When the 2013 Crowd Gave Bad Advice
Of course it would take a sample size of so many decades of FanGraphs Contract Crowdsourcing before The Crowd could be properly evaluated for their salary-creating acumen. I will not hesitate, however, to deliver swift and mighty judgement upon The Crowd based on just their (your) inaugural year of estimations.
A few weeks ago, one Carson Cistulli shared an exhaustive chart showing how The Crowd’s aggregate contracts compared against the actual salaries that 2013’s free agents actually received. This is only one dimension with which The Crowd’s collective wisdom can be judged. Now that the free agents of 2013 have played the 2014 season, we can also begin to make strong guesses (and/or swift and mighty judgements) as to whether The Crowd or The Real-Live GM proposed a contract that more accurately reflects a given player’s competitive value.
Tomorrow, a look at The Crowd’s successes: five contracts that are perhaps even more discerning than the contracts proffered by The Real-Live GMs. But today, a look at The Crowd’s collective failures — five contracts that, were they executed by The Real-Live GMs, would be cause for those Real-Live GMs to question (and/or have already lost) their job security.
These five contracts have been selected not out of a set of rigorous objective criteria. However, their short length (all five players were signed by Real-Live GMs to one-year contracts) invites the arrival at conclusions. None of these five players had what one would call a “good” or even a “mediocre” season in 2014. While The Real-Live GMs who designed these contracts no doubt regret their decisions, in each case The Crowd proposed a much larger commitment, in terms of both years and dollars, for these players. In the financial tables below, one can see The Crowd’s aggregate proposals for the number of years (cYRS), average annual value (cAAV), and cumulative contract totals (cTOT) offered to these players, compared to the contracts that they would agree to with The Real-Live GMs (aYRS/aAAV/aTOT).
Jesse Crain – Houston Astros
The Financial Situation
| cYRS | cAAV | cTOT | aYRS | aAAV | aYRS |
|---|---|---|---|---|---|
| 2.1 | $6M | $12.2M | 1 | $3.3M | $3.3M |
In the first half of the 2013 season, Crain pitched a sterling 36.2 innings of relief for the Chicago White Sox, posting a 0.74 ERA and 1.52 FIP. Even though Crain was injured from the beginning of July and onward — with the Tampa Bay Rays giving up Sean Bierman and Ben Kline so that Crain could rehab for a few months in Florida — The Crowd saw it fit to give Crain one of baseball’s larger middle-reliever salaries.
The 2014 Performance Situation
| IP | K/9 | BB/9 | HR/9 | FIP | WAR | FG$ |
|---|---|---|---|---|---|---|
| 0 | 0 | 0 | 0 | 0 | 0 | N/A |
In 2014, Crain’s rehab tour moved on to Texas, where the Astros made Crain their third-highest-paid player (behind Dexter Fowler and Scott Feldman), only to see him never take the field. This winter, it’s unlikely that Crain will receive a contract any larger than $3.3M, with his last in-game appearance now approaching 18 months ago.
Stephen Drew – Boston Red Sox, New York Yankees
The Financial Situation
| cYRS | cAAV | cTOT | aYRS | aAAV | aYRS |
|---|---|---|---|---|---|
| 3 | $11M | $32.8M | 1 | $10.1M | $10.1M |
Following his performance as starter for the World Series Champion Red Sox, Drew was presented with a qualifying offer by Boston, and The Crowd believed that Drew would either be resigned by the Red Sox, or was a valuable enough option to one of the other 29 teams that he would be worth the lost first-round draft pick. After a winter of contractual stalemate, Drew re-signed in Boston in the middle of May.
The 2014 Performance Situation
| PA | AVG | OBP | SLG | wRC+ | WAR | FG$ |
|---|---|---|---|---|---|---|
| 300 | .162 | .237 | .299 | 44 | -1.1 | -$6.3M |
If Drew and his representation thought that it was difficult to sign a contract last winter, then this winter will pose a real conundrum. There is perhaps an argument to be made that the lack of conventional Spring Training, plus the midseason arrival in Boston, plus the midseason departure for New York all combined to throw off Drew’s internal ballplayer circadian rhythms. Or: Drew will not be entering negotiations from a place of strength.
Josh Johnson – San Diego Padres
The Financial Situation
| cYRS | cAAV | cTOT | aYRS | aAAV | aYRS |
|---|---|---|---|---|---|
| 1.6 | $10M | $16.3M | 1 | $8M | $8M |
With just 56 combined starts from 2011-2013, not to mention a history of long-term injury going back to 2007, 2008, The Crowd still favored giving Johnson a multi-year deal. Even though the Padres were confident enough in Johnson’s health to sign him (remember the Orioles passing on Grant Balfour), they still felt the need to get creative with the contract. Johnson would have received a not-insignificant bonus ($1.25M) if he made at least 26 starts in 2014, and a 2015 vesting option of $4M if Johnson were to make less than seven starts on the season.
The 2014 Performance Situation
| IP | K/9 | BB/9 | HR/9 | FIP | WAR | FG$ |
|---|---|---|---|---|---|---|
| 0 | 0 | 0 | 0 | 0 | 0 | N/A |
Much like Crain, Johnson was the third-highest-paid player on his team, with nary an inning to show for it. The Padres have already declined that $4M 2015 option, meaning that Johnson’s future in the game must be frightfully limited. It must be colossally frustrating for Johnson to see his elite performance get undercut with injury after injury — here’s hoping that he can call Brandon Webb for moral support.
Paul Maholm – Los Angeles Dodgers
The Financial Situation
| cYRS | cAAV | cTOT | aYRS | aAAV | aYRS |
|---|---|---|---|---|---|
| 2.1 | $7.1M | $14.8M | 1 | $1.5M | $1.5M |
As a totally serviceable starting pitcher for the Braves in 2013, The Crowd saw it fit to give Maholm a serviceable starting pitcher’s money. Seeing as the Braves voluntarily picked up Maholm’s $6.5M 2013 option at the end of the 2012 season, the numbers proposed by The Crowd did not, at the time, seem unreasonable.
The 2014 Performance Situation
| IP | K/9 | BB/9 | HR/9 | FIP | WAR | FG$ |
|---|---|---|---|---|---|---|
| 70.2 | 4.33 | 3.57 | 1.02 | 4.96 | -0.8 | -$4.1M |
For reasons that I’d think are very inside baseball (i.e.: confidential), nobody signed Maholm until February, at which point the Dodgers signed him to an uncharacteristically prudent contract. Even though Maholm’s 2014 season mostly just served as a foil for the early-May arrival of Clayton Kershaw, Maholm was the 23rd-highest-paid member of the Dodgers last year, so the lack of value in his contract was not terribly noticeable.
Kendrys Morales – Minnesota Twins, Seattle Mariners
The Financial Situation
| cYRS | cAAV | cTOT | aYRS | aAAV | aYRS |
|---|---|---|---|---|---|
| 2.5 | $10.8M | $27.3M | 1 | $7.5M | $7.5M |
Although the two players are quite unalike in appearance and style, Morales followed nearly the exact same trajectory as Drew. Morales was presented with a qualifying offer, turned it down, and was forced to wait until almost midseason to finally get signed. Also like Drew, Morales was traded from his original slumping team to a team that was trying darn hard to make the playoffs before pulling up just short.
The 2014 Performance Situation
| PA | AVG | OBP | SLG | wRC+ | WAR | FG$ |
|---|---|---|---|---|---|---|
| 401 | .218 | .274 | .338 | 72 | -1.7 | -$9.5M |
The similarities with Drew continue: Morales followed up that jumbled timeline with easily the worst performance of his career. Seeing as the Twins only paid Morales for six weeks or so before flipping him for a prospect in Stephen Pryor, they are perhaps the only party mentioned in this post who can be proud of their actions here. Meantime, the Mariners fell one game short of the playoffs after giving Morales and his below-replacement bat 239 late-season plate appearances — it is they who most obviously have egg on their faces.
Miles Wray contributes sports commentary to McSweeney's Internet Tendency, Ploughshares, The Classical and Hardwood Paroxysm. Follow him on Twitter @mileswray or email him here.
It’s probably more fair/more interesting to drop out (pre-season) injured players & players snagged in the QO mess. The crowd hardly has any info on such inside, off-field factors?
This. I don’t think these contracts tell us anything about the wisdom of the crowd. Except that we don’t have access to team doctors or medical reports, and that armchair GM’s don’t have to worry about losing a draft pick.
It’d be much more interesting to know where the crowd simply misjudged the talent/projections/performance.
I agree with regard to player health information, but not in regard to the qualifying offer. There’s no point in critiquing the crowd for not having information that teams had. However, the crowd is capable of judging how much losing a draft pick would impact their interest in signing a player and how much money they would give him. Projecting how the value of a draft pick will impact a free agent contract is certainly not as straightforward as projecting how a player’s performance will impact the same, but I feel it is in line with the purpose the exercise: instead of creating a complex contract projection system, just ask the crowd and have them try to work out the variables.
However, if we’re talking about comparing the projected contract to player value, then I think misjudging performance of Morales and Drew are not something that the crowd did wrong per se. The fact that both of them were so bad points to the possibility that taking that much time off could have an effect on the player we don’t really understand. Since the crowd had no reason to anticipate those player wouldn’t sign until June, nor that doing so would impact performance this way, I don’t think fair to just say the crowd was totally wrong.
Crowd contracts and projections of performance are essentially based on the assumption that the player is signed during the offseason and that they get a full spring training. In the case of Drew and Morales, the moment that they started to sit out Spring Training and the regular season, all bets were off. You can’t invalidate a necessary condition and then talk about the result.
In the case of Morales and Drew, we at least know that the amount they were offered prior to the start of spring training was less than they were willing to settle for. Would Morales have taken a $28M contract in the offseason, like the crowd projected? I think probably. Would Drew have taken $33M? Absolutely. So, if they would have taken the crowd-projected contract, but they in fact did not sign a contract, we can at least surmise that no team offered anything at or above what the crowd predicted. It would have been excessively difficult to predict what actually ended up happening, but I think one could reasonably say that the crowd underestimated the impact the QO would have on marginal free agents.
There’s the thing about players and agents misjudging the market, though. Remember how Ervin Santana was seeking something like 100 million?
For Santana, I have to think that was mostly talk. I mean he wasn’t really expecting $100 million. Was he?
the contributors are running out of things to discuss…
For Johnson yes, for Crain, no (at least on fairness). Crain’s injury situation was well-publicized, and that he was most likely unavailable until at least May or June was well-known last offseason. The un-wisdom of the crown in that particular instance was the 2nd year, and any real GM who had offered a guaranteed 2nd year would have had some ‘splaining to do
Pitchers with pre-existing injury issues, e.g. Johnson and Crain, are probably the type of players to end up with the least reliable crowdsourcing numbers since so much of their ultimate contracts will be determined by physical exams, to which the crowd has no access.
That’s it. Next time I put in my crowdsourcing numbers, I’m going to first go to each player’s bed in the middle of the night and run it through an MRI. Sure hope none of them have any metal bedframes…
Hey everyone –
Take a look at Kendrys’ stolen base numbers!
I don’t get it.
Idiots. This place is full of idiots.
I’m not an idiot because I know I’m an idiot, while you delude yourself into thinking you aren’t an idiot, therefore making you an idiot.
Nothing that has happened is idiocy; the fact that it happened proved it to be correct.
Don’t put me on the spot, man!
I’m sorry.
I must be the biggest idiot of all. I’ve never participated in the crowd sourcing or paid too much attention to it but I thought the point was to predict what contracts players would get, not what we estimated the players would actually be worth. Otherwise I assumed the crowd sourced numbers would almost always come in lower than the actuals, especially for the biggest deals.
Am I the only one that’s completely missed the point of this exercise?
No, you understood correctly. This article is a bizarre exercise for several reasons.
Yeah that’s my problem with this. I guess we suck because we thought that front offices are dumber than they are for these cases?
Exactly – the crowd sourcing should probably have a couple different values:
1. Anticipated contract (with and without qualifying offers)
2. Anticipated actual value
Yeah, I similarly thought this article was pretty strange. I mean, I would’ve never given Morales the crowd-sourced contract, but I fully expected Jack Z or someone to give it to him.
What we believe a player is actually worth (in terms of how we expect him to perform) and how we expect the market to value him are two very different questions.
How is Josh Johnson not a crazy super deluxe happy fun buy low bounce candidate unless his arm is actually off?
It is, actually. Don’t be fooled by Josh’s clever disguise. It’s him:
https://www.youtube.com/watch?v=2MiFHZ-VpcU
Maybe we should have a with-QO-attached and without-QO-attached contract guesses in the future. Because obviously (or at least very frequently) the ones that the crowed failed at the most are the ones in which an unexpected QO was extended and then declined.
If you take out the contracts that were hindered by a QO, I wonder how this would look. If we take out some of the lesser free agents with QOs, the crowd projections tend to be very good (Like Drew, Morales, Cruz, Kuroda, Napoli, Santana). I did exactly this to Carson’s table you linked to an, ta-da, the average is now nearly exactly right in terms of years and now about 4% too low rather than 11% on total value!
And regarding the distribution, it gets really hard to predict things when near a boundary, like the 1-2 year contracts. Essentially, the crowd is pretty good at figuring out which player will get a 1 or 2 year deal, but we don’t do a particularly good job at differentiating between 1 or 2 years (if you do a regression of crowd Years to actual Years for all crowd Years <= 2.2 the R^2 is just 0.02). Then, on the lower year deals, we’re worse (too low) at predicting the AAV of the deals than the higher end deals. While we seem to underestimate the total years and thus total dollar value on the higher end deals.
Worse, those short contracts tend to have bundled option years that make them very hard to quantify. I mean… what am I supposed to put when I think a guy will get a 1yr with a second year subject to a reasonable buyout? Many of those short ones are voodoo contracts that can only be described using a decision tree of some sort.
I hope this guy isn’t a teacher. “You failed to correctly answer a different question than you were asked.” One can almost picture him doing the superior dance.
I love that anonymous readers are actually offended by this article.
I thought it was interesting, Miles.
This applies to all comparisons of crowd sourced contracts to actual contracts given out by GMs:
Does it make sense to compare the average of the crowd sourced contracts to the actual contract (likely the highest valuation by one of the 30 GMs)?
I think it would be interesting to compare the actual contracts to the average (or median) max offer from several different randomly chosen 30-contract selections from the crowd sourced data. Or maybe use samples of 10 since that might be a more realistic estimate for the number of teams in the market for a given player.
If the valuation of a player had higher variability would this not favor the player and lead to higher contracts? I don’t think the crowd sourced contracts take this into account.