Shifting the Florida Infield
The Florida Marlins are a very potent offensive club. Unfortunately, the same cannot be said for their defense. Last season, their starting infield featured Mike Jacobs, Dan Uggla, Hanley Ramirez, and Jorge Cantu. Though these four players combined for approximately +90 runs with the bat, Uggla’s +0.3 UZR topped the bunch defensively. Despite defensive improvements from Ramirez and Uggla, their projections for next season still place them around a half-win below average with the glove.
One member of this quartet will not return in 2009, as GM Larry Beinfest sent Jacobs to the Royals this offseason. According to recent reports, however, a new infielder might be joining the team: Orlando Cabrera.
In a recent analysis of Cabrera’s future, I surmised that his value for next season is right around +3.1 wins. He isn’t a terrific hitter but is durable and plays solid defense at the toughest non-catcher position. The problem here is that Hanley Ramirez plays the same position. While it may seem obvious to place the better defender at the position, the more likely scenario involves Cabrera manning the keystone corner.
Except Dan Uggla plays that position. Again, even though Uggla had an average year defensively, Cabrera would likely play much better defense. With Cabrera at second base, where does Uggla go? He could play first base and replace Jacobs in this scenario, or he could head over to the hot corner. Assuming the signing of Cabrera results in Uggla moving to third base, Cantu gets displaced to first base.
Signing Cabrera would most likely see him playing out of position, Uggla moving to third base, and Cantu moving to first base. Unless Cabrera plays remarkably better defense at second base than at shortstop, his win value decreases to somewhere in the +2.7 to +2.9 range. Uggla’s suspect defense could be exploited less at third base, though, perhaps resulting in a higher UZR rating. Cantu, however, loses plenty of value with the shift to first base due to the positional adjustment.
How would their infield look under different circumstances? If Cabrera is not signed and league average play at first base occurs, the foursome is worth right around +13.5 wins. With Cabrera at second, Uggla at third, and Cantu at first, the infield is worth approximately +14.1 wins. If Uggla plays first, with Cabrera at 2nd, they are worth +13.3 wins. Therefore, the most productive alignment would be the proposed shift assuming Cabrera signs.
Of course, if he doesn’t sign, this has all been one monumental waste of time, but I always find it very interesting when values shift with different defensive alignments. As we saw here, merely moving players around the diamond could see an improvement of up to +0.8 wins.
25
Interesting article. It would be fascinating to see something like this for other teams that looked into how much their fortunes could potentially improve by shifting around various players to different defensive positions. I’m think specifically of the Reds. There has been a lot of talk about moving Encarnacion, Phillips, and Votto to new positions. I wonder how their values and the team’s fortunes would change depending on how these players were utilized.
What about playing Cabrera at SS and moving Ramirez to 2B?
If you play Cabrera at SS with Ramirez at 2B, Uggla at 3B, Cantu at 1B:
Cabrera: +3.1
Ramirez: +6.1 (pretty much the same at SS. I figure he’ll earn a few runs fielding but lose them right back in positional adjustment)
Uggla: +3.4 (at 3B, his defensive shortcomings could be hidden more)
Cantu: +1.6 (at 3B, his value is increased, but the adjustment for 1B hurts him here).
That totals to +14.2, virtually equivalent to the proposed shift.
It’s an interesting exercise but for a team as extraordinarily cheap as the Marlins are, a free agent signing of this type seems really unlikely. Isn’t it more likely that they play Cantu at 1B and Dallas McPherson at 3B?
Definitely a feasible scenario, but this all came to be based on the idea that Orlando Cabrera is rumored to be on their radar, which would shift the defense.
It’s interesting to project Hanley’s future defensively. Peter Gammons was discussing this at the Winter Meetings and mentioned that the Red Sox were skeptical that he could remain at SS, but would rather have to eventually shift to 3B or CF. With Maybin scheduled for CF, what would the infield value be with Hanley at 3B, O-Cabrera at SS, Uggla at 2B, and Cantu/MacPherson at 1B?
I think this is an example of positional adjustments getting in the way here. Assuming you have the same 4 players, their offensive contribution is going to be the same no matter what defensive position the play. Therefore, the best way to get the most net runs out of them is to align them based on the highest net total runs they can prevent. Docking Cantu 12.5 runs by swinging him to first is somewhat useless, unless you factor in signing some other 1B to play the position. But then you’re bringing in new players into the equation. Ultimately, if Uggla can play 3B better than Cantu, Cantu can play 1B at least as well as whatever they had last year, Cabrera can play SS better than Ramirez, and Ramirez can play 2B better than Uggla, then that’s the alignment you should have. Factoring in positional adjustments into that calculation makes things unnecessarily complicated. Cantu may become less valuable compared to the rest of the league, but not less valuable compared to the Marlins, he doesn’t subtract 12.5 runs from the Marlins just by jumping to 1B.
James,
They aren’t getting in the way though I understand what you’re saying. The positional adjustments are there to level the field of play so a -1 SS isn’t considered equal to a -1 1B, etc.
With Cantu, I have him as something like +7.5 batting, -6 defense, and +25 in adjustments at 3B for 2009, which would place him right around +2.7 wins.
If he plays 1B, he gets the +7.5 batting and might even improve defensively. Say -2 runs at 1B. He still gets the +22 for PA adjustment, but instead of a +2.3 pos adj, it’s -12. At 1B, +7.5-2+22-12 = +15.5, right around +1.6 wins.
Your argument is basically that it’s common sense that Cabrera > Ramirez at SS and Ramirez > Uggla at 2B, and Uggla > Cantu at 3B. I’m generally of the same opinion, but I chose to quantify it here.
Aren’t they based on the fact that the average defender would lose/gain that amount of runs switching positions fielding-wise?
Does Uggla have a particularly good arm? I thought the general consensus from the 2B / 3B defensive comparison was that the positions were of about equal difficulty unless the particular player’s skills were much more suited to one than the other.
Uggla sure doesn’t look pretty out there, but if his UZR is about 0 then he’s obviously not all too bad, either. It’s odd to me that over the last 3 years his UZR is very inconsistent. I wonder what that’s about.
I have a great deal of trouble envisaging a Marlins team, which traded Kevin Gregg because his salary was getting too high, paying the going rate for a FA SS of Cabrera’s calibre AND giving up a 1st round pick as well. I rather think that the A’s are much better positioned to go after Cabrera and their need is definitely greater which suggests they’ll be more aggressive bidders than the fish.
Before you brought the other Cabrera into the conversation, I had thought that the Marlins did need to move Cantu off 3b. Ramirez and Uggla both had ugly defensive numbers in 2007, not so bad in 2008. Was it the pitching? But Cabtu has always been terrible, earlier at 2b and now at 3b.
The trades of Willingham and Jacobs opens up lf and 1b. Put McPherson at 3b, move Cantu to lf, put Gaby Sanchez at 1b.
Eric, not a waste of time. Keep em comin!!!
Hey, I have a few problems with the way you went about this.
First off, you say that Cantu would be poor at first because of a position adjustment. However, he has played first base in the past and has been above average at it. Meanwhile, the notion that Uggla would do better at third base doesn’t hold much ground because he struggled at the position previously, as well as the fact that he does not have a very accurate arm. If there is a player that will struggle with a positional change, it would be Uggla to third base.
I also think you’re downing on Cantu’s defense at third too much. The guy continually got better at third as the year went on. Now THAT was a position adjustment for him. He will still be below average, I agree, but not as bad as he was last year.
I also certainly don’t understand why OC would be such an upgrade defensively over Uggla. The guy used to be above average, sure, but in the past three years he’s put up a -9 +/-. Yeah yeah most of that was his bad ’06, but even then the past two years he’s been the definition of average. Yes SS is the epitome of IF defense and putting him somewhere else raises his defensive value, but that also just puts even more of a burden on his ~.700 OPS.
Best regards.
I honestly don’t think you understand what position adjustment means
Even if that was my misunderstanding of the sentence, the point still stands. Cantu has been an above average 1b in his career but is docked points. Uggla has been a below average 3b in his career with a poor, inaccurate arm but is awarded points.
And moving Uggla to third base makes 0 sense. He has league average 2b defense, and he’s one of the top offensive 2b’s in the league. Having Dallas at third (who has been an average defensive player) and Cantu at first would be better. Having the like scenerio of Gaby at first (who is above average defensively) and Cantu at third is still better than going out and grabbing a player like OC, who’d barely be above average defensively at the position while being well below average with the bat even though we’d have to deal with Cantu’s below average defense at third.
There’s also the fact that this will very very likely be the last season of Uggla in a Marlin’s uniform. It would completely deprecate his trade value.
Well if that is how you feel, I am sorry you are blogging about the Marlins.
I agree some with Nny here.
If you are signing Cabrera, it probably makes more sense to move Hanley off SS than to move Uggla off 2B. Hanley is probably more of an improvement over Uggla at 3B as he would be at 2B, and if you are bringing in Cabrera for his defensive value, it doesn’t make sense to play him out of position where his defense would initially suffer.
Keep in mind, a big part of the improvement you are going to see from upgrading the defense is psychological, as far as the impact on a young pitching staff. The Marlins are trying to do something similar to what Tampa Bay did last year with Jason Bartlett. If you try to measure it objectively, Bartlett, at 1.7 “value wins” was only 0.6 wins better than Brendan Harris. Defensively, he was only better by about 7 runs.
But, something made the Rays a whole lot more than 7 runs better iin 2008 defensively. They allowed 273 less runs in 2008, going from 14th in the AL to 2nd in runs allowed. Can we objectively account for all of those runs? Judging by FIP, the pitching staff improved by about 46 runs. That still leaves 227 runs to account for, and you won’t find more than a fraction of them in measurable defensive stats.
Regardless of what the stats say, though, the impression of many around the team seems to be that the addition of Bartlett (as well as Longoria at 3B) helped improve the confidence of a young pitching staff. I’m not sure how to measure that.
If the Marlins are similarly trying to boost the confidence of their young pitchers, in addition to finding some measurable savings in runs, then having everyone on the infield learning new positions probably doesn’t accomplish that.
With Uggla, I would either keep him at 2B, or trade him. If you do bring in Cabrera, put him at SS, and move Hanley (3B? 2B? OF?). And Gaby Sanchez is probably the best choice at 1B.
I am a big Marlins fan and have been a proponet of moving Uggla to third, and Cantu to first and finding a 2B answer for a while to my friends and on my blog.
The overall comments that moving players may adversely effect them, psychologically, may have validity, especially to a young pitching staff.
However, it seems some of you do not understand positional adjustments and I perhaps need to explain it better.
The reason we factor the adjustments in is basically in order to show that certain positions are much harder to play than others. A +10 SS is NOT equal to a +10 3B. With a +7.5 run adjustment per 162 games for SS and +2.5 run per 162 for 3B, the SS is actually +17.5, while the 3B is actually +12.5 in the field.
The adjustments do not reward or subtract actual fielding runs from the players, but they allow us to compare different positions much more accurately, since the UZR compare players to their position, not everyone in the league.
Cantu may have been above average defensively at 1B, but 1B is much easier to play than 3B, which is why he is docked runs. It has nothing to do with HIS defense, but rather the conversion of going from 3B to 1B. For all we know, he might be -8 at 3B with a +2.5 adjustment, for a net of -5.5 in the field. Moving to 1B might allow him to be a +8 fielder, but +8 at 1B is not tremendous defensively compared to other positions. With the adjustment, even with the improved defense, he would be -4.5 runs in the field, a slight improvement.
Matt – I’m not “blogging” about the Marlins and to be quite frank, I think blogging is idiotic way of saying “My opinion is better than yours about a team, so shut up and listen.” Considering your attitude here, where you can’t debate a point and instead lash out insults, you seem to fit into that demographic nicely. I mean seriously?
Eric – Thank you for the clarification, as I do not pay much attention to fangraph’s stats (to be up front, I really dislike the UZR stat, and MUCH prefer +/- and RZR/OOZ, which it self invalidates a lot of fangraphs win statistics since they then depend on defensive statistics…much like how I despise BP’s WARP because I think FRAA is a bad stat).
However, I don’t really understand that. If you are + runs in the field, aka above average, I don’t see how that would then cost runs. Although this is also under the assumption that 0 = average, so if it’s going by something other rate, than obviously things will differ.
Although it’s also assuming that the +wins is based off average. If it’s just a strait +win, than yeah, I’d understand it. Glossary doesn’t really state if it’s based off average or not.
(Also just an aside, the blogging comment wouldn’t relate to you, as this would not be a blog about an opinion towards team(s). Rather a blog pertaining to information about baseball. Difference between displaying information to fans, and saying “this is what I think and it’s right and if you don’t think it you’re dumb”)
NNY,
The best way I can explain it is that the defensive metrics calculate average on a position by position basis. So, if you see that a 3B is +5 runs, it means he is 5 runs above an average 3B…NOT 5 runs above an overall average defender. The same for SS. A +5 SS is 5 runs above an average SS, NOT everyone else.
The positional adjustments are there so we can compare a +5 SS to a +5 3B. Clearly, SS is tougher to play, but how much tougher. If we have an SS and 3B with equivalent above average marks, they are not equivalent fielders because the positions are not equal in difficulty.
So, the adjustments come into play in order to normalize the field of play. SS gets +7.5 runs per 162 gp while 3B gets +2.5 runs per 162 g. If both players play 160+ games, then the +5 SS is actually, overall, compared to ALL other positions, not just SS, a +12.5 fielder. The 3B is, all told, a +7.5 fielder.
If we had a 1B in there, at +5 runs in 162 games, he loses 12.5 runs because it is the easiest position relative to others. He is actually, all told, a -7.5 run fielder. He might be great compared to other first basemen, but if we want to compare the fielding of a 1b to that of an SS, we need an adjustment to level the playing field of positional difficulty.
Ok, now I understand the point of it, about being able to defensively compare different positions…but then I don’t really see the point of it when it comes to added wins though.
i.e. you have a SS and a 1B, both are exactly +0 in offense and defense.
You can either add a SS that will save 10 more runs than the average SS (aka your SS), or a 1B who will save 10 more runs than the average 1b (aka your 1b).
Going by 10 runs = 1 win, that means that each is adding 1 win over their replacement. In other words, the +10 from SS adds as many wins as the +10 from the 1b
The way we evaluate players is four-fold. First, find the offensive runs above average which comes from park-adjusted wRAa. Second, incorporate UZR for that season. Third, prorate the positional adjustment per 162 games. Fourth, prorate th +20 runs per 600 PA for value above replacement, not average.
If we have an SS and 1B who are both league average offensively and defensively, who amass 600 PA in 162 games, their values are:
SS: 0 + 0 + 20 + 7.5 = 27.5 = 2.75 wins
1B: 0 + 0 + 20 – 12.5 = 7.5 = 0.75 wins
The adjustments have to be included in order to normalize fielding. A player’s true fielding contribution is Fielding Metric + – adjustment.
look – we don’t need the offensive positional adjustments here because we’re not comparing the individual players to other players. we have our 4 players and we can assume that our team will score the exact same number of runs regardless of what defensive positions they play, because the lineup will be exactly the same regardless. so defense is the ONLY thing we need to quantify.
if you want to compare our 4 man infield to last years 4 man infield, then you figure out the offensive value of our 4 and compare it to that of last year’s 4. (and then figure the defensive values of the various alignments this year.)
but it’s useless to say that cantu will be less valuable if we move him to 1st, because it just isn’t true. here’s another way to think about it: he doesn’t have any positional advantage playing 3B on the marlins because the marlins have at least 2 other good options at 3B. 100% of positional advantage at 3B over 1B comes from being able to find better replacement offense at 1B than 3B. but we have enough information with this marlins team to make that concept useless, because we know exactly who the replacements will be!
ohhh. nevermind. i must have this all wrong. when you talk about positional adjustment you’re trying to adjust fielding value!? wow. i think i know what you mean but that’s a really confusing way to think about it. i mean, i think it’s much easier to say that a +5 1B is worth exactly as much as a +5 SS defensively, all else equal. (that is, if your replacement at each position is one guy who can play both SS and 1B at the league average) and then compare their offensive contributions to their position. i think you’re doing the same thing, just by a different method (correct me if i’m wrong). i think you’re NOT normalizing offense, and de-normalizing defense, we i am leaving defense normalized and normalizing offense to compensate. so i THINK (though i’m not 100% confident) that my point still applies.
“With Cabrera at second, Uggla at third, and Cantu at first, the infield is worth approximately +14.1 wins. If Uggla plays first, with Cabrera at 2nd, they are worth +13.3 wins”
obviously this has to mean that 3B Cantu/1B Uggla is 0.8 wins better than 3B uggla/1B cantu
i’m going to say that you can’t know that that’s true. you can say that cantu is a better 3B if you think it’s true – but it’s not something you can measure in advance. and the only thing it has anything to do with is the defensive abilities of those 2 players. (because everything else is exactly equal)
With the improvements Hanley made last year, it seems like it would be stupid to move him off of SS now. If the guy can be an average SS defensively you keep him there. If the goes back to a -20 UZR, then you have to move him. But if he hovers around a 0 UZR, you take it and get a bigger bat at an easier position. The guy is 25 years old and amazingly athletic and talented…I wouldn’t be the least bit surprised if his improvements were real. To me, it looks like 2007 could have been the outlier, as his RngR #’s go -1.1, -13.3, 2.5. He has been error prone all three seasons, but the range looks like it is league average.