Revisiting Ottoneu Arbitration Strategies: Part I

I recently read The Midrange Theory, by Seth Partnow, which I would highly recommend if you are interested in basketball analytics, but that is not why I mention it. I mention it because his chapter about the NBA draft got me thinking about Ottoneu arbitration and drove me to do some research that turned into this article.

Partnow, citing psychologist Daniel Kahneman’s work, notes that it’s extremely difficult to become an expert-level intuitive decision-maker regarding the draft because the environment changes year-to-year, you get limited practice, and feedback is not immediate (nor is it particularly accurate or easy to interpret). The same issues apply to Ottoneu arbitration – every year, every league, and every team is different; we only do it once a year, and you don’t really get feedback on how effective it was for at least a few months, and even then it is hard to interpret.

As a result, some of the advice I have given in the past – avoiding rookies, pushing prices to the “keepable but just barely” level, etc. – is heavy on unvalidated intution. To try to adjust for this, I wanted to take a look back at some data and try to gather better information on how to effectively handle allocations in arbitration.

One of the challenges is that data is limited – there is no one place to see universe-wide arbitration from past years, let alone make judgment calls on what was and was not effective. So a caveat: doing this research, I relied on one league of data – League 1, the original Ottoneu league. I picked this one because a) I am in it, so I am well-positioned to evaluate what happened in edge cases, b) it is the longest-standing league and has had almost no turnover in years, so the managers in this league have more experience handling allocations than almost anyone else, and c) I needed to pick a league and #1 seemed as good a place as any to start. Extrapolating from this one league is not a perfect approach, but it is the best option I have for now.

After receiving the allocation data from 2021 (thanks to help from Niv Shah), I went about evaluating each player that was hit with at least one allocation and categorized each of the allocations as “fully effective” (i.e., the player was kept last year and will still be kept this year), “effective so far” (i.e, the player was kept last year and either cut this year or likely to be cut this off-season), “arguably effective” (i.e, the player was cut last year but either auctioned for more or was picked up by a team that placed the arbitration on him – more on this below), or “ineffective” (i.e., the player was cut last year and auctioned for less).

I used that as a starting point to answer questions and evaluate strategies.

This strategy ties into the “arguably effective” tier I mentioned above. I have always contended that if you assign arbitration allocations to a player and they are cut before the cut deadline, you wasted that arbitration. There are managers out there who disagree, arguing that pushing a player back into the auction can have value.

I understand this concept, in theory, but it requires a very specific set of circumstances to play out.

  1. You have to know or at least have a very good guess at what price a manager will cut a player. If a player costs $30 right now and you push them to $32 via arbitration, and they are then cut, you need to know if the cutting manager values the player at $31 exactly. If they value them at $30, you could have put $1 allocation instead of $2 with the same effect. If they value them at $29 or less, your arbitration allocation had no impact.
  2. You need the player to auction for more than they were cut. Using that same example, if the player auctions for $35, you have actually had a multiplying effect on your money – you allocated $2 and pulled $5 out of the auction pool – making all of your purchases slightly less expensive. Alternatively, if the player auctions for $25, your $2 allocation actually increased inflation, putting $7 back into the auction pool ($30 put in for the player being cut, $2 put in for the arbitration allocation that didn’t stick and $25 taken out when the player was bought back). You saved another manager money and made all of your purchases slightly more expensive.

I know some managers will say there is a second version of #2 – you could be the manager who buys back the player for less. In that example, if you get the player for $25 (and are happy with that price), that is a win. The problem is that makes the first item on the list less likely – if the other manager valued the player at $30 or $31, they would likely be willing to bid up to $30 or $31 at auction and you would not get the player for $25. If they were not willing to bid $26, they were likely unwilling to keep the player at $30, which means your allocations had no impact.

There is some wiggle room here – if in that scenario you win the player at auction for $30+ and the original manager was still in the bidding up to ~$30, I think you can claim victory. Honestly, even if the other manager bails from the bidding around $28, I think there is a case to be made that they would have kept the player at $30 (certainty, plus changing strategies over the off-season changing their valuation on the player) and that you have still “succeeded” by winning the player at $29+. But it is a narrow window.

In league one last year, there were four cases of players receiving an allocation and then being cut. Let’s take a look at those and see if any of them reflect a “win” for the allocating manager:

  • Buster Posey – Allocations: $1; Post-Arbitration Salary: $8 – We can be quick on this one. This was just a mistake. The manager likely put $1 on Posey early in arbitration and then forgot to edit it when Posey retired before the end of arbitration. Had Posey not retired, he would have been kept at $8.
  • Cavan Biggio – Allocations: $4; Post-Arbitration Salary: $14 – Biggio was cut at the deadline and then picked up at auction for $1. The manager who won Biggio was one of the managers who made an allocation to Biggio, but it is hard to call this a success. Given no one (including the original Biggio manager) bid even $2 on him at auction, I don’t think it is controversial to say this manager could have allocated that dollar to another player, Biggio still would have been cut, and he still could have picked him up for $1. Plus, he cut Biggio again before the end of April. Not great.
  • Randy Arozarena – Allocations: $2; Post-Arbitration Salary: $18 – This is an interesting one. Two managers put $1 each on Arozarena, who was then cut at $18 and picked up in the auction for $19 by the same manager who cut him at $18. This feels like a win for the teams that made the allocations. They basically turned their $2 of arbitration into $3 of impact. Not a bad outcome. And given the manager was willing to bid $19 to get him back, it is fair to assume that he would have kept him at $16 or $17 (we can’t know for sure, but it isn’t unreasonable).
  • Mookie Betts – Allocations: $2; Post-Arbitration Salary: $49 – Betts is another interesting one. He was cut at $49 and won at auction for $54 by one of the teams that placed an allocation on him. This meets the second criterion above – he auctioned for more – but of course, we have no idea if it met the first.

So we have two clearly ineffective cases, one likely effective one and one possibly effective one. But there is one other element to the “likely” effective case – it relied on the original manager to make a mistake. Cutting Arozarena at $18 was a mistake. It might be an acceptable gamble to take (“I like $18 Arozarena but maybe I can get him back cheaper”) but it didn’t pay off.

This is a pretty low success rate. Of the $300 in allocations in this league last year, $179 are still negatively impacting teams this year (i.e., those dollars are still on clear-keepers), while only $2 of the $9 placed on cut players are still having an impact, and that is only true because the Arozarena manager made a poor decision. This might be a little unfair to the managers that allocated to Biggio and Betts – maybe they are happy with the outcomes – but it gets to the root of the problem with this strategy: you never know if it worked.

From my point of view, I don’t want the success or failure of my actions to be dependent on another manager making a mistake. And we still don’t know if these managers would have cut at the lower, pre-arbitration prices. I still would not use or advise this strategy.

One other quick angle on this – among the other 65 players who received arbitration, half of the players allocated to are going to be kept again, meaning those arbitration dollars from 2021 will still have an impact in 2023. That is not the case for at least three of these four players – two were cut last year, Arozarena was cut in-season and picked up cheaper, and Betts is not a clear keeper for 2023 at his current price.

Don’t Allocate to Rookies
I have always advocated against allocating to prospects or rookies and prospects. But as I looked over the arbitration results, I realized that advice was too narrow. Rookies and prospects are really just a sub-category of a larger population that I would avoid allocating to: players who have not already shown their value. The shorter the track record, the more volatile the projections and the riskier it is to make an allocation. If you allocated to Julio Rodriguez at this time last year, you are probably still happy with that. If you allocated to C.J. Abrams, you probably don’t feel as good about it.

Last year, league one allocated to only three players with no MLB experience and three more with MLB experience but who needed a breakout to justify the allocations:

  • Julio Rodriguez – $1 pushing him from $7 to $8
  • Bobby Witt, Jr. – $4 pushing him from $8 to $12
  • Zac Veen – $1 pushing from from $2 to $3
  • Tarik Skubal – $1 pushing him from $8 to $9
  • Josiah Gray – $1 pushing him from $4 to $5
  • Alex Kirilloff – $1 pushing him from $6 to $7

All six of these players were kept through the auction last year, so that part is good. But it gets less pretty after that. Rodriguez and Witt are the big successes here. Both were excellent in 2022, though Witt’s 5×5 performance doesn’t fully translate to Ottoneu. League 1 is 4×4 and neither his SLG nor his OBP helped this year.

The others though are less positive. Gray was cut early in the season. Kirilloff contributed nothing and I don’t know that he is a keeper at $10 going into 2023. Skubal looked much improved but now is hurt and the timeline for his return still isn’t clear. Veen only made it to Double-A (where he was really, REALLY bad) and $4 is a high price to pay for a guy who won’t have a fantasy impact until at least 2024 and whose stock is trending down.

As mentioned above, overall about half of the players who received an allocation are going to be kept again into 2023, extending the impact of those dollars. Among the breakout candidates, it looks likely that two will be kept, while the others may all be cut. Skubal is the big wildcard for me. He will be $11 and while I still like him a lot, he carries a lot of risks, as well.

If Skubal is kept, that gives this group the same success rate (in terms of dollars being kept into 2023) as the overall group. That would suggest that league 1 was fairly appropriately averse to assigning dollars to this class of players.

The issue for me remains that those dollars assigned even to Witt and Rodriguez were high risk and low reward. Rodriguez is likely to get pushed to $35+ this year. Witt, I am less sure – I suspect he gets pushed quite a bit, but I also don’t think he was worth more than maybe $15 in 2022. In both cases, not allocating to those players last year wouldn’t have stopped us from pushing them way up this year. If you take the $9 allocated to these players and put them on other players, they would have been equally if not more effective, and we could still push Witt to the upper teens (if desired) and Rodriguez past $30 (as we should) this year.

Based on this, I’ll update my advice to be “don’t allocate to players who haven’t already demonstrated that they can earn their new salary.” That isn’t as pithy as “don’t allocate to prospects” or “don’t allocate to rookies” but I think it is more accurate. Make a guy prove it before you assume he is worth targeting in arbitration.

A long-time fantasy baseball veteran and one of the creators of ottoneu, Chad Young's writes for RotoGraphs and PitcherList, and can be heard on the ottobot podcast. You can follow him on Twitter @chadyoung.

Comments are closed.