Free Agent Predictions Retrospective | FanGraphs Baseball

Free company has come and gone this offseason. Sooner than we’re used to in recent times, we will look again on this 12 months’s class and make some conclusions. To some extent, that’s loads of our offseason protection; what are ZiPS projections and positional energy rankings, in spite of everything, if not catalogs of how groups have modified their fortunes within the offseason? Immediately and tomorrow, I assumed I’d do one thing barely extra navel-gazey, and maybe barely extra helpful in the long term, by trying again at my contract predictions to see what went proper and what went mistaken.

First issues first: let’s take an accounting of each the group’s and my predictions. I took the contracts signed by every of the highest 50 gamers in free company. A number of clarifying remarks: I eliminated gamers who accepted a qualifying supply, as each the group and I made our predictions earlier than qualifying affords had been prolonged. I thought of solely assured years, ignoring choices of any sort, be they vesting, group, or participant. I additionally ignored incentives and commerce kickers. Lastly. I’m utilizing Carlos Correa’s rumored take care of the Mets — 12 years, $315 million — even whereas it’s not but official and could also be amended.

With that out of the way in which, I grouped the gamers into positional teams and in contrast our predictions to actual life. How’d it go? Fairly effectively, really, for each side. Constructive numbers right here imply we under-estimated, and destructive numbers signify an over-estimate:

Good job throughout. Each units of predictions didn’t miss by a lot, notably in the case of common annual worth. These numbers overstate my accuracy, although, as a result of I used to be too excessive on some gamers and too low on others, particularly on the AAV that each the group and I did so effectively at predicting within the mixture. Right here’s absolutely the worth of misses for each units of predictions. In different phrases, lacking low by $2 million on one prediction and excessive by $2 million on one other would end in a mean absolute miss of $2 million, not $0:

I did meaningfully worse right here than I did within the mixture, and I’m going to clarify the precise particulars of why in a separate article that focuses on particular person predictions that I’d like to enhance on. For at the moment, although, we’ll concentrate on greater image particulars, so these numbers are principally only for illustration’s sake.

With that out of the way in which, I’d like to speak about one thing that I did effectively in my predictions: approximating common annual values for beginning pitchers. After I was developing with my estimates, I cut up the inhabitants of free brokers into three sorts: starters, relievers, and hitters. I then in contrast 2022 WAR and 2023 WAR estimates, together with some home-cooked getting older elements, to earlier years of free company.

I feel this works greatest for starters, as a result of their demand is rather more predictable than both high-leverage relievers or place gamers. Who wants a middle fielder? It’s onerous to say prematurely. Who wants a stud beginning pitcher? Actually everybody, yearly, may add Justin Verlander to their group and get significantly better. That broad-based demand implies that earlier market conduct may give a fairly good thought for present conduct.

One other strong class, for each the group and me: common annual values given out to hitters in mixture. I suppose you may simply say that all of us did effectively on common annual values usually, apart from the truth that we each missed the aid market considerably, however I struggled to determine particular hitter contracts regardless of getting the final market proper. Once more, that may should do with the much less predictable marketplace for place gamers.

For each of those classes, making use of a formulation labored pretty effectively. I took the full quantity of common annual worth dedicated in earlier years, in addition to the full quantity of year-one WAR out there in free company in every class, and used that to create an estimated open-market value of WAR. From there, I utilized an inflation time period. Mine was based mostly on a mix of three elements: the long-term drift in contract values, a post-CBA-year bump roughly based mostly on the previous few CBA negotiations, and general CPI inflation. My considering in making use of common inflation on prime of regular baseball drift was that whereas 2% inflation possible wouldn’t have a lot impact on baseball contracts, the runaway inflationary setting of the previous 12 months would bleed by means of considerably. I pressured the general numbers for every group to match this formulation, which is why I feel I did effectively at predicting broader markets; if I wished to allocate a little bit more money to 1 hitter, I subtracted some elsewhere and vice versa. That served me effectively when it got here to predicting the market talking broadly, and guarded me from having one or two missed particular person participant predictions torpedo the general end result.

I’m fairly happy with that general mannequin, however I’m aggravated to have missed the pattern in whole greenback outlay. The large pattern of this 12 months was groups providing gamers longer contracts than ever earlier than. That was the case on the prime of the market, certain, but it surely was additionally the case all the way in which down the road. The common contract size for the highest 50 free brokers was a half-year longer than each crowdsourced estimates and my predictions. That’s lots, provided that the typical contract size for the highest 50 was 3.5 years. That additionally made the AAV predictions look higher than they really had been; as a substitute of bumping up per-year payouts, groups prolonged contracts out additional into the longer term.

As I wrote about final month, I consider that greater rates of interest led to those longer contracts. Not foreseeing that was an enormous miss, so far as I’m involved; rates of interest have been excessive for fairly some time now, and I ought to have seen the impact that may have on future financial outlays. I’m going so as to add a time period for change in long-term rates of interest within the subsequent iteration of my predictions in an try to higher seize altering group conduct.

One other miss: I merely didn’t do an excellent job of predicting reliever contracts. I’m not notably distraught about this one; given that there have been solely three relievers within the prime 50, this principally boils right down to Kenley Jansen getting a king’s ransom in Boston. Relievers had been typically in demand this offseason, and if I may do the train over once more, I might have put Robert Suarez on the record, however I’m typically okay saying that I ought to account extra for latest type in my future reliever estimates and go away it at that.

Lastly, I really feel higher about my potential to foretell contracts on the prime finish of the market than on the backside. I used to be 20% low on ensures prolonged to the highest free brokers. That sounds unhealthy, till you think about the truth that the group was 36% low. The highest finish of the market blew the doorways off this 12 months, and I principally anticipated that. Then again, I did a a lot worse job of dealing with the underside half of the highest 50.

That is actually linked to the longer-than-expected contracts handed out all winter lengthy, however I’m going to spend extra time subsequent 12 months combing by means of mid-tier free brokers to get a deal with on that phase of the market. I’m not saying I’ll get it completely proper, however placing additional concentrate on that phase of free company can’t damage. My early plans for a way to do that contain breaking out predictions into extra tiers based mostly on anticipated contract dimension, however I’ll spend a while honing my methodology earlier than repeating this train subsequent 12 months.

Total, I’m pleased with my efficiency and impressed by that of the group. This 12 months was meaningfully completely different than the previous two years of free company. To the extent that we based mostly our estimates on previous years, we’d have been mistaken, and certainly, the pattern in the direction of longer contracts caught us each flat-footed. However general, in the event you’re on the lookout for contract estimates, this was an excellent 12 months to take a look at FanGraphs.