The Scouting Report, By the Fans, For the Fans - 1B Report (March 24, 2004)
A look at 1B, and how you fans and UZR sees them. Lots of good stuff to check out here.
--posted by TangoTiger at 11:30 AM EDT
Posted 1:52 p.m.,
March 24, 2004
(#1) -
villageidiom
Great stuff, as usual.
One danger in regressing to the fans is that, sometimes, the fans' evaluation could be based on reputation or distant skills than on current demonstrable skills. On the whole I'd expect it to work, but the danger comes once you start looking at individual players.
From a ballot perspective, I'd expect such reputational voting to come through on ballots where players on many teams were evaluated. We're all fans of baseball, but most tend to watch games of one or two teams often enough to make a judgment. If someone is evaluating players on 20 teams, in most cases they either (a) have TiVo and more than 24 hours in their day to watch it, or (b) have evaluated players based on a small sample of observation.
To test this, I suppose you could look at the average number of players evaluated per ballot that a given player appeared on. (Perhaps variance would be better?) For example, how many different teams are represented on each ballot where John Olerud was evaluated? The assumption is that higher averages (or variances) would indicate greater "reputational" evaluation.
As your sample of fans gets larger, I'd hope that actual witnesses will far dominate the guessers out there. Until then I'd fear that small samples will give nutty results at an individual player level. And, if that is happening, then regressing to the fans' rating would be akin to the Gold Glove award balloting.
BTW, could you add sample size to your table?
Once again, on the whole I think the fans' scouting report is fantastic, endlessly fascinating, and a reliable and independent confirmation of UZR. I'm just not sure how much credence should go to the fans' rating at an individual player level. IMO a simple 1/3 weight isn't enough. Perhaps the weight should vary by sample size, with a max weight of 1/3 and a min of 0?
Posted 2:14 p.m.,
March 24, 2004
(#2) -
tangotiger
Excellent comments. Let me just take a few.
One danger in regressing to the fans is that, sometimes, the fans' evaluation could be based on reputation or distant skills than on current demonstrable skills.
I was thinking about that as well. The equation should probably be: Fans * .25 + UZR * .50 + mean * .25, or some such.
Perhaps the weight should vary by sample size, with a max weight of 1/3 and a min of 0?
Absolutely. With UZR, the regression towards the mean is 420/(420+BIP). So, I'd have to use some similar thing here, and not the above equation I just listed.
If someone is evaluating players on 20 teams
This is nowhere near the case. I think the most I had was 3 teams from 1 person. 95%+ of the people voted for 1 team.
BTW, could you add sample size to your table?
Ugh. I keep meaning to do this, and I keep forgetting. The reason is that some players have 10 evaluations for instincts but 12 for speed, etc. I'm thinking of just doing the sum of the total evaluations and divide it by 7.
Posted 3:42 p.m.,
March 24, 2004
(#3) -
dlf
MGL's excellent defensive metrics (UZR, arm, dp) seem to miss two skills of 1Bs: the "scooping" ability and the 3-6-3 / 3-6-1 double plays. I don't know how much impact either would have on a complete evaluation. But the fan's vote has 4 categories (hands, release, strength, and accuracy) that are not directly reflected in UZR. It should be no surprise that those four categories show the lowest correlation to UZR.
Posted 4:05 p.m.,
March 24, 2004
(#4) -
tangotiger
Great point!
I'm trying to come up with the best weighting scheme (by position). Right now, for 1B, I'm double-weighting Acceleration and Speed (because of their high correlation to UZR), and Hands (because of the scooping ability which is not captured in UZR). I'm not sure how important the DP-type skills are for a 1B. After all, only about 20% of PAs have a DP in effect, I believe. So, even if it was very important, it would only apply 20% of the time, thereby dragging it down to very unimportant.
In terms of covering for the sac bunt, I see Speed, Hands, and the 3 throwing categories as important. But again, not many bunt attempts to begin with.
Are we satisfied with my current weighting scheme? I'm tempted to half-weight the throwing categories, simply because of lack of opportunities. Maybe keep Release as single-weighted to capture the 3-1 plays.
Thoughts?
Posted 4:07 p.m.,
March 24, 2004
(#5) -
MGL
Actually, the "DP" part of UZR is as a pivot man AND as the starter of the DP. So the 3-6-3 and 3-6-1 DP IS included in the first baseman's UZR plus DP. Actually, I don't know whether Tango is including the GDP for IF'ers (and arm for OF-ers). I don't think so. Scooping is NOT included in UZR for 1B'man (I wish that it were - perhaps next year)...
Posted 4:17 p.m.,
March 24, 2004
(#6) -
tangotiger
MGL: you know, I never noticed that 1B had that. To confirm, I am only looking at UZR, though in these positional reports, I should include the Arm as well.
For 1B with at least 40 games, the range is from +2 to -2 runs for the arm. Bumping up the game requirement to 140, and it goes from +1 to -1. Interestingly, Minky has a -1.
In any case, this goes to the heart of my point. Among 1B with at least 120 games from 1999-2003, the SD for UZR is 9.6, while for the DParm it's 0.6. I think giving the 3 throwing categories a half-weight might even be too generous.
Right now, I'm thinking:
1: Instincts
2: Acceleration, Speed, Hands
1/2: Release, Strength, Accuracy
Total weights = 8.5, of which 1.5 are for throwing (or 18%).
Posted 6:27 p.m.,
March 24, 2004
(#7) -
dlf
MGL, Thanks for the clarification. The data is presented in your charts, but not in Tango's summary here.
Is there a source for GIDP started by the 1B somewhere on the web? If not, can someone suggest a good book that includes that data? Anecdotally, it seems like guys like a Don Mattingly who played off the line and had a decent arm started many, many more 363 and 361 DPs than did the ones like Steve Balboni, who couldn't throw the ball 90 feet without a cutoff man relaying it. Certainly many more seasonally than would be reflected in a +/- 2 runs. But I have nothing concrete to support that observation.
Posted 10:27 p.m.,
March 24, 2004
(#8) -
MGL
As is often the case, I think that the "value" of the 3-6-3 or 3-6-1 DP is much overrated. Basically to calculate the DP portion of UZR I take all the DP opps (runner on first and less than 2 out) and see how many DP's a fielder starts (say at 1B) and how many they are a pivot man on (for SS and 2B of course). For every "extra" (plus or minus) DP's above or below expected DP's (based on the league average number of DP's per DP opp at that position), the pivot man gets "credit" (plus or minus) for half of the .45 runs (or whatever the average diff is between a DP and only one out) and the fielder that started the DP gets the other half "credit." Based on the low SD for first baseman (.6 runs), we have to pretty much conclude that the ability to turn the 3-6-3 or 3-6-1 DP is just not worth very much (probably because it is relatively rare despite what appears to be the case). Even the DP's for the SS and 2B, which include starting the DP AND the pivot, aren't worth all that much. FWIW, the y-t-y "r" for GDP defense is around .35 for an average IF'er per 125 games (obviously more for the SS and 2B and less for 3B and then again 1B). The y-t-y "r" for OF arms is around .4 for 130 games...
Posted 10:39 a.m.,
March 25, 2004
(#9) -
dlf
Just curious: what is the basis for weighing the player who starts the DP equally with the pivot man? How much would the ratings change if the player who commences the play got 75%? Or if he got 25%?
I'm trying to follow the math. Runner on first, no outs, the run expectation is somewhere around 0.8 runs. If the 1B turns a 363 and there are no runners on, two outs, that expectation drops to roughly 0.1 runs. If the 1B gets the lead runner, its about 0.5; if he only gets the batter, its about 0.7. So for each 363 or 361 that is turned, the advantage is between 0.4 runs and 0.6 runs. Roughly similar numbers would apply with 1 out or runners at the corners. (I assume very few 363 / 361 turned when the runner at 1st isn't held on, i.e. when a runner is on 2nd as well.) At one half run per occurance, it would take less than one extra 363 or 361 DP per month to go right to the top of MGL's charts. Are there really that few 363 & 361 DPs?
Posted 10:58 a.m.,
March 25, 2004
(#10) -
tangotiger
I spoke with MGL about this a while ago. You should not do the "divide by 2". We don't do that for pitchers/fielders, or hitters/pitchers, and we shouldn't here.
If the average OBA is .34, and the batter gets on base, he doesn't get +.33 times on base, while the pitcher also gets +.33 times on base. If the average BIP is .30, and the batter gets an out, the pitcher doesn't get -.15 hits and the fielder doesn't get -.15 hits. That's not how it works.
Suppose you have the following combos:
Jeter/Sori: .40 DP / opp
Jeter/Wils: .40 DP / opp
ARod/Sori: .50 DP/opp
ARod/Wils: .50 DP/opp
avg: .50 DP/opp
What this means is that AROd, Sori, and Wils are all average, and that Jeter is -.10 DP/opp.
If the Jeter/Sori combination did 40 DP per 100 Opp, you would give Jeter -10 DP and Sori 0.
For a more extensive explanation, you should see what I did with catchers:
http://www.tangotiger.net/catchers.html
On the other hand, with such little impact with DP, MGL's process is ok. But, I wouldn't do it that way.
Posted 11:02 a.m.,
March 25, 2004
(#11) -
tangotiger
Note: my example here was very simple. Sori and Wil could be +.1, Arod -.1, and Jeter -.2. This is why I suggested reading the catcher article.
Posted 11:05 a.m.,
March 25, 2004
(#12) -
dlf
Tango,
I would take it a step further and look at 463 and 643 for each of the player combinations seperately. Similarly, to find the defensive value of the SS at turning two, look at the 543 DPs to help sort out how much of the value comes from the 2B. (I'm not sure if that is what you were saying by looking at A-Rod above; I interpreted it as looking at his performance historically as a SS, not tomorrow as a 3B.)
I have no data to back this up, but anecdotally, reading biographies and autobiographies of good defensive infielders, it almost always seems that they credit a quick and accurate throw to the pivot man as the key to turning two. My observation would reflect that as well - although I recognize that I may be seeing only what I expect to see having read all those Graig Nettles, Ozzie Smith, etc. articles and books.
Posted 11:15 a.m.,
March 25, 2004
(#13) -
tangotiger
Yes, absolutely.
You figure ARod-[specific2B]-1B, and repeat this for all SS (like my Gary Carter example). That's for the 6-4-3 DP skill.
Then, [specific2B]-ARod-1B, and repeat. This is for the 4-6-3 skill.
You would probably want to limit it to a 4 year period, as aging will probably come into play here alot more than with pitchers/catchers.
Posted 12:33 p.m.,
March 25, 2004
(#14) -
tangotiger
I updated the 1B report for the following:
- added DParm to the UZR
- resorted based on Fans Runs
- did the correlation between Fans Runs and UZR162 (I actually had the correlation between Fans Runs and the UZR162 for 2003 only; now I have it for both years)