Fantasy Football Best Ball

Print

Optimizing Our Way to $1.5 Million: Looking at Best Ball Mania V Through an Optimizer

Coming to Grips With the Current Best Ball Market

150 Puppies. 15 Best Ball Manias. All slow drafts. All at once.  

That’s how I kicked off Best Ball Summer on Underdog.  

I’ve been playing best ball since DRAFT was a thing, but this marked my first foray into the 2024 market. What I thought would be one of the more +EV spots of the summer (maxing the Puppy after the schedule was released) quickly began to feel like lighting $750 on fire as I went through the five stages of grief with the inflated WR market. The resulting bargain-bin RB market only added to the confusion.

Bijan Robinson and Breece Hall in the back half of Round 1? A few years ago they probably would have had ADPs of second and third overall.  

Saquon Barkley, Derrick Henry, and Josh Jacobs all now on good offenses go in the second, third, and fourth rounds, respectively.  

Young talents in good situations such as Isiah Pacheco, James Cook, and Kenneth Walker go shortly thereafter.  

How to Handle WR-Crazy Draft Rooms

I use a lot of different draft strategies, and Zero RB is one of my favorites. But each round seemed to present an even greater RB value than the last as talented backs tumbled and questionable WRs got pushed further up draft boards. 

We’ve never seen a market like this before. The general discourse seems to be: Keep drafting WRs early and often or you’re doomed.

But everything we’ve learned about what works in best ball is still from small-sample size data in vastly different markets. 

I couldn’t help feeling like we are too confident that the current market is efficient. I started thinking: “What are 2025 best ball articles going to look like?”  

Optimizer to the Rescue!

I watched a 2022 interview with DFS legend "nerdytenor" on optimal best ball strategy. He mentioned that his strategy in tournament best ball would be to find optimal lineups and match exposures to these as best as possible.  

And from that, this idea was born. Take Draft Sharks’ industry-leading ceiling projections and run them through an optimizer to see which players appeared most often to figure out what’s optimal in this current market.

With a little help, I built that optimizer and was able to generate more than 6,000 lineups.

I ran every player with an ADP in BBM V through an optimizer 6000+ times again using @SmolaDS 's #1 ranked (ceiling) projections

Here are your top 25 in order of % appearance

Caveats and top QB, RB, WR, TE pic.twitter.com/bKPGTy8LUR

— Steve Myette (@in3usDFS) June 6, 2024

What You Should Know About the Optimizer

But before we get into the results, let’s talk about the limitations of the optimizer and this study to provide important context. 

What the optimizer does really well is exactly what an optimizer sets out to do in its most basic form: Create the optimal lineup based on the projections it’s fed. There are two important constraints:  

1. It selects one player per round based on their ADP
2. It adheres to common-sense roster construction (in this iteration: 2-3 QBs, 4-6 RBs, 7-9 WRs, and 2-3 TEs). 

However, the optimizer also has some limitations that impact the results: 

1. Ceiling projections typically do not account for contingent upside, a pillar of the Zero RB strategy. This will naturally skew optimal selections in later rounds towards WR when RBs might be better picks based on ceilings that projection systems don’t account for. 
2. The optimizer looks at cumulative scoring (so it isn’t optimizing for Week 17 or weekly scoring) and it doesn’t account for draft capital. This has both positives and negatives.  On the positive side, the optimizer is unaware of the current market “meta” and avoids potential roster-building pitfalls related to this. Alternatively, the optimizer does not know how many of each position we start in a week, so when the first 5 lineups the optimizer spit out wanted us to draft both Travis Kelce and Mark Andrews, I noticed the process was not perfect. While I love both guys this year in the depressed elite TE market, I’m pretty confident having them on the same team when we are almost always only starting one each week is not optimal. That’s why the study looks at individual players’ appearances in the outputs and not individual optimal teams. 
3. The optimizer does not force any stacking so as to not skew individual player values by forcing a suboptimal player into the pool to complete a stack. 
4. Setting ADP ranges by round and not by pick boosts some players with ADPs toward the front of rounds and hurts some players with ADPs toward the back of a round, especially in earlier rounds where ADP is more efficient. This by no means suggests these players should never be drafted. (My sincerest apologies to Jaylen Waddle and his 23.6 ADP at the time the optimizer was run for this article.) 

What the Optimizer Teaches Us

OK, now we’re ready for some results. For those just interested in the data or player takes, I’ve uploaded the output to a Google Sheet that I will update throughout the summer as ADP and projections change.

So what conclusions, if any, can we draw from what the optimizer says is optimal in the 2024 market?  

To quantify that at a draft-capital and roster-construction level over an individual-player level, I grouped the results into six buckets (Rounds 1-3, 4-6, etc.) and looked at the percentage the optimizer spit out by position. 

For example, the optimizer is suggesting that we generally select two RBs and one WR in the first three rounds. The suggestion for rounds 4-6 is one RB, one TE (Travis Kelce or Mark Andrews), and likely a WR over a QB. 

I’m still not convinced how much weight to put in these results. We know the market. We know we are using some of the best projections available. We know we are constructing rosters with positional allocations that make sense.  

We also know the drawbacks, mainly that this exercise is inherently biased against Zero RB.

So to see results skew toward RB early and WR late isn’t all that surprising.  

But are the results meaningful?

There is a reason the hyper-fragile RB strategy exists.

Underdog’s own Hayden Winks showed that RB-RB starts worked in BBMI and II. Pat Kerrane took down BBMIII with an RB-RB start. The data hasn’t been as promising recently, but we’ve also never seen a market like this.  

This next visual even further contextualizes the new gap we are seeing between RBs and WRs. 

We can see RBs are projected to outscore WRs at their ceiling at every stage of the draft until the Round 7 or 8 range. I even adjusted this for PPR scoring and RBs still came out ahead (although the gap was much narrower).  

Still, the projections in this market are a far cry from the actual points we’ve seen scored by these positions at past ADPs. Take a look at the graphic below from a 2021 article by Establish the Run’s Jack Miller, the most recent comparable data I could find.

WRs were previously outscoring RBs at every stage of the draft. If we trust the Draft Sharks projections (and we probably should), then that doesn’t look to be the case in 2024. 

Maybe the question we should be asking ourselves now isn’t “can I get good enough RBs later?”, but “can I get good enough WRs early?”

What Does This Mean for 2024 Best Ball Draft Strategy?

Having said all that, as the author of this article, I went into this with no agenda and I remain cautious.

These aren’t the answers to the test. They are only one piece of the puzzle. I’m still terrified of not leaning into this WR-heavy market.

We know it was inefficient previously, especially when it came to rookies. Now the question is just if we’ve overcorrected. I will still use plenty of Zero RB and Hero RB builds this summer.

I am also convinced that in a market that prioritizes WRs as much as this one does, there is still some validity to mixing in hyper-fragile RB builds – and, perhaps in the hungriest of WR rooms (gulps), robust RB builds. 

But what if we’ve created too much division on both extremes in the RB bro vs. WR bro holy war?

Recently, Pat Kerrane pointed out the drawbacks to hyper-fragile drafting while Pete Overzet did the same with Zero RB. The issues are the same. Extreme draft strategies on either end lock you into a position early, limiting flexibility and opportunities for value later in the draft.

Like with many things, the answer likely lies somewhere in the middle.

We all know the mantra of “letting the draft come to you.” The current market exemplifies that. I don’t think we should be scared of these WR prices and can’t afford to fall behind at the position. But we’ve also never seen quality RBs come at such a discount and shouldn’t be afraid to take them either.

So now I’ll also mix in a more balanced strategy of WR-heavy starts followed by targeting RBs in a dead zone that now feels very much alive. 

Maybe the title of this article in 2025 will be “Everything in Moderation."

Ready to Draft Your Best Ball Team?

Boost your chances of winning with the most powerful draft tool on the market. Watch the video below to learn more.