THE BOOK--Playing The Percentages In Baseball

The first two are not on the same level of importance as Tango’s suggestions, but:

* There’s a difference between a park’s effect on home runs and its effect on runs scored

* If you insist on using RBI, please look at runs scored too. 

most importantly:

* Offensive production must be evaluated in the context of outs.

writers tend to not be the worst offenders of this stuff, but i really wish broadcasters had a better understanding of sample size. a solid percentage of the stats show during a game are pretty meaningless. a pitcher’s numbers against a particular team (where the roster isn’t even constant), against a particular batter over 8 PA.

Patriot’s last point is also key. i really think the entire perception of getting on base would be different if outs made were listed as a counting stat. too often the idea is that an out is just a missed opportunity to add something, rather than the actual hit to your team’s RE or WE that it is.

It’s kinda difficult to come up with wishes that are more important than those; well, at least the first two, like Patriot says.

- Errors/fielding% are not a good way to evaluate a player’s defensive prowess. But I feel like we should tell the managers and coaches that, since they are the ones voting on Gold Gloves.

- Please stop describing players as clutch as if it has any predictive value at all.

- A player’s gross stolen base total is not as important as his SB%. I think many writers realize this, though.

Well, the third stat has to be position or position-adjusted defense for batters (position is a crude measurement of defensive value but it’s really easy to understand and measure, and we have it for hitters at every level).

Playing off Patriot’s, my third stat would be:
R+RBI-HR

1) There is no such thing as an “RBI man” as a skill. It’s a function of overall hitting skill and the guys ahead of the RBI man.

2) Good (not Hoffman/Rivera level, but just good)relief pitchers are a dime a dozen.

3) The free pass is almost always a bad idea.

Personally, on batting average, I wouldn’t be opposed to making OPS to BOPS, which compensates partly for the undervaluing of OBP in the equation but is still able to be tallied in one’s head in a couple of seconds.

* You are not a psychologist

@3: Nobody actually evaluates fielders by fielding percentage, and certainly not the Gold Glove voters. It’s a better stat than raw range factor—it really does tell you (how often a fielder makes errors) what it purports to do.

[1] Sample Size - Seriously, a batter being 2-for4 against a pitcher does NOT mean he “has had his number”.

[2] “Luck” is often used in place of “we don’t know yet” or “we don’t have a complete understanding”. I can just see this coming. When the media/broadcasters get a hold of BABIP, every pitcher with either be lucky or unlucky, and it really will be that simple. *Shrugs*

[3] I want the media to understand the value of defense* and position scarcity. In other words, give Chase Utley his due.

* Provided that we are certain that we are measuring defense accurately and translating it into an accurate number of defensive runs.

All the stuff about pitcher Wins and RBIs seem like mountains out of molehills.

@7: Oh, that’s good. And there’s no such thing as a hot streak, and a cold streak is still probably bad luck as well.

For the third offensive number you could use some kind of net steal number. Perhaps weighting CS to emphasize not getting caught.

charles,

don’t some of the individual skills of overall hitting lead to being an “RBI man” though? i think there is such a thing as an RBI man - heard theo epstein say that drew will never be one because he doesn’t put the ball in play a ton and walks a bunch. the important point is that the skills that lead to high RBI totals are not the only ones that lead to total team run scoring.

yes, to those who worship the RBI i try to get them to look at it as a % based on opportunity, but doesn’t even THIS comparison bias toward certain types of players (and not necessarily the best offensive players)?

Oh, I change my mind, and vote my Phylan/7 as #1.  Easily.  Brilliant.  My god, you just removed 50% of articles on that suggestion alone.

And, yes, I also agree, showing split or streak data that uses 20 or 30 PA: that’s a big no-no.

Forget my sh!t.  You guys are much better at this.

@8 I don’t know, it seems to be a factor.

Rolen had the 2nd highest fielding% in his league (and Polanco was hurt for a while).

Phillips and Cano had the highest (except for Eckstein, who zero errors—maybe that destroys my argument).

Jeter had the highest in the AL—and look back at all the past years. Tulo had the highest in the NL.

Pujols and Teixeira had the highest (ok Kotchman was higher, but 35 less games).

Longoria is the one outlier.

Although, it’s probably easier to just chalk it up to reputation in most cases.

Everyone did a better job than me anyhow.

"don’t some of the individual skills of overall hitting lead to being an “RBI man” though? i think there is such a thing as an RBI man”

Yes, it’s called an elite power hitter.

While many probably don;t want to hear it, but if you take a 45-HR Ryan Howard and put him in SEA’s horrible lineup, hitting cleanup, he probably gets 70 RBIs off HRs alone (assuming half are solo shots.

Guys with a lot of extra bases are going to be RBI guys just because so many of their hits score runners, even those on 1st base.

The point has already been made in regards to guys that will take a walk with men in scoring position. But, IIRC that was an issue way back in the Musial-Williams debate .... Musial would expand the zone to drive in a run, Williams wouldn;t swing at a pitch an 1/8 of an inch off the plate and take a walk. I don’t know whether any of that is actually true, I only bring it up because this subject goes back at least to the 1940s.

I don’t understand the fervor against RBIs. Some folks act like there’s 10 Tony Bautista’s every year, or if Ryan Howard only gets 140+ RBIs because he has the bases juiced every time up.

Whenever I see someone like Mike Pagliarulo leads the decade in RBIs instead of a future Hall of famer, then I’ll accept that RBIs do not tell us anything about a player’s performance.

Single season examples, as exceptions, for me, do not ruin the general trend or statement.

1) Ignore RBIs in evaluating hitters (especially for MVP voting).

2) Past performance by a hitter against a specific pitcher tells you nothing.

3) Today’s baseball players are the best who ever lived, and a LOT better than the All Time Greats we learned about as kids.  (Corollary:  expansion has not “diluted” talent in the game.)

Maybe you should hold a vote after we generate some more ideas?

If a baserunner is caught stealing with batter X up, and afterwards batter X hits a home run (or anything that would have possibly scored the runner), you cannot say that 2 runs should have scord (or whatever).

1) The basics of the run expectancy table and its inferences (e.g. bunting is rarely a productive activity)

2) Hot and cold streaks are descriptive, not predictive

3) That there is a way to consider a players’ overall contribution to the team, the WAR framework (the specific implementation isn’t really that important—just have them understand the framework)

And if I had a 4th:
4) Counting stats are meaningless without context.  At minimum, tell me how many PA a guy had if you’re going to talk about his HR, RBI, SB, etc.

Regarding the use of AVG on the scoreboard, I think we sometimes forget to consider the purpose when discussing stats.  If we’re trying to evaluate a hitter’s overall worth/production, AVG doesn’t help us much once we have OBP & SLG.  But I don’t think that’s necessarily the reasoning behind what goes up there.  Rather, it’s to set a context for what events are likely to happen.  That context is the backbone of a fan interacts mentally with the game being watched.

In that context, AVG actually is meaningful.  As a fan watching the game, a hit and a walk are very different outcomes in terms of excitement/enjoyment.  Sure a .400 OBP guy is a .400 OBP to a large extent.  But knowing if he’s a .330/.400 or a .250/.400 completely colors the way I watch him hit, the expectations I’ll hold and judge what I’m observing against.  The same can be said about strikeouts vs other types of outs.  While similar from a value perspective, those events are very different experientially.  A ball in play gives me a jolt of excitement even whens it finds a fielders glove.  A strikeout, by contrast, is notable for denying me that jolt.  Same outcome, very different experience.

I’d like to see AVG/BB%/ISO.  Of course, I’d also like to the Frankenstein creation “at bat” stat die a quick, immediate death.  Rather, I’d use “hitting average” (H/PA) and walking average BB/PA which can be added together to get OBP.  You could then change ISO to Bases/Hit and you’d have a really nice trilogy of stats that summarize the guy’s ability at the plate.

I wish baseball writers would realize that having a closer mentality is a bunch of hooey and the only thing that makes certain people cut out to be closers is the ability to be extremely effective in short outings. The length and color of one’s beard is no indicator of a predisposition for 9th inning effectiveness.

My saber heresy in this discussion is I don’t have a big problem with BA remaining part of the discussion.  High-average hitters who don’t have real value are fairly scarce these days, and low-average hitters who do have value almost always have power stats that get them appreciated.  Obviously we want OBP to be part of the discussion, and often these days it is.  But hitting for average is a real skill and I still care who wins the batting title (even though I know I shouldn’t).

Reliever ‘record’ counting stats are all pretty useless - saves, wins, losses, holds, blown saves - since the rules on how to get them are so absurd and, to some extent, dependent on the tendencies of the official scorer.

Actually, I guess that works for starters too…

If I had three wishes from a saber genie, I wouldn’t waste them on the writers.  There are data sets I’d love to see…

My biggest pet peeves - hitter-batter matchup data and hot/cold streaks dictating who should be in the lineup - are really just a subset of not understanding sample size, though I’ll add the following caveat:  if you have additional information that might help explain a seemingly odd result in a small sample size, give it to me.  E.g., if the Phillies have fed McCann a steady diet of high fastballs--and other teams haven’t--and it appears McCann is having trouble handling them, then maybe I’d like to know McCann only has 1 hit in his last 6 games vs the Phils.  Or if Heyward is 2-for-25 since he jammed his thumb sliding into second and came up wincing, tell me if you think he’s hurt.  Even then I know I’ll still have to wade through a lot of meaningless noise when you try to fit a story to data, but I can handle it.

RBI’s taken together with RBI% is a more meaningful stat than RBI alone.  But the only time the scoreboard changes is with a RS and 95% of the time that is result of an RBI.

Also, rate stats can hide a players lack of playing time or the fact that he is benefiting from a platoon advantage.

OBP is fine, but walks don’t always result in a run, and the value of a walk is also dependent on batting order and teammates.  For example, you have a decent hitter batting 7th, with 2 very weak hitters behind him, and there is a man on 1st and 3rd with 2 outs, a walk has little value.  A walk at the top of the order by non sluggers is more valuable, and walks are more valuable on teams that have deep lineups and not so much on teams like the Mariners.

Ichiro has put up a 372 OBP over the past 2 years. Despite the relatively high OBP and great speed, he has scored only 81 runs per year with 700 PA.  His OBP is not much value to the Mariners since they don’t have guys who can drive him in.

"OBP and SLG with men on base instead to get a better indicator (than RBI)”

Alas, it ignores the beauty of a HR with nobody on base, especially valuable in low scoring games when base runners are hard to come by.

It also says nothing about how many runs ACTUALLY scored with ROB as a result of that OBP.
A walk is less likely to result in a RBI than a hit.
Not all events that lead to a hitter getting on base are equal.

Totally agree with Guy in #21.

It’s not that BA tells you much about quality once you know OBP/SLG, but it’s a great descriptive stat.  Ichiro and Nick Swisher might have similar OBP/SLG, knowing the difference in average tells you that Ichiro is going to try and put the first good pitch into play while Nick is going to work the count until he gets a pitch to drive.

#8 - I did a piece before 2009 at FanGraphs using Jeter as an example of how the Gold Gloves usually go to the fielders with the best out rates on groundballs, pretty much ignoring range.
http://www.fangraphs.com/blogs/index.php/the-great-derek-jeter-conspiracy/

There’s been much talk this year about Jeter leading the league in Fld%, but if you add in infield hits (and the difference is very subjective, trending towards hits) Jeter was only 22 of 30 in 2010, when he’s usually been in the top 5 in getting outs on the balls he’s reached.