What if He's Not the Guy you Need?

A Case for more Discerning Player Takes in 2024 Best Ball Tournaments

I have a rather long introduction to this column so I want to give you a quick guide post for what this covers:

I. Introduction on the challenges to archetypal drafting in 2024
II. The Confounding Case of Jake Ferguson
III. Where Spike Weeks Come From
IV. Applying The Data to BBM5 Drafts

My most difficult internal conflict to reckon with as a person conducting analysis for probabilistic gaming is to determine the appropriate balance between humility with respect to outcome, and humility with respect to process.

The dominant tenet of all our work should be an understanding that the game we play depends on highly uncertain outcomes. The player who tilts the environment such that the winds of uncertainty thunder at their back is the player best positioned to emerge victorious. But how do we weaponize this respect for the great unknown without succumbing to nihilism?

For instance, as referenced in my last piece, my first ever article in this industry was an application of the Pareto Principle (aka the 80/20 rule) to fantasy football. The thesis was that a small number (roughly 20%) of a starting lineup produces a large portion (roughly 80%) of the value over replacement in a given fantasy team. It turned out this thesis was broadly accurate. However, the more actionable element of this was that of the small group of players who produced that substantial chunk of value, the majority were drafted early with an expectation of such production, and those who were not were often those we knew the least about.

Thus, the takeaway was: once you’ve reached a portion of the draft in which elite-level production is no longer expected, shift away from prioritizing the likelihood of a hit, and toward prioritizing the scale of a hit. In other words, don’t draft players with a substantial track record of middling production, chase the unknown once the known elites are gone.

This piece was my first example of the dilemma we often face in this industry: how do you balance a proclamation that on one hand is sweeping, dismissive, and confident and on the other hand is based on the premise of humility and uncertainty?

It’s the same paradox that analytics twitter accounts face when preaching about positional value or trading back: on one hand, that doctrine is premised on an acknowledgement of our inability to confidently evaluate players, but is often viewed by opponents as over-confident and dogmatic.

I am keenly aware of the risk that any heuristic, no matter how well-founded, may run the risk of being too broadly or rigidly applied, or outlive its usefulness if all of our opponents begin to follow it. However, I also fear that when we don’t draw any line in the sand, and continue to invite more and more plausible processes in a game like fantasy football, the inevitable endpoint becomes non-actionable and nihilistic.

This is especially true of best ball tournaments, where we have so few bullets and so many competing incentives.

The Illusion of Diversification

I’ve made this case before, but our draft exposures are largely artificial. In order to turn a profit in Best Ball Mania over a large portfolio of teams you need to make the finals. And in order for any entry to have positive expected value given how top-heavy the tournament is, you really need to be optimizing for first place.

The odds in a vacuum of any one team advancing to the BBM finals is 0.08% (539 / 672, 672). Or, less than 1 in 1,000. If you max-enter the tournament, your odds increase to 12%., or roughly 1 in 8.

Even if you are a particularly sharp drafter with a legitimate edge in this tournament, you would need to be better than 4X the average drafter in order for it to be likely you make the BBM finals this year. I’m not convinced anyone falls into that category.

Now if you enter several bullets into all the other similarly-structured Underdog Contests from Puppies to Poodles to Pomeranians to Dalmatians and so on, you may be favoured to have something to root for come Week 17. But at most we’re still talking about a handful of teams in most cases. It can feel re-assuring to click 2-4% of several different players just to “mix them in” in the event you’re wrong. And you probably will be on several of your fades! The issue is, the odds that player you took 3% of just in case will probably not show up on any of your actual finals teams!

I don’t think there is a clear cut answer on how much diversification is *optimal* and largely it just comes down to how you prefer to play, and your level of risk tolerance. But no matter how you choose to approach it, you should do it with a clear-eyed understanding that no matter what exposures you draft in the summer, by the time the money is handed out, your exposure to most players in the pool will be 0%.

That steers my approach toward taking more aggressive stands on my targets, in hopes that it (a) increases my odds of finding the right combination around players I contend will be the best picks of the season, and (b) gives me more opportunities to get those targets through the finals.

This approach of course is only helpful if I’m right about those said targets. And I very well may not be - it’s virtually certain I won’t be on at least some of them. But if we’re entering thousands of dollars into this tournament, I posit we should have some degree of confidence that we’re bringing a player selection edge to the table.

Competing Incentives

The difficult part of course is finding ways to balance these desired stands with the competing incentives we face in drafts: stacking, team structure, and cost. While picking the most impactful players is the single most important driver of your season’s success, drafting correlated teams, allocating positional value appropriately, and making selections at or after ADP are all levers that increase the expected value of your team. And unlike player selection, we can know immediately after the draft how well we accomplished those goals.

It’s simply not possible to have 50% exposure to a player if you only take them after ADP, only take them correlated, and only when they fit your structure. It’s easy to wind up so focused on these other edges you become a “sophisticated auto-drafter.” After all, anyone could be the “guy you need.”

Perhaps in Best Ball Mania I or II, when the concept of team stacks was lightly emphasized and Week 17 match up correlation was derided as galaxy-brain dorm room philosophy, simply emphasizing these micro-edges at the expense of player takes was a sufficient advantage over the field. By all means there are still completely dead teams in every lobby. But there are also several well-prepared drafters building 150 live teams who have all read the Best Ball Manifesto.

Unless you are comfortable playing educated plinko with your 150 entries just waiting for the one time every 8 years your ball drops into the right hole, you’ll need to have some degree of conviction on who to take, and you’ll need to be correct on enough of those stances with a high degree of impact to materially improve your odds.

But who do we take?

Back to my point about heuristics out-living their use… just two years ago I wrote in favour of “archetypal drafting” for Best Ball Mania 3.

The idea at the time was that the market was still being driven to a large extent by median projections, which just aren’t a particularly useful tool in this format of tournament where advancement to the playoff rounds is largely based on a small number of outsized hits, and playoff round results are driven by weekly ceiling.

If prices are being driven by medians, winning is being driven by ceiling, and you accept that access to a higher ceiling relative to a median largely comes with a lower floor, a broad heuristic to apply is that we should simply be focused on drafting the players with the widest ranges of outcomes: both from a weekly and season-long perspective.

As discussed last time, we have come a long way in how we price players in this game - especially when it comes to rookies.

In that article, I devoted an entire section to exploring why Garrett Wilson was being drafted behind Marquez Valdes-Scantling, and made the case that while ‘MVS’ had a strong archetype from a weekly perspective, Wilson’s potential ceiling from a season-long perspective should be valued more heavily. Betting on a player with that level of talent opens up the possibility he is a driver of your advancement season-long, while also providing outsized spike week probability from the mid-late rounds of your draft come playoff time.

Suffice to say, the market has become infinitely sharper since then, and i suspect we will never see another No. 11 overall pick priced behind a player with the talent level of MVS for a very long time.

However, while rookie prices have (correctly) gone up across the board since 2022, another humerous to look back on player who was commonly drafted ahead of Garrett Wilson that season was fellow rookie Jalen Tolbert. While the market was willing to steam up Tolbert based on the allure of the unknown, it may have got too carried away with prospecting nihilism to have priced such a vastly inferior prospect drafted two full rounds later above Wilson, based on opportunity and environment.

I think Tolbert’s rookie season was a forseeably comical rookie ADP, but we may look back on Keon Coleman and say similar things in a couple years.

You simply cannot look at the 5th - 7th rounds of drafts in 2024 and say that median projections are still driving ADP. The market has made the correct adjustment toward these wider range of outcomes players, which again makes our job more difficult.

If we want to pick up a sustained and substantial advantage over our opponents can’t only draft based on structure and correlation, and we can’t simply draft the wide bands… what indicators *should* we be looking for to determine where to take our stands?

It’s not who you take… it’s who you don’t

My inspiration from this column is largely due to a desire to articulate my response to a few pieces of feedback I’ve gotten on my recent rankings and explanations of those rankings on various podcasts.

I’d characterize that into a few categories:

1. Why do you hate our lord and saviour Jake Ferguson?

2. How can you be so low on ____ if you’re also low on ____ / high on _____ [insert player in same offense] ___?

3. Why are you so high on Jaylen Waddle?

4. Why are you so high on Xavier Legette who clearly, absolutely sucks?

[Allow me the tongue-in-cheek characterization… everyone who brought up these questions were quite respectful, I enjoyed engaging with you, and I think everyone asked legitimately good questions with valid points!]

Let’s kick this off with everyone’s favourite player in 2024: Jake Ferguson. (I will eventually come full circle around to the other 3 items of feedback… just let me cook)

Big Ferguson has truly taken over the industry. Every year there are players like this (or at least in my opinion comparable to this): a middling talent with an uninspiring track record who finds themselves in a favourable spot - sometimes after their own min-breakout in equally favourable circumstances, and sometimes simply due to the presumption of a potential breakout - and an ADP that clearly correlates more to their environment than to the player themselves.

Recent candidates that stick out in my mind include Allen Lazard, Marquez Valdes-Scantling, Gabriel Davis, Dalton Schultz, Logan Thomas, Mike Davis, Myles Gaskin, Alexander Mattison, Dameon Pierce, Miles Sanders, and - in fairness - some other players who shared similarities that did pan out such as Isiah Pacheco, Rachaad White, and assuredly more.

The Tight End position in particular has been known to produce one-season wonders out of nowhere, who are then drafted as mid-round TE1s the following season in a similar offensive environment, who often fade back into obscurity sooner rather than later. In fact, it’s happened so frequently I’ve named the scale by which we weigh such a TE’s likelihood of succumbing to this fate ‘The Barnidge Index.’

If you’re wondering why I’d place Mr. Ferguson highly on the Barnidge Index, I would encourage everyone to read this fantastic thread on the indicators of TE upside put together by David Gautieri, and spot how many times Ferguson is the outlier among the group.

The TLDR is that compared to TEs who have historically sustained as high-end options over time, Ferguson is relatively un-athletic, has a poor collegiate profile, struggles to earn volume on a per-route basis, and has poor pedigree.

One of the arguments I’ve seen promoted by Big Ferg is that he actually under-perfomed his usage last year and is due to regress positively. My pushback on that is that while efficiency stats such as fantasy points over/under expectation are often random, the premise for why we prefer volume stats to efficiency stats is that the best players are the ones who consistently sustain volume. Those players who flash with efficiency in small samples may not be able to hold up to the demands of a larger role, and see their efficiency vanish. With a player like Ferguson, I’d posit that taking the hollistic view of his profile, he’s simply a fish out of water getting the type of role he had last year (and projects to have this year) and thus under-performance from an efficiency perspective should probably be expected!

Among TEs with at least 200 routes, Ferguson was TE15 in YPRR (1.49), even including his post-season spike game. His TPRR was a middling .188. For comparison sake, here is every other TE going ahead of or around him in YPRR / TPRR:

• Sam LaPorta: 1.77 / .240

• Travis Kelce: 2.02 / .234

• Trey McBride: 2.03 / .258

• Dalton Kincaid: 1.51 / .196

• Mark Andrews: 1.93 / .214

• Kyle Pitts: 1.43 / .185

• George Kittle: 1.99 / .178

• Evan Engram: 1.56 / .227

• David Njoku: 1.75 / .235

• Brock Bowers: (College)

The only players who are comparable to Ferguson are Pitts and Kincaid. Kincaid is notably a rare early TE I’m not particularly interested in at cost, but it’s certainly easier to envision the ceiling materializing for a year-two breakout with Kincaid’s college profile and pedigree than Ferguson. Pitts is his own vortex, and honestly I think a fade stance is reasonable. But we *have* seen elite upside in Pitts’ per-route peripherals before, and there is a plausible injury excuse for his poor play in 2023.

Being asked to take Ferguson ahead of David Njoku who is a pedigreed, athletic TE with long term team commitment and a season of terrific play doesn’t compute to me. Neither does Ferguson ahead of transcendent rookie Brock Bowers. Maybe those environments fail those players, but I’m also buying the upside of environmental uncertainty and talent at their prices, while Ferguson is far closer to maxed out. His price assumes his environment will carry him along for the ride.

What’s surprised me about Jake Ferguson most, and why I’m devoting this much space in an article to a Round 8 TE as the proxy for my larger take here, is that with other players I’ve viewed similarly in the past, most people who view the game similarly to the way I do have been in lockstep making fun of this style of pick - including Cowboys TE Dalton Schultz as a 6/7 turn pick just two years ago. [Who scored 7.6 Half PPR points per game that year and I will not be gaslit to agree that is good]

With Ferguson, I’ve seen several folks I think the world of back him at cost, or at least defer to his single-game upside given the Cowboys environment, and his proven spike week ability based largely on the 2024 NFC Wild Card game.

If you’d like to see a spectrum of Ferguson takes debating this question by some great minds in the industry: check out minute 58 of this Ship Chasing episode (though as I’ve experienced with my subs and they make reference to here… Ferg Nation is large, loud, and proud).

And then… about 70% of the way through having drafted this article I got the opportunity to discuss Ferguson with the ADP chasers. My thoughts were not widely held.

I acknowledge fully that in this tournament, weekly ceiling *is* the most important driver of expected value… I question our ability to meaningfully separate that weekly ceiling out from season-long ceiling, such that we would prefer to bet on a weaker talent profile overall, because their environment is conducive to weekly spikes.

This led me down the path of looking up exactly who had the major spike weeks in Best Ball Mania history anyhow. I focused on Week 15-17 over the past four years and included any RB/WR who scored 25+ points and any TE who scored 20+ (since it’s a lower scoring position). While it reduces the sample, I chose to only include Weeks 15-17 (or 14-16 for BBM I) because I do think aspects of the playoff weeks are unique with respect to the distribution of points across a season from rookies vs. veterans, and the effect of injury attrition on the player pool at the end of a long season. We are solving for the playoffs here, and I think those weeks have enough unique characteristics that focusing specifically on what happened during those weeks is appropriate.