Sabremetric philsophy.

[Left-handed Badger]

First off, I know in our VORP is mularkey thread, we know I wasnt a fan. (And still not)

Second, I dont want this to end in a whole 'nother battle about sabremetrics.


But, I am genuinely curious as to how a team would be built by a sabremetrician. I know Billy Beane and to some extent Theo Epstein are known to be disciples of it. But, I dont know for sure if they are totally or how extensive. I am really looking for more how they would order a lineup, what type of players (I have some ideas of hitter type, but I want to know more). I really dont know much about sabremetrics and pitching and fielding. So, that would be nice to know more with those 2 areas.

Course, my ideas probably want change much, but I would like to know more.

And to reiterate please stay civil all.


Not sure exactly what I expect but just try to remain polite.

[andymac]

I don't think there is any typical sabremetric way to construct a lineup. In fact, most studies show that lineup order is pretty much insiginificant. The general idea is simply to have your high on-base guys in the first couple spots, your high power guys in the middle spots and just fill in the rest in decreasing order of hitting ability.

Pitching from the sabremetric view is basically admitting that a pitcher has little control over what happens on a ball in play. He can control strikeouts, walks and home runs to significant degrees but that is it. So, you want to look for pitchers who excel consistently in those areas (i.e. striking hitters out, avoiding walks and avoiding home runs). A secondary skill to look at is gb/fb percentage and which type fit better according to your teams defensive strengths and weaknesses.

As for fielding, there are a number of different systems out there but none are considered to be particularly reliable, although reliable enough to show that Derek Jeter's gold glove awards were ridiculous.

There is a sabremetric wiki that is in its' early stages you may want to poke around at if you are truly curious.

http://tangotiger.net/wiki/index.php?title=Main_Page

[Left-handed Badger]

Anybody, whoever saw Derek Jeter play a game of SS, wonders how he won the Gold Glove.

But, yes, I do want to learn what I can, probably mostly for curiosity though if nothing else.

[Qwerty75]

One thing that the A's have done for the last couple of years with their lineup that bucks traditional baseball thinking is batting Nick Swisher in the 2nd spot. A "conventional" 2-hole hitter would have his primary strength in making contact with the ball to move the leadoff man over into scoring postion for the heart of the order, so the premium is on bat control and power is usually disregarded (think Placido Polanco). Swisher, on the other hand, strikes out a good deal (152 and 131 Ks the last two years), and has above-average power. He fit in the 2-hole in the A's lineup because of his high walk total and subsequent high on-base percentage (.372, .381) despite a low batting average (.254, .262). So, the A's in general 1) Don't put their leadoff man in motion because they don't have someone with speed who also can steal at a high percentage, 2) Don't believe in the "small-ball" practice of sacrificing an out for a one-base advancement, and 3) Put a premium on power for much of their run-scoring. It's in line with what studies have found about the value of not making outs vs. "small ball" and the value of walks and homeruns vs. high batting average. Some lineup studies have found that putting the "best" hitter in the 2-spot is optimal, although this isn't necessarily put in practice with either the A's or Red Sox. I seem to recall that McGwire was put in the #2 spot occasionally in his heyday with the Cards under LaRussa, but not sure what the thinking behind that was.

For pitching and defense, you should learn about DIPS theory (Voros McCracken is the originator) as andymac alluded to. It's the starting point to understanding the run-prevention side of sabermetric studies which still generates a decent amount of debate (e.g. can some pitchers control the outcomes of balls in play, and how?), while conclusions on the run-scoring side have pretty much been set down as orthodoxy and don't generate strong challenges within the researching community.

[wlight1]

Lineup order is easy. Current theory holds that the impact of lineup order is either none at all, or so small as to be insignificant. Therefore, you order your hitters from best to last 1-9 because the one thing that stats do prove is that the higher up a hitter is in the lineup, the more PAs he gets. I don't recall off the top of my head but I think that each lineup slot is worth something on the order of 15-20 extra appearances per season. So there's that.

Your ideas on hitter type are probably 'OBP is king', or follow something along those lines (I apologize if I'm mischaracterizing your perspective; you didn't really say). This is a very common misperception brought on by the wave of publicity following Moneyball's release. The idea isn't that one certain characteristic is better than any other, it's that smart GMs look for the qualities that other GMs haven't noticed yet. In 2003, that quality was OBP for hitters. I believe that the 'next big thing' on the forefront of smart GMs analysis, following the explosion in popularity of OBP, was defense.

Defense, of course, is extremely difficult to quantify. Most MLB teams (probably all, by now) have in-house stat teams with proprietary defensive metrics. In this respect, the baseball establishment is now moving ahead of the alternative press in terms of statistical sophistication. Money, as always, talks.

[1998 Yankees]

Originally Posted by Left-handed Badger

First off, I know in our VORP is mularkey thread, we know I wasnt a fan. (And still not)

Mr. Badger, your spelling is "malarkey" (var. "malarky").

[toxicavenger74]

Originally Posted by Left-handed Badger

First off, I know in our VORP is mularkey thread, we know I wasnt a fan. (And still not)

Second, I dont want this to end in a whole 'nother battle about sabremetrics.


But, I am genuinely curious as to how a team would be built by a sabremetrician. I know Billy Beane and to some extent Theo Epstein are known to be disciples of it. But, I dont know for sure if they are totally or how extensive. I am really looking for more how they would order a lineup, what type of players (I have some ideas of hitter type, but I want to know more). I really dont know much about sabremetrics and pitching and fielding. So, that would be nice to know more with those 2 areas.

Course, my ideas probably want change much, but I would like to know more.

And to reiterate please stay civil all.


Not sure exactly what I expect but just try to remain polite.

Looking at sabermetrics from a broader view, I work in the process engineering field and sabermetrics really is using process engineering techniques/statistics to attempt to get a higher quality product result at a more efficient or cost effective rate. Using more general statistics is attempting to achieve the same product but with the risk of doing so at less efficeint or cost effective rate.

RBI's for example are more of a symptom than a productivity measure. Sabermetrics attempts to look only at productivity measures. This is something that is done in all fields of manufacturing and it is known as Process engineering, continuous improvement, process improvement etc. etc.

There is the arguement that "Nobody cares about VORP at the end of the game only who scored the most runs". This is very true. The same can be said of any product. Nobody cares about the statistical analysis of the product they are using only that it is getting them the result they want. RBI's in effect get wins for a team, wins for a team gets championsships. However, what is the process in getting those RBI's?

Behind each product is a statistical analysis to attempt to get that same positive result with limited variation and deviation, in the most cost/time efficient manner. Overall, this is what sabermetrics is, it is not devaluing RBI's as many times high RBI guys are high VORP guys but it is evaluting the process of how those RBI's are generated or potential RBI's could be generated.

So, back to your question, in my opinion a sabermetric team would typically:
1. Value high OBP, power hitting players at C, SS, 2B as these are very costly to obtain and in limited supply.
2. Place very limited value on low OBP OF as these are easy to obtain (I'm looking at you Juan Pierre).
3. Place high value on pitchers with high K:BB strikeout ratio.
4. Have a short leash with players over the age of 30 as their productivity typically decreases.
5. Often will sacrifice defense, except for Catcher, if there is an increase in OPS.
6. not have low OBP, high RBI guys for more than one season.
7. shuffle closer role more often.

[turdfurgeson]

Here is a pretty easy primer on pitcher evaluation http://ussmariner.com/2006/08/29/eva...itcher-talent/

[NCBeachBum]

Originally Posted by Left-handed Badger

But, I am genuinely curious as to how a team would be built by a sabremetrician. I know Billy Beane and to some extent Theo Epstein are known to be disciples of it.

I would say the most useful idea GMs such as those two have brought to the forefront isn't really a sabermetric thing, but is probably worth mentioning. That is the targeting of qualities that are being undervalued by the rest of the market and stocking up on it. You can then use your surplus in that area to trade for those overvalued areas in which you have come up short.

[CBL-Commish]

Originally Posted by Left-handed Badger

First off, I know in our VORP is mularkey thread, we know I wasnt a fan. (And still not)

Second, I dont want this to end in a whole 'nother battle about sabremetrics.


But, I am genuinely curious as to how a team would be built by a sabremetrician. I know Billy Beane and to some extent Theo Epstein are known to be disciples of it. But, I dont know for sure if they are totally or how extensive. I am really looking for more how they would order a lineup, what type of players (I have some ideas of hitter type, but I want to know more). I really dont know much about sabremetrics and pitching and fielding. So, that would be nice to know more with those 2 areas.

Course, my ideas probably want change much, but I would like to know more.

And to reiterate please stay civil all.


Not sure exactly what I expect but just try to remain polite.

You'd have a team of 17 guys who hit for average, power, and will take a walk while playing great defense. And you'd have eight more guys who average over a K an inning on the mound with great control, keep the ball in the park, and have good endurance. Some might quibble with the hitter/pitcher breakdown, but you get the idea.

[ctorg]

Originally Posted by CBL-Commish

You'd have a team of 17 guys who hit for average, power, and will take a walk while playing great defense. And you'd have eight more guys who average over a K an inning on the mound with great control, keep the ball in the park, and have good endurance. Some might quibble with the hitter/pitcher breakdown, but you get the idea.

I think any team with an unlimited budget would have those things, sabermetrically inclined or not.

I think it's more interesting to talk about what one would do with limited resources. Let's say you have to make choices between some of those things. You can get the guy who strikes out 10 per 9 innings, or you can get the guy who blasts 50 homers a year (a little simplistic, but that's the idea). Which one do you pick? Why?

Note: this is rhetorical and intended only to display the sort of question that I think would lead to an interesting discussion of different team construction strategies.

[wlight1]

Originally Posted by ctorg

I think any team with an unlimited budget would have those things, sabermetrically inclined or not.

Correct me if I'm wrong, but I think that was sort of CBL-C's point. Sabermetrics doesn't represent a different philosophy of baseball, it's just about identifying & describing more accurately how to go about winning games.

* Although it's worth keeping in mind that some baseball notables actually don't believe in batters taking walks (I think "clog up the basepaths" is a famous phrase?).

[NCBeachBum]

Joe Morgan, Dusty Baker, and Mickey Hatcher are all proponents of slower hitters swinging away at bad pitches rather than walking for that very reason.

[RonCo]

Originally Posted by NCBeachBum

I would say the most useful idea GMs such as those two have brought to the forefront isn't really a sabermetric thing, but is probably worth mentioning. That is the targeting of qualities that are being undervalued by the rest of the market and stocking up on it. You can then use your surplus in that area to trade for those overvalued areas in which you have come up short.

Sabermetrics is about finding value that others haven't found.

[BoofBonser26]

Originally Posted by wlight1

* Although it's worth keeping in mind that some baseball notables actually don't believe in batters taking walks (I think "clog up the basepaths" is a famous phrase?).

Originally Posted by NCBeachBum

Joe Morgan, Dusty Baker, and Mickey Hatcher are all proponents of slower hitters swinging away at bad pitches rather than walking for that very reason.

Sure. You'd rather have them make an out.

[wlight1]

Originally Posted by BoofBonser26

Sure. You'd rather have them make an out.

The funny thing is, I think they would.

And not a strikeout, either. Strikeouts are what bad hitters do. They'd rather have the ball put in play for an out, preferably with men on base and fewer than 2 outs.

[NCBeachBum]

Yup. Brilliant, ain't it?

[Syd Thrift]

Originally Posted by BoofBonser26

Sure. You'd rather have them make a productive out.

Fixed, at least to the extent that I can fix the quote but not Dusty Baker's head. FWIW the latest Hardball Times Annual had a study of which managers have historically done a good job of being there when their players did better than normal career progression would suggest, and Baker was near the top. IOW his ideas may seem to suck to us stat-heads but apparently he gets good results with them.

Another guy who sometimes gets maligned by the stat community who did well on that was Lou Piniella. With Lou, all his goodness was concentrated in hitting; pitchers did about average under him. That's actually pretty on par with what I saw while he was the M's manager in the 90s; he'd say some dumb things sometimes but it seemed like everybody and their grandmother improved their walk rate under him. He was just plain brutal with young pitchers. Mike Hargrove was rated about where I thought he'd be, which was near the bottom. The worst manager at this metric was Art Howe.