Best ball is probably making you worse at redraft

Fine, let's get into it

I’m 10 projections into my process, but yesterday was a slog, and I feel more like letting the fingers fly this morning, as evidenced by my RJ Harvey writeup earlier. Everything in moderation, so I’m going to not force myself to project this morning and instead get an answer out to something I’ve been meaning to.

In my post “I wonder if the paths for rookies are trickier” two weeks ago, I went off on a tangent about how I have a ton of conceptual posts I’m thinking about, and I threw in a note asking if any of them stood out as something you guys wanted to read. My truly deranged way of writing that was:

I’ve been slow to start my projections in part because I have 15 half-written pieces about how ball-knowers have a pattern-matching problem, and that concept of everyone just updating their own priors, and what we actually look for in “evidence-based” football research (and how so many think of that as the same type of evidence and reject others), and more about what I’m starting to call “the 60/40 rule” in my head where analysts get so confident in the 60% explained by their models they try to fill in the 40% “unexplained variance” I talked about in my last post, and how the people who will throw out intelligent analysis as biased are probably the most biased, actually, in thinking their numbers tell the whole story in a way observational analysis cannot, or shouldn’t be trusted to, and about where we’re at with AI and fantasy football(!!), and last but definitely not least how it’s probably Right, Actually to be biased, and we should all probably stop being slaves to ADP and correlation in the ways best ball probably does require on some level — when we’re not playing best ball, and most of my content has been geared away from that as I’ve realized how best ball is more different than redraft than almost literally anyone understands, and a related point I could write about how best ball has almost assuredly made the market meaningfully worse at redraft over the past half decade (which a lot of my closest friends in the industry are some of the biggest names in the best ball space, so don’t get it confused and think this is a commentary on extremely sharp people like Kerrane and Overzet who are good enough to know the difference, but is instead a broader point) — because if you’re actually doing the work, and you do understand collegiate profiles correctly, and you know what it is we’re actually looking for — which could maybe be defined as actual football players not theoretical ones — the big takeaway is the correct way to process everything is to be biased toward the profiles that matter. The ones who hit that take their time almost always have something in their background that was pretty elite, and the ones who people get too far out over their skis on are almost always in that other bucket.

Sorry for that run-on paragraph, but these half-written notes are all over the place, and it was kind of cathartic for me to get them all in one place. You can see where trying to tackle all those subjects might be difficult — that’s weeks of writing, and I have so much else to do before September — but also when I lay it out like that, I do kind of want to get into those topics… (Maybe let’s do this: If you saw a topic or a few topics from that last paragraph you really need my expanded thoughts on, leave a comment as sort of a vote, and if there’s interest and consensus I’ll try to write a few of those up in between my annual exercise of being overly statistically precise, i.e. projecting an NFL season I’m going to be very wrong about.)

In response to that, I didn’t get anything about the list of half-written concepts, but got multiple questions about best ball and redraft, which was this statement I made sort of as an aside as that run-on sentence really got out of control. And I’ve given that take before, and I’ve gotten the request to dig into it before, so let’s do it, not perfectly, but in my stream-of-consciousness, vibes-based writing style that I’m feeling this morning.

Michael wrote:

Just chiming in to say that a post elaborating on the differences between best ball and redraft and how you think best ball has (negatively?) affected seasonal drafters would be great.

And Luke responded to Michael to add:

This definitely. They are of course mechanically different from the perspective of waivers, and tournament play - but I’m intrigued as to what other elements of BB vs redraft make them so drastically different? That line really caught my eye as I was reading, not least because anecdotally my redraft results in recent years have been much worse and I’d like to know if I have any of these leaks

So I’ll try to address those specific thoughts.

I also want to start with Madison Parkhill’s comments, and if you don’t know Madison, he’s a very sharp fantasy football mind who writes a Substack newsletter called “Reducing Uncertainty” and contributes to Establish The Run. Madison started by quoting part of what I wrote, and I’m including that mostly because he agreed with me and it stroked my ego, but also a little bit because it’s relevant to where I’m going to jump into the discussion:

“Actually to be biased, and we should all probably stop being slaves to ADP and correlation in the ways best ball probably does require on some level — when we’re not playing best ball, and most of my content has been geared away from that as I’ve realized how best ball is more different than redraft than almost literally anyone understands, and a related point I could write about how best ball has almost assuredly made the market meaningfully worse at redraft over the past half decade.”

Couldn’t agree more with this.

As an example:

During the year where we had a bunch of “elite TEs” going in the 5/6 round, you drew the conclusion that they were largely a miss for redraft (and they were). For best ball, that rocky regular season mattered less, as they hit a few playoff spikes in “have to get first 10+ person groups” that demonstrated why they are still so important for best ball.

So I want to start here because this comment is almost too niche, but it explains one of the biggest points, which is that even though everything is weighted to the playoffs in redraft, the degree to which everything is weighted to playoff week ceilings in best ball is exponentially more significant, leading to the idea of “what if he’s the guy you need” and way more fixation on the theoretical. And I’m not saying that’s wrong in best ball, because there’s evidence of it, where random guys spiking in the playoffs is everything.

But because of it, I do think the top best ball players probably do have a sort of agnosticism to player profiles and willingness to buy anyone at any ADP that fits correlation or a given build, and again, I think this is probably correct. But I’d argue for a 12-team redraft league where you can navigate adds and cuts, that mentality is extremely damaging, because you really need to be building teams that fixate on a specific type of play that has the ability to work out or not, and brings something to a cohesive draft construction that is very different than in best ball. The objectives, and how your teams tell stories, are very different.

For redraft, one note is you prefer answers much earlier in a season, and when I say that I mean hitting a ceiling that is not defined by some unknowable thing in Week 15, 16, or 17 — which is relevant to best ball because nothing can change in Week 3, or Week 8, or Week 12 — but is defined by more knowable things in Week 3, or Week 8, or Week 12.

This also leads to specific player types being very different in both. In redraft, I’ve hammered volume-based RB plays in the RB Dead Zone since Underdog was a small company, or maybe my writing on that even outdates Underdog, I’m not sure. But one of the most important takeaways from Michael Leone’s multiyear manifesto work at Establish The Run was the importance of not taking zeroes in best ball, and how valuable floor is to a setting where you can’t change anything. You should be targeting discounted volume-based plays! This is also true of WRs who just run a lot of routes, because being on the field is going to lead to some touchdown-scoring games and “usable weeks” that aren’t even spike weeks, but in best ball, usable weeks are valuable. This doesn’t mean you can’t target rookies in best ball, and there’s considerable research into how important the inherent upside of rookies is, but there’s far more layering of a build that is probably necessary to actually get a best ball roster right, whereas for redraft you can legit just hammer rookies until the cows come home and then cut them and find boring routes-based plays on the waiver wire if you need a usable week when the bye weeks and injuries give you a roster crunch in Week 7. And because of that option of playing the floor-based play only when you need it, you get a roster-clogger element of any draft pick in redraft that doesn’t actually have upside, where people hold these guys on benches for “insurance” they don’t need because they can find the same replacement-level production the week it comes up as a need. It’s one of the biggest leaks people make in redraft, and the importance of not taking zeroes in best ball has only convinced people to pay more attention to that class of player. I see it a ton in my own leagues, in my discussions with people, etc.

So anyway, to wrap this part up, it’s not everything, but there’s a whole different pool of viable picks in both, with the best ball pool being much wider, and there are specific groups like traditional Dead Zone RBs and late-round WRs who lack real upside outside the WR Window (but will run enough routes and are good bets for a 750-yard, 5-TD season) that are actually really useful groups to be looking at in best ball — because of their discounted costs relative to their usefulness in best ball, and we know the ADPs wind up similar across these two disciplines (too similar, which is part of the problem) — but where in redraft they are the exact types of picks that shave off win percentage the second you select them. And that part of it really comes down to opportunity cost, where in best ball these boring players are foundational pieces for a build because they are cheaper than they should be and give you some floor of production at a lower opportunity cost than similar options going higher, but in redraft their cheapness doesn’t matter because the opportunity cost is you just gave up a chance to hit a real home run, and you only get 18 chances at it so you need to be swinging for the fences. Redraft is a Home Run Derby — you hit a nice line drive into the outfield and it’s not a double. That counts as an out here, because everything that’s not a home run can be cut, and you should be thinking about like 6 of your 18 draft picks being on your roster when you make the fantasy playoffs, and the other 12 being churned by then. Hopefully you have a better hit rate than that, but it’s that kind of mindset — that two-thirds of your picks will be trash anyway — that frankly makes you a better redraft player, and will also make you worse at best ball.

That said, there’s an element to the best ball mindset that does embrace this idea of things being trash. The central premise of best ball is you’re drafting a ton of teams and each team can be thrown in the garbage at really any point. There’s an understood worthlessness to each build that allows for some creativity of thought, frankly, just to try some things, because if a 1% seasonal outcome hits that creates some really weird weight on some positional thing we haven’t really seen before in fantasy football history, well it’s actually +EV to have dabbled in that weird thing a few times if you’re going to draft 1,000+ teams. But no one outside maybe Chad Schroeder is drafting 1,000 managed, redraft teams, and so the ways you’re trying to get exposure to different outcomes are different. You can’t be chasing every possible outlier result, and need more confidence in the stands, but you also do want some different configurations of the highest-upside plays if you do have a portfolio that for most of us is going to range from like two teams to a couple dozen, at most.

In best ball, to say, “Yeah, I should probably get exposure to that guy” is fine. In redraft, it’s probably pulling from your exposure to other plays that are your actual targets, and limiting your outs if you’re right on the actual targets, because you have a limited number of builds. It’s almost certainly wrong if you’re on the lower side of number of managed leagues you’re playing, where in that case you should be trying to jam in all the upside plays you can into a single build, not playing some balancing act born of understanding each build is insignificant that then overemphasizes our knowledge of the unknown elements of NFL seasons (which are real). In redraft, you need to have more of an ethos in your player targets and your draft structures, or else you’re just leaving things to chance and letting circumstances out of your control like what happens in your draft rooms dictate too much of your decisions to be consistently successful. Central to these points is the idea of ADP value, and how in best ball extreme ADP value is an opportunity to diversify, and there’s probably merit to drafting a huge chunk of your portfolio prioritizing great ADP values, and maybe layering in a few key stands where you thumb the scale a bit, whereas in redraft I’d argue you really need your target list and the willingness to reach when a room puts you in a bind.

All these things are interrelated, but another similar way I’d put this is in best ball, because of the larger portfolio nature of it, there’s the element that any one pick you make on a given team isn’t everything, because no single team means everything. And then because of the extreme ceiling needed in playoff weeks, it’s almost not about builds at all; last year, you kind of just wanted some Kyle Pitts in Week 15, somewhere. The year or two prior, it was George Pickens in Week 16, and before that Amari Cooper in that game in Houston with Joe Flacco. It doesn’t literally come down to just that, but it winds up having so much weight that you do sort of want to open your pods and find you have “a well-constructed Pitts team,” essentially. (Or maybe not, because what you probably want more than anything is to just not have Pitts teams in your pods that you’re playing against, so you can get non-Pitts teams through, because now Pitts is going to have such a high advance rate that not having him is pretty great, provided you get that lucky.)

Anyway, that “what if he’s the guy you need” YOLO logic has to exist in best ball, and it’s grown in that space over time, and rightfully for that space, but in a way that leads to an agnosticism that I’d argue is damaging on multiple fronts in a more concentrated landscape like redraft.