See also part 2: Sleeper opportunity costs
Every year, when fantasy managers go to the draft they really worry about 2 things. First, not screwing up their first 2 picks. While this is certainly important, it is largely out of a savvy manager's control as they should basically take the best player available and hope that they don't get injured. The other is sleepers, players who are low draft picks but who have a chance to become top options. Today I want to talk about sleepers and Busts. How common are they? How much do they impact your chance of winning?
The Data
For this post, I will be combining 2 data sources. The first is the ADP and final ranking data from Asteroid Databases. This dataset includes players over the last 5 years and will be providing information on who were the sleepers and busts and what they look like. The second will be the draft results from ESPN leagues in 2020 (this will be the same dataset I used in my 2020 snake draft recap post and comes with the same large caveat that it is only 1 season of data). We will use this to begin to understand what impact a sleeper or a bust may have on a team's outcome. In both cases, I will only be looking at QBs, RBs, WRs, and TEs as it is rare that a K or DEF makes enough of a difference to matter (though not impossible, looking at you Patriots DEF of 2019).
Sleepers
Let's start with the positive option and explore the impact of sleepers. The first challenge is defining a sleeper. For our purposes, I will define a sleeper as a player drafted as a non-starter (QB/TE 13+ or RB/WR 24+) but finished as a top-half starter (QB/TE Top 6, RB/WR Top 12). These are the surprise studs, players that were drafted to fill in who became weekly must starts. (Note, I played with a lower threshold of a sleeper as a player drafted outside of Top 36 but finished in Top 24, but it did not change the overall statistics much. This is actually the first interesting finding of this analysis, most true sleepers really jump to stud status, few end up as just ok starters over the entire season).
Given this how many sleepers are available each year?
Looking at the table above, we can see that there are about 10 sleepers available every year. The other interesting aspect is that there is a pretty even split between QBs, RBs, and WRs in terms of the number of sleepers. Additionally, about 2 of those sleepers every year actually go undrafted, so remember to stay active on the waiver wire.
So who were the sleepers of 2020?
Last year was a slightly down year for sleepers with only 9 players qualifying. on top of that, we had a whopping 4 of them go undrafted! Interestingly, they tend to be young (under 25), with a new team (Stefon Diggs and Logan Thomas, or in a new situation (Mike Davis due to the McCaffrey injury, Robert Tonyan as a new starter). That being said, they do all feel like sleepers which helps give me confidence in my definition.
How much does drafting a sleeper help you?
We can take our 9 sleepers and look back at how teams performed if they drafted this player (for a complete explanation of the methods here, see my previous posts on 2020 Championship Players and my post on 2020 Snake Draft Strategy). Given that there is a mix of 8, 10, and 12 team leagues included in the dataset, the base level odds of winning the championship (i.e. random chance) was a little over 10%. The results here are pretty surprising! 3 of our sleepers significantly helped (with Stefon Diggs being the second most valuable player overall in terms of championship odds), but T.J. Hockinson did not help at all and Kareem Hunt actually lowered your championship odds! Overall drafting a sleeper gave you a 16.6% chance of winning the championship, about a 66% increase in your odds!
Billing Department
Busts
As with sleepers, let's start by defining our busts. For our bust definition, we can say any player drafted in the top 2 rounds (ADP less than 24) who finished the season as a non-starter (QB/TE 13+ or RB/WR 24+).
How many players bust each year?
So a couple of interesting points. First, not surprising there are more RB or WR busts simply because there are more of them drafted in the first 2 rounds compared to QBs or TEs. Surprisingly, QBs have a 40% bust rate when taken in the top 2 rounds. Just another point that taking an early QB isn't worth it (in single QB leagues). Most interestingly, WRs actually had a higher bust rate than RBs over the last 5 years. This flies in the face of the zero RB strategy which proposes WRs to be safer bets. In total, we should expect about 7-8 players out of the top 24 to bust each year or about 1/3 of players.
I'll be honest that is a much higher number than I expected, so lets look at why they bust. Specifically, I am curious how many busts were due to injury/missing games (something a fantasy owner can't predict) vs just poor play (which may be predictable on some level).
Bust Cause
Injury is the obvious cause so let's look at the number of games played for busts played vs non-busts.
Box plot of the number of games played by players drafted in the top 2 rounds split by whether or not they busted.
And we can immediately see that missing games is an obvious explanation for why a player busts. On average a bust misses 6 games in the season, whereas almost all non-bust players miss less than 2. Unfortunately, we can't control injuries, so lets take a closer look at the players who missed no more than 2 games, but still busted.
Non-Injury Busts
A couple of points come out of this list. First, some good news, even though these are all classified as busts, 6 of them (C.J. Anderson, Eddie Lacy, Randall Cobb, Odell Beckham Jr, Allen Robinson, and Deandre Hopkins) were likely playable, if as a flex at least. Obviously, this is not what you want (these are busts after all), but it's not like these players should have been dropped. Second, we can see that no QB or TE that was drafted in the top 2 rounds busted for any reason besides an injury. Not sure if that should give you confidence or not to take one (especially given QB's 40% bust rate), but it at least reduces the number of risks you have to worry about.
Interestingly, the main reason for these busts is poor QB play. C.J. Anderson in 2015 had an old injured Peyton Manning and Brock Osweiler at QB (while also splitting time with Ronnie Hillman), Allen Robinson in 2016 had Blake Bortles, Deandre Hopkins in 2016 had Brock Osweiler (2nd bust caused!), Amari Cooper in 2017 had Derek Carr (up and down at best), Brandon Marshall in 2016 had Ryan Fitzpatrick (same as Derek Carr), and Jordy Nelson in 2017 was missing Aaron Rodgers (injured). We can include Odell Becham Jr in his first year in Cleveland with a still too young Baker Mayfield in that list too. Jay Ajayi found himself in a 3-headed rushing game with Kenyan Drake and Damien Williams with Jay Cutler at QB and Adam Gase as Head Coach (A clear disaster looking back). That only leaves Eddie Lacy and Randall Cobb in 2015 whose bust cannot be explained by poor QB play, but both suffered from a Green Bay team that massively underperformed. The takeaway, try to avoid players with bad QBs in the top 2 picks as it is only adding to the number of ways the pick can go wrong. Players that may concern me this year are Alvin Kamara (we'll see what the Winston/Hill combo looks like), Jonathan Taylor (Carson Wentz may be renewed with Frank Reich, but maybe not), and Antonio Gibson (there's Ryan Fitzpatrick again).
What impact did drafting a bust have on your season in 2020?
We had 8 players bust in 2020, with all of them being due to injury of some form. Using the same methods as above, we can calculate their playoff and championship odds. Interestingly, we see that while Austin Ekeler and Chriss Godwin did not necessarily hurt or help your championship odds. Other players were much worse (looking at you Saquon). Overall, drafting a bust gave you a 7.35% chance of the championship, dropping your odds by about 3%. It's obviously possible to recover from this, but you are cutting your odds of a championship by a third!
Limitations
An obvious limitation to this is my definition of Sleepers and Busts, there are certainly other ways to define these, and that will have an impact. Significantly, I am only looking at end-of-year position ranks which ignores players like Jalen Hurts last year who came from nowhere to be a major contributor on championship rosters, but only played a few weeks and therefore was not considered a sleeper. On top of that, all of the championship odds analysis is based on 1 season only, and a covid season at that and should therefore be taken with a major grain of salt. Finally, none of this analysis considers opportunity cost, just the upside of hitting on a sleeper and the downside of hitting on a bust. (Note I have added a post on opportunity cost which you can read here) Despite these, I do think it is valuable to have a sense of how many sleepers/busts we can expect and what impact those players have on our fantasy season.
Conclusion
A couple of key takeaways to this analysis. First, we should expect about 10 sleepers and 7-8 busts in any given year. Second, Busts are often the result of injury or poor QB play, so try to avoid spending a top 2 pick on a player with a bad QB. Finally, remember that successfully drafting a sleeper increases your championship odds by 60% while drafting a bust cuts those odds by a third.
Finally, just a plug for previous work, but if you wanted to check the playoff and championship odds of a player I did not mention, you can see most players in my 2020 Championship Players post.
Questions? Comments? Let me know at ac@alexcates.com. Want to read more breakdowns like this? sign up for my newsletter. Finally, like what I do? Consider supporting me on buy me a coffee.