In last week’s installment, we looked at the 2013 partial-season performances of Khris Davis and Scooter Gennett — specifically, their walk rate, strikeout rate, and home run rate — to see what those stints told us about their actual ability. Even factoring in some regression next year, we saw that, among other things, Davis is expected to hit fewer — but still a lot — of home runs and that Gennett projects to improve his walk rate, although he still likely will be below league average in that category.
I won’t recap everything from last week’s article, but for those who are new to this topic, regression analysis tries to strip random variation out of a baseball player’s performance to better understand his true ability. This is important when looking at small samples from new arrivals, who may be benefiting from luck as much as sustainable baseball skills. Heading into 2014, the Brewers are looking for players that will make ongoing contributions, rather than a flash in the pan.
We regress a player’s performance by recognizing, first, that baseball players tend to either improve or decline back in the direction of the “mean,” or average league performance, from wherever they currently are performing. We then focus on certain peripheral statistics — among them walk rate, strikeout rate, and home run rate — that stabilize fairly quickly and allow you to get much of a player’s measure without a full season’s work. When you properly weigh and combine the player’s actual performance with the average league performance, you get a better picture of a player’s true ability, stripped of some of the luck that affects everyday baseball. I used Derek Carty’s method, which extends the work of Tom Tango and others, to make the calculations.
Last week, we discussed Khris Davis and Scooter Gennett. Here are the regressed batting statistics for the rest of whom I labeled “young” Brewers hitters:
|Name||Age||PA||HR||K Rate||Reg. K||BB Rate||Reg. BB||HR Rate||Reg. HR||2013 Actual HR Pace||2013 Regressed HR Pace|
I like this table in part because you can see the differing effect of regression depending on the number of plate appearances a player has received. Jean Segura, for example, had over 600 plate appearances this year. That sample so far exceeds what is necessary to stabilize his peripheral statistics that the regressed statistics (“Reg.”) are virtually identical to his actual ones. So, in 2014, you should expect Jean Segura to walk a little more, strike out about the same (not very often), and, if he makes about 500 plate appearances, to once again hit double-digit home runs, although not by much. (FYI: Segura’s home run “pace” for 2013 is lower than his ultimate number because the pace assumes fewer plate appearances). Logan Schafer is basically in the same boat; his statistics don’t move at all, even with regression. You can contrast that with Sean Halton, who, with a mere 111 plate appearances, has all three peripheral statistics experiencing a notable adjustment.
Of the other players on the chart, the two most interesting to me are Caleb Gindl and Juan Francisco. I’ll conclude by discussing them both.
If Khris Davis had not stormed onto the scene, Caleb Gindl would probably be the leading option to supplant Norichika Aoki in the starting lineup next year. Like Davis, Gindl experienced significant playing time in 2013 as a result of Ryan Braun’s injuries and suspension. During just over 150 plate appearances, Gindl drew walks at a significantly above-average rate, demonstrated equally above-average ability in avoiding strikeouts, and also produced home runs at a healthy rate, even for an outfielder.
Regression tempers Gindl’s performance a bit, but still suggests he has the core skills to contribute in 2014. He should continue to draw walks at a double-digit rate (10%), which will cushion any hitting slumps and keep getting him on base. His low strikeout rate may be useful for pinch-hitting, and he projects the ability to hit 15 to 20 home runs over the course of a full season. Still, it seems unlikely Gindl will spend much time in the majors without injury to some other player. Braun, Carlos Gomez, and Khris Davis all project to be more productive, and Logan Schafer provides superior defense as well as a left-handed bat when the Brewers seek a platoon advantage in particular situations.
Juan Francisco is, as Ryan Topp frequently describes him, a lottery ticket. The Brewers bought low on Francisco, claiming a hitter with enormous power on waivers who had failed to put it together with both the Cincinnati Reds and the Atlanta Braves. Francisco caught some attention early when he muscled several home runs over the fence during the summer months. His peripheral statistics, even regressed, also show why the Brewers remain interested in his progress. Although he strikes out way too much (31%, regressed), he shows the ability to draw walks at an average rate (9%) and is probably able to channel his power to generate around 27 home runs. Normally, that would be considered workable for an aspiring first-baseman, at least in the offense department.
But Francisco also displayed red flags that make him a dubious choice for first base going forward barring signs of significant improvement. The most obvious is his high strikeout rate, which soared at the end of the season as opposing teams seemed to figure Francisco out. Primarily a fastball hitter, Francisco could not adjust as teams peppered him with a steady diet of sliders and other off-speed pitches. Tom Haudricourt and others noted repeatedly in September that Francisco looked entirely lost at the plate. Adam McCalvy of MLB.com reported that the Brewers have adjusted Francisco’s swing and sent him to winter ball to see if he can show improvement.
The second major problem is that Francisco, despite coming over to the Brewers as a third baseman, showed no ability to play competent defense at first base. Although the Brewers’ offensive woes receive most of the press attention, an equally serious problem for the Brewers at first base was poor defense. As I read it, according to FanGraphs the Brewers’ awful defense at first base cost them 21 runs, or more than 2 wins last year. Two wins! Two wins lost by not being able to play even average defense at the easiest defensive position in baseball.
And although the sample size is small, Francisco was a major part of this problem. His UZR/150— a statistic which projects the effect of a player’s defense over the course of a season — suggests that his first-base defense would have cost the Brewers 27.1 runs — or almost 3 wins had he played the position all year. UZR is volatile, and you can’t take one year’s performance as the final word on anybody. But Francisco’s defensive statistics — which certainly agree with what most of us saw on the field — demonstrate that Francisco unlikely to improve the defense at first with more playing time. His poor defense, in turn, puts Francisco in a huge productivity hole, requiring him to generous considerable offense (5 wins or more) simply to reach the threshold of being a 3-win, or good overall player. On balance, Francisco is unlikely to be the answer at first base unless he shows tremendous improvement in his batting skills during winter ball.
Follow Jonathan on Twitter at @bachlaw.
- Derek Carty, “When Hitters’ Stats Stabilize” (Baseballprospectus.com 2011)
- Adam McCalvy, “Francisco Mulling Adjustments in Winter Ball” (2013)