Let’s Talk About That Eric Thames Projection
Back on November 29th, the Brewers decided to non-tender first baseman Chris Carter, despite the fact that he hit 41 home runs for them last year. Despite his big time power, they wanted to open the position for Eric Thames, a free agent they signed the same day. Thames has been a star over in Korea the last few years, and the Brewers decided to bet on his uncertainty, hoping that some of his success over there is based on real improvements, and not just evidence of what a minor league slugger can do to inferior competition.
Normally, when teams make moves, we like to cite the projection system data as a baseline, to give us a rough understanding of what a player might reasonably be expected to do going forward. Because Thames was playing in Korea, though, we didn’t have a Steamer projection for him at the time, and in fact, we didn’t have one for him until yesterday, when Jared Cross finished running his forecasts for international players. And so, starting today, we’re officially displaying a projection for Thames here on the site. And it’s a pretty good one.
| PA | BB% | K% | ISO | BABIP | BA | OBP | SLG | wRC+ | WAR |
|---|---|---|---|---|---|---|---|---|---|
| 575 | 10% | 24% | 0.243 | 0.313 | 0.272 | 0.350 | 0.515 | 124 | 2.5 |
The last time Thames played in MLB, he ran an 82 wRC+, as he struck out six times for every walk he drew, and didn’t have enough power to compensate for his lack of strike zone control. This forecast suggests his walk rate might be double what it was back in 2012, while his strikeout rate is only a little above league average, and Steamer thinks the power displayed in Korea is definitely for real. For reference, here are the top 10 ISO projections that Steamer is forecasting in 2017, for players currently listed on one of our depth charts.
| Name | AVG | SLG | ISO |
|---|---|---|---|
| Giancarlo Stanton | 0.269 | 0.560 | 0.291 |
| Mike Trout | 0.303 | 0.568 | 0.265 |
| Nolan Arenado | 0.290 | 0.541 | 0.251 |
| Bryce Harper | 0.286 | 0.533 | 0.247 |
| Anthony Rizzo | 0.279 | 0.524 | 0.245 |
| Chris Davis | 0.231 | 0.476 | 0.245 |
| Eric Thames | 0.272 | 0.515 | 0.243 |
| Josh Donaldson | 0.275 | 0.514 | 0.239 |
| Kris Bryant | 0.275 | 0.513 | 0.238 |
| Manny Machado | 0.294 | 0.530 | 0.236 |
That’s some pretty great company to be keeping, as it’s a bunch of All-Stars and MVP Candidates, plus Thames and Chris Davis. Guys who can hit with this kind of power are valuable, and often highly compensated for their skills. Thames, though, will make less over the course of his three year deal with the Brewers ($15 million) than Davis will make in 2017 alone.
That’s because the market was still pretty skeptical of whether his skills will translate back to the big leagues, and clearly, there weren’t teams with internal projections this high, or else the bidding war would have been a bit more intense. So I know that it’s natural, when you see a guy sign for $5 million a year, and then Steamer comes out and says he might be better than Edwin Encarnacion, to not take the projection all that seriously. It’s easy to be suspicious of the calculation, and think that maybe Steamer just has something broken in the adjustment from the KBO to MLB.
Except, if we look around, it’s actually pretty similar to the other forecasts for Thames out there. Back when he signed, Dan Szymborski tweeted out his ZIPS projection for Thames.
ZiPS for Eric Thames in Milwaukee. pic.twitter.com/CUJhF3TJTr
— Dan Szymborski (@DSzymborski) November 29, 2016
The numbers aren’t quite as optimistic, pegging him for a 116 OPS+, which is more like a 114 wRC+ (since OPS overrates the SLG component, and Thames’ OPS is slanted towards SLG), about 10 points lower than what Steamer has for Thames now. But ZIPS is also buying the power to the same degree that Steamer is, projecting a .246 ISO, almost exactly the same as what Steamer has. The difference between them is that ZIPS sees fewer walks and more strikeouts, so the question is more about how much of his plate discipline from the KBO might carry over, not whether Thames has learned to drive the ball better than he used to.
Okay, that’s two projections, one pretty good and one pretty great. But that’s still just an n of two, so let’s look around for some other forecasts. Clay Davenport has been doing projections forever, and was one of the first public analysts to put together translations for players from foreign leagues. Here’s his page for Thames, which includes translated batting lines for his years in Korea, and a 2017 projection: .285/.358/.533.
That’s even more optimistic than the Steamer forecast, and similar to that projection, puts Thames’ future projection in a similar range as Edwin Encarnacion’s numbers. In other words, basically every available projection system that attempts to translate Thames’ numbers from Korea to MLB see him as an elite slugger, and thinks he’ll control the strike zone well enough to be one of the best hitters in the league.
Now, of course, projections aren’t gospel. And we know that pretty much every team in baseball has their own projection systems at this point, and if even a few teams had similar forecasts and put a lot of stock into those numbers, Thames would have gotten more than 3/$15M. So the fact that the market just valued him like a part-time bench player is also a data point we should consider, and is perhaps a good indicator that there might be reasons to think the public forecasts might be a bit too optimistic.
But while the Steamer forecast looks staggering, it’s worth remembering that it’s not really an outlier among public projections. And Jung Ho Kang’s success after getting ignored by the market is a nice reminder that MLB teams don’t have this figured out perfectly either. Everyone’s doing the best they can with imperfection information.
But if Thames comes anywhere close to these projections, the Brewers very well may have gotten the steal of the winter. And if he actually lives up to these forecasts? Then the rebuild in Milwaukee might not take as long as we think.
Dave is the Managing Editor of FanGraphs.
If Thames does start producing, will Milwaukee still make any significant progress? They’re projected to be one of, if not, the worst team in baseball in 2017.
Biggest obstacle for the Brewers improving in 2017 will be replacing Jeffress, Thornburg & the Fresh Prince in the bullpen.
Rotation should be pretty solid with Guerra & Davies turning in solid 2016’s, Nelson & Peralta still offering some promise of being mid rotation starters, Garza & Anderson for veteran depth at the back end and Hader/Lopez/Ortiz/Woodruff waiting in the wings for a shot.
Lineup should hold its own with Braun, Broxton & Santana in the outfield; Thames, Villar, Arcia & Shaw around the infield and some combination of Susac/Bandy/Pina behind the dish.
I’d take the over on the current 67 win projection but at the same time wouldn’t be surprised or disappointed in any way by a bottom five record.
I’d call that rotation more fun than truly good, so I think that’s the actual biggest hurdle to getting back in the contention discussion. I like that team, I just think they are a couple of big pieces away from worrying too hard about the bullpen.
I would still imagine that Guerra and Braun get moved, which would set them back a little this season, but I more or less feel the same. Pretty decent up & coming pitching, and pretty well positioned for the near future (2019?).
The E5 Steamer projection is conspicuously light and feels a bit like a cherry pick to compare Thames’ projection against:
4.3, 4.0, 3.6, 4.5, 3.9 WAR in the last five seasons for e5 projects out to 2.2?!? in 2017….
After reading this piece by MGL recently I am more convinced than ever that projection systems are penalizing “older” players too severely:
https://mglbaseball.com/2016/12/21/a-new-method-of-constructing-more-accurate-aging-curves/
On the other hand I recall Jose Abreu having some very favorable translation projections and he turned out to be nearly as good as a lot of them (with the bat anyways).
Projection systems are building in risk that old guys turn to nothing instantly into their projections. Obviously, not all will collapse, but enough do to drag the estimates down.
The market probably overrates youth, but not in an unwarranted sense. Mike Leake is mediocre at best rubbish and John Lackey has been great forever, but how do you know if you are getting 2016 Lackey or 2016 James Shields?
If you read the MGL piece you’ll see that he’s trying to correct for some of the effect you mention.
Yes, older players can/do have precipitous production drops but they happen both due to performance and playing time drops. His method looks to “add back” the lost PA’s of players that hit hard times and lose their jobs before they get a chance to play more and (potentially) regress back toward the mean. His adjusted method is much less steep on the 30+ side of the curve and I am willing to believe the truth is at the very least somewhere between. It’s not hard to believe that some of the worst performers to start the season who then go on to lose their jobs played under their true talent levels for the period observed.
Short version: Lots of players that show major declines in year two using the delta method had particularly bad luck and are given no more chances to regress positively so their numbers are artificially dragging the delta method year two bucket numbers down. Adding back the remainder of their year 2’s projected stats to the observed year 2 performance softens the 30+ side of the aging curve pretty significantly.
I probably should have taken into consideration this: If a player who starts poorly in “year 2” gets reduced playing time or is dropped completely it MAY be likely that his team knows a little more than a projection knows. Not that the bad performance is an exact reflection of his true talent. It isn’t. There is no doubt that these players got unlucky overall before their PT is reduced or they’re released or demoted. But it also may be true that we have to “nick” their projection a little, which I didn’t do, to account for information that may be relevant to the player’s true talent and somewhat independent of the stats – that is, the team is telling us that they think the player’s true talent has dropped to a point where he is no longer playable or deserves less playing time.
I think there’s an agency issue when citing teams’ perceived lack of interest in Thames based on their internal rankings. It’s very easy for a GM to convince an owner that this restoration project from Korea came out of nowhere and your talent identification just missed. It’s tougher to explain why the millions you shelled out to get him were a worthwhile risk to take as he struggles to meet expectations.
In other words, I expect there’s some GM self-interest driving selective skepticism of internal projection systems.
That’s simply risk-aversion, which I’m not sure applies to many front offices.
I’d suspect that most front offices are pretty risk-neutral: they understand that the $5M prospect might not pan out and the $15 free agent might not pan out, but as long as their process results in enough hits, then they’re doing okay.
I wonder if several teams weren’t interested in Thames, but Thames wanted to sign somewhere that he was basically guaranteed a full time job right away. The Brewers were maybe the one team willing to say sure, we will hand you the keys the day you sign on the dotted line.
I’m guessing that some other teams would have paid the same or even a bit more, but they were likely asking Thames to come to camp, get ready to play a few days a week, a bit at 1B and a bit at DH or LF, etc.
He was a superstar in Korea, so Thames likely wanted the assurance of a full time job right out of the gate. Good for the Brewers for giving him the opportunity, and to Thames for earning it.
He was interviewed on The Ringer. Seemed pretty shocked to be coming back. https://www.youtube.com/watch?v=xi3sTT8VC38
This is going to be one of the more interesting ongoing storylines of the year, quite possibly the move that is a tipping point in giving teams a push towards looking at the Korean league a little more like the Japanese league.
Having watched Thames look lost for a couple years as a Jay, though, it’s just really hard to see an old dog like him learning new tricks, but maybe there’s something to be said about the mental effect on a player in his own abilities when he’s having such great success for such an extended period of time.
Or maybe Barry Bonds would have hit 500HRs in a season in the Korean league…
I would take the under. Bonds would of hit a home run if pitched to about 1 in every 4 at bats in Korea
Looking at the full Steamer projection I was surprised to see a 13 SB projection for him, especially given his numbers displayed on Fangraphs. Surprise, he stole 40(!) bases in 2015 in the KBO, sandwiched between years of 13 and 11.
It seems that Davenport’s page is missing his 2016 stats, so that projection may not be updated yet.
Is his name pronounced the American way or the British way?
I believe Eric pronounces it the American way and Marcus the English. Or do I have that switched?
The British way is the only way to say it!
Goodbye Eric Thames 2017 fantasy sleeper!
Staggering projections. Were they similar for Kang?
Most east Asian players that come over are pitchers and they have had good success in recent years, but the hitting track record is very sparce.
I think he’ll produce this year.