We Simulated 12,000 IDP Best Ball Teams. This is What We Learned.

Want to give yourself the best chance to win? Don't draft THIS position early...

May 04, 2023

(This guest post is written by David Kelly, a Canadian fantasy football player and occasional contributor to IDP Guys.)

If we are being honest with ourselves, our favorite thing to do in fantasy football is drafting. That’s what makes best ball such a popular format. You draft a team (or 50 teams), and then you’re done. You don’t have to worry about any roster management during the season. With no major best ball providers supporting IDP best ball drafts, I’m happy that The IDP Show has taken a lead in the format by running over 25 best ball drafts this offseason.

Best ball has provided us with data sets of thousands of drafts we can analyze and see what draft strategies give us a better chance of winning. That kind of work is completely missing in IDP. Even worse, IDP leagues are so varied in scoring and lineup requirements, that we are often left answering questions about roster construction by saying “it depends on your league” and then falling back on strategies that made us comfortable or have been successful for us in the past.

With that said, I believe that what is being presented here is the first attempt to really understand the structure in IDP leagues, best ball or not. I’ve taken the draft data from the six best ball leagues organized by The IDP Show last year and used it to randomly generate 12,239 different teams, scored them with their total points for the season, and filtered those teams by various roster construction strategies to see what strategies were more successful than others. We’ve also calculated win rates for each player to try and see what the optimal draft ranges were for each position.

I used the setup of Best Ball League 6, which was a 12-team, 25-round draft, where each team was required to start 3 DL, 3 LBs, 3 DBs, and 1 flex. (There are some limitations to this setup that I discuss at the end of the article.)

All the data for this was provided by Adam, who I am tremendously grateful for. If I had to collect the draft results and weekly scores for 300 players, I would have just simply not bothered to attempt this.

Results

This is the points distribution of every team generated from ADP. It is a pretty normal distribution, however, the part we are interested in is the top 8.3% of scores, the teams that scored at least 2,876.3 points. That roughly corresponds to winning a 12-person league. As we apply various filters to teams, I’ll refer to the “win rate” as the percentage of those teams that scored higher than 2,876.3. Win rates higher than average (8.3%) are good and lower win rates are bad.

Drafting Defensive Backs Early

In the ADP from last year’s best ball drafts, 24 defensive backs were drafted in the first 10 rounds, so to keep pace with your draft room, you needed two at the end of round 10. Here’s a look at the win rates for teams by the number of defensive backs drafted in the first 10 rounds.

Staying with or behind the field at defensive back was good for win rates while going heavy at defensive back very quickly hurt teams. Those teams missed out on top defensive linemen and linebackers and were more likely to miss values at the position later in drafts, because they were using some of those picks to catch up at other positions.

It’s especially relevant here as drafters were already fading the position. Teams drafting zero, one, or two defensive backs in the first 10 rounds were all filling their bench spots before defensive back spots on their roster.

Drafting so few defensive backs is not typical of other IDP drafts, so it’s worth keeping in mind even if you find yourself in a non-best ball league where defensive backs are going early.

Not every defensive back heavy start is bad of course. Best ball league six, which these randomly generated teams are based on, was won by IDPGuys writer Matt Record, who drafted four defensive backs with his first 10 picks. He benefitted from later-round defensive line picks like Nchenna Nwosu (round 13), Brandon Graham (round 20), and Javon Hargrave (round 21).

Defensive Line or Linebacker?

The most common answers you will see to the question “Which IDP position do I draft first” is Linebacker and Defensive End. The common argument you will see for drafting linebackers early is their consistency. On the other side, defensive line advocates usually argue that fantasy success for the defensive line is more skill-based, while linebacker is volume-based, making it easier to plug and play new players with others net injured.

The lack of waivers in best ball should tilt the argument toward linebackers here. You can’t add Jack Sanborn in the middle of the season in these leagues. Here’s a look at the win rates of teams based on this split of defensive linemen and linebackers in the first 6 rounds. For context, by this point in the ADP, there are 37 defensive linemen, 26 linebackers and 9 defensive backs drafted. As we’ve already seen that drafting defensive backs early is not optimal, these filters were limited to teams that drafted zero or one defensive back

While acknowledging that some of these constructions were exceptionally rare, spending early picks on the defensive line has a clear edge. It is also worth noting that there were substantially more defensive linemen taken than linebackers during these rounds. This is definitely something to look for in other IDP drafts linebackers sometimes drafted more often, or even equally with defensive linemen, even if they aren’t best ball leagues. The win rates are slightly lower across the board for teams with one defensive back than none. That’s not a surprise given how poorly early drafting defensive back teams fared, however, it suggests that the one and two defensive back builds that were somewhat successful above were taking those players after round six.

Estimated Player Win Rates

This chart is the estimated win rate of every player and their ADP. As the teams are randomly generated, the number of teams each player gets picked to can vary, so the win rates here are calculated as the number of times a player appears in the top 8.3% of teams divided by the total number of times that player appears amongst all the teams. Individual win rates are mostly affected by a player’s score during the season and their ADP, with overachieving players at low ADP’s getting higher win rates than some higher-drafted players.

There are a couple of things worth noting in the chart. The first is clear ranges where one position is better than another. Defensive linemen are the best choice within the first 100 or so picks, before gradually tailing off toward the end of the draft. Linebackers are the 2^nd best choice early, and peak between picks 120 and 185, before cratering. That is likely the time when all projected starting linebackers are gone, with the remaining picks being projected backups, many of whom scored no points for large portions of the year. Defensive back is the worst position early, rising through the middle rounds before seeing the same peak as defensive linemen at the end of drafts.

This won’t be a surprise to anyone, but having a player bust from the early rounds is catastrophic to a team. The lowest individual win rates come from players who didn’t play much like Shaquille Leonard (1.7%) Joey Bosa (1.8%), and Chase Young (2.0%). Even T.J. Watt (5.2%) who played 10 games in 2022 had a very low win rate. Injuries are unavoidable, but there should be caution when spending a high pick on a player who’s coming into the season with injury concerns like Leonard and Young.

A couple of factors explain the differences between positions in the chart. Defensive Linemen score points with a higher variance as they typically have fewer tackles and more sacks and tackles for loss than the other position. Big 3 scoring enhances this variance, with one sack being worth about 9 solo tackles. Pretty much any player with a sack makes it into your lineup, and multi-sack games generate huge spike weeks that are mostly not attainable for the other positions.

The other factor is our accuracy in picking the right players. Drafters as a collective aren’t too bad at figuring out who the best defensive linemen and linebackers are. We are far worst at picking the best defensive backs. I’ve written about the IDP Invitational drafts over at IDPGuys, and noted that the top 5 Safeties in their tournament were Rayshawn Jenkins was the only top 5 scoring Safety to be drafted, and that happened in only 1 of the 12 tournament drafts. A similar thing happened in the best ball drafts, where of the top 10 scoring defensive backs, three players - Marcus Jones, Rayshawn Jenkins, and Desmond King – do not appear in the top 300 players by ADP. For comparison, Frankie Luvu is the only top 10 Linebacker to be outside the top 300, and none of the top 10 defensive linemen have an ADP higher than 212.

Add in that Big 3 scoring scores a pass defensed much higher than a tackle, raising cornerbacks and further flattening and very flat position scoring wise, higher drafted defensive backs are terrible picks.

A good example of this is Derwin James. Drafted with an ADP of 13, he finished at the DB9 on the season. His win rate however is 6.6%. By contrast, Bobby Wagner was the LB9 and had an ADP of 14, one spot after Derwin James. Wagner had a win rate is 9.4%. Defensive back is too deep, and to have success spending a high pick on a defensive back, they must hit a much narrower range of good outcomes than other positions.

Drafters in the 2023 best ball leagues are already catching on how poorly defensive backs did in 2022 and are fading the position even further, with some drafts only having 12 defensive backs taken in the first 10 rounds instead of 24. There is at least one draft where Derwin James was a 5th-round pick. At some point, pushing defensive backs down draft boards will make them better picks, but it’s hard to say at what point that will happen.

The top 10 win-rate players at each position repeat most of what was said above. Jalen Ramsey was the only defensive back in the top 10 rounds, as the DB24 by ADP. On the other side, just two of the top 10 win-rate defensive linemen were picked after the first 10 rounds, with several high picks making the list. Linebacker is in between the two, with mostly middle-round picks.

Conclusions

Every NFL season is different and will produce different results from the past season. However, the two main takeaways are:

1. Do not draft defensive backs heavily in the early rounds
2. Prioritize defensive line over linebackers in the early rounds.

Despite this analysis being specific to the IDP show best ball leagues and their scoring, I think these takeaways are both applicable to some extent in other leagues. League scoring, starting requirements as well as the direction of the draft room should be considered, staying flexible to any changes.

Given how bad early picks on defensive backs were in these drafts where the position was already being faded, the strategy to fade the position should work in other leagues where the position is being drafted normally.

For defensive linemen, best ball scoring should allow that position to be drafted later, as we don’t have to choose who is in our lineups. But in managed leagues you do have to choose, so fading the position in those leagues adds the challenge of getting your lineup right more weeks than not.

Thanks for checking out this article. If you want to read more of my work, please check out my author page at IDPGuys.org.

Subscribe to The IDP Show so you don’t miss an episode!

Limitations

As you might imagine, there were several challenges and limitations to this work that affect how it is applied to best ball drafts, or more broadly to other leagues. Some of these issues I tried to address, and others I couldn’t.

Each of last season’s best ball leagues was a little bit different. Some drafts were 10 teams while others had 12. Some leagues started 4 DL, 4LB, 4DB, and 4 flex, while others started 3 of each position and one flex. I chose the setup I did because I think if best ball is ever adopted by a major platform, it will likely look more like this than the drafts where we had more players in our weekly lineup, and also because I just like that format better. My thought for larger format drafts is that you can be more extreme in ‘good’ roster constructions because there are more lineup spots to fill and more rounds at the end to backfill with defensive backs. I’ve yet to be in a draft where I’ve felt like there are no more decent defensive backs to pick. That’s just not true for the other positions.
Randomly generated drafts are “dumb”. There is no logic to the sequence of picks or even counting the number of players at each position. I addressed this by only using drafts that had at least six players at each position.
Players drafted in the first two rounds had narrow ranges where they were drafted. T.J. Watt and Myles Garrett were always drafted by the 4^th pick, while Derwin James and Bobby Wagner had ADP’s in the early 2^nd round, but were never drafted after pick 18. While Watt & James or Garrett & Wagner teams would likely occur in a large number of drafts, they never occurred here. I adjusted for this by having about half the teams where any first-round player and any second-round player could be selected, and half the teams with players in the top half of the first round & the bottom half of the second round, or the other way around. This had some impacts on win rates for players in the range, as the top 6 players by ADP as a group out preformed the next three six-player groups. Nick Bosa ended up having a slightly higher win rate than Maxx Crosby, despite Crosby scoring more points. Similarly, Isaiah Simmons and Derwin James both had round 2 ADP’s and scored similarly, with Isaiah Simmons having a slightly above-average win rate (8.9%), while Derwin James was below average (6.6%).
On the opposite side, once we got past the fourth round, player ADP varies so much that it no longer made sense to confine players to one round or another. So rounds five and six were combined, with two players taken from the group, same with rounds seven and eight, and so on, with groupings getting larger until the last six rounds were grouped together.
Dual positions players were only counted as one position. I just could not find a way to allow dual position eligibility, so all the DL/LB and DB/LB players lose the scoring flexibility that makes them popular with some drafters. Players were assigned to the position that’s more beneficial, which is defensive line for DL/LB players and defensive back for LB/DB players. Positions were counted as of the start of the season, so Isaiah Simmons is considered a defensive back despite being classified only as a linebacker while all six drafts were occurring. Similarly, Micah Parson was a defensive lineman despite only having that eligibility for some of the drafts. Having dual-position players should help those players even more, as some of the best draft strategies involved drafting more defensive linemen than lineup spots for those players.
The players are scored with Big 3 scoring as that’s what was used for these leagues. This makes sense as the drafters were presumably drafting to that scoring setting. Even though I believe a lot of IDP scoring settings are more similar than different, it’s worth noting if you plan on taking any findings from here to other leagues.