity cost standpoint that we could draft quarterbacks late was because the last hypothetical starter in leagues was drafted far later than what we see at running back and wide receiver, the cost and waiting wasn't as substantial at the quarterback position. If we run under the same assumptions as we did before, where the QB one to the QB twelve at the same point differential as the RB one to the RB twenty four, and the wide receiver one to the wide receiver twenty four, then all of a sudden, we're looking at adding the difference in value from the QB thirteen to the QB twenty four because now you're starting one more quarterback per team. You're getting a point advantage of x from QB one to QB twelve, but from QB thirteen to QB twenty four you're getting x plus y. Basically, if this new added point differential outweighs any difference in average draft position, then it could favor drafting quarterbacks early. But it doesn't, and it goes back to predictability. We've already talked through how there's no correlation between early round ADP and fantasy points scored at running back, wide receiver, and quarterback, or at least there's not that much correlation. Things are a little easier to compare in Super flex leagues, because you're essentially starting two players at each position, you're probably starting more at running back and wide receiver. But let's make things favorite quarterbacks as much as possible here. So remember we can look at predictability in two ways, through the lens of season long predictability and through the lens of weekly predictability. If you chart the top twenty four quarterbacks, running backs, and wide receivers by ADP versus their fantasy points scored in that season over the last seven years, you get pretty similar trend lines across all three positions. To put this another way, as you move from one positional slot to the next and EIGHTP, there's not that much of an edge to be had at one position to the next in terms of yearly scoring. For example, at quarterback, you generally will lose four point two points per positional slot, So the QB one will score four point two points on average across the season more than the QB two, and the QB two will score four point two more points than the QB three, and so on. At running back, that slope is four point one four, so slightly less than quarterbacks. A wide receiver, it's three point five to one Honestly, this sort of goes to show that we're pretty bad at predicting things. If you look only at those numbers, they seem to point to the fact that you can get your biggest edge at quarterback versus what you can get from a wide receiver or a running back, because the quarterback trend line slope is largest, but is point zero six more points per season that significant? So if you know that you're not gaining any real sort of value over replacement player advantage by selecting one position over another, when your next goal should be to wait on positions that are predictable. And the reason for that is because if the bad players are at least predictable, and if we know those bad players will sometimes have big weeks, which is what we saw when we were looking at scoring variants, then we can then predict those big weeks. Remember, guys, the QB nineteen to QB twenty four group last season had a coefficient variation of fifty seven point seven percent. Based on historical average draft position data, those quarterbacks will be drafted between picks seventy nine point four and one hundred and seven point three in superflex leagues. That's the beginning of the seventh round to the end of the ninth round. In the twelve team league, the only groupings that were clearly more predictable on a week by week basis than those low end QB twos last season were the RB one to RB twelve group and the wide receiver one to the wide receiver twelve group. And since these lower tiered quarterbacks are indeed still posting usable in elite weeks, you can find an edge in your superflex league with weekly predictability because those quarterbacks are still valuable in super flex leagues. I did the same exercise with usable and elite weeks for superflex leagues as I did with regular single quarterback ones, and while it did make the quarterback position slightly more valuable, the position itself still didn't stand out above running back and wide receiver without going deep within the data, and you can see that data within the study on number fire. It basically showed what we saw before that it's rare for a quarterback to have some significant edge over his competition, even when you change the baseline to twenty four quarterbacks instead of twelve. If I were to sum all of this information up as neatly as possible, it would be that top quarterbacks and superflex leagues aren't giving you any huge advantage over other positions. That's the first thing to remember, especially when you consider ADP Because lower tiered quarterback starters will be drafted later than lower tiered running back and wide receiver starters in a superflex format, that alters the opportunity cost equation. All the while, we know that there are going to be a lot of different players in fantasy football at all positions who provide usable and elite performances each week. We also know that quarterbacks are more predictable than other fantasy football positions, so that means when we're building our lineups, we're going to feel more confident that we know what type of output a player is going to give us at quarterback than at running back and wide receiver. Now, this doesn't mean that you should wait literally forever to get your quarterbacks in a super flex league. We understand supply and demand, so we should understand that, unlike running back and receiver, quarterbacks can dry up faster. Hypothetically, if everyone in your league drafted a quarterback with their first three picks, you'd have zero players to confidently play each week, because all starters would be rostered, there would be no more starting quarterbacks left. That's why the strategy for a super flex league can't be like it is for a single quarterback one. I don't think it should be like a single quarterback one. Instead, it just makes sense to try and continue to win at these more unpredictable positions. Early on, it makes sense to get the elite running backs that are giving you a significant edge. It makes sense to want Antonio Brown and DeAndre Hopkins and Julio Jones over top quarterbacks because those players are giving you both a scoring variance edge as well as a predictability one versus the rest of the players at their position. Maybe it gets to your pick in the fourth round and you can't stand the running backs and the wide receivers that are available. You see no difference between them and the ones that you can get in the sixth round. Sure, get your first quarterback, get a top ten guy. But when you do this and when you move through your draft, I just ask that you remember the concepts that we talked about today. Know that someone like Ryan Tannehill, while not very valuable within his position, is still a more predictable piece than some random wide receiver or running back that you get at his ADP. And maybe it's not Ryan Tannehill, maybe it's Andy Dalton, or maybe it's Philip Rivers. The point is we know these lower tiered quarterback starters will be starters. They'll throw the ball thirty or more times per game. They'll provide a floor for fantasy points, and in good matchups, they can provide a decent ceiling too. Because of the large sample size we get from these passers each week. They also provide more predictability. You're not going to have a flawless roster in any competitive fantasy football league. It's just not gonna happen. If you want to take your chances on finding late round wide receiver and running back gems while winning the quarterback battle in order to obtain that flawless roster by all means, go for it instead. I think the smartest thing to do is draft running backs and wide receivers early while platooning the quarterback position. Get three or maybe even four quarterbacks who you feel good about when it comes to having a starting job all season long, and then play two of them by matchup each week. Because what you're Essentially doing then is attempting to win at running back and receiver while maximizing the predictability aspect of the game of fantasy football. You're thinking beyond just fantasy points scored. You're thinking about how you, the fantasy owner, actually interact with your lineup, how you actually make decisions, and once you start thinking this way, you increase your chances of winning. That's it for today's show. Thanks to all of you for listening. If you've yet to subscribe to the Late Round podcast, make sure you are by searching for it on iTunes, Google Plays, SoundCloud, tune in, and Stitcher. And when you're on iTunes, do me a favor and leave a rating and review. Like I said I said before, if you give me a rating and review, and if I ever meet you, I RL beers are on me. I'm also on Twitter at Late Round QB. If you have a question for this week's mailbag show, feel free to hit me up on there or email me at jjat number five dot com. Thanks again, everyone, I'll talk to you later in the week.