THE BOOK--Playing The Percentages In Baseball

I agree with the overall premise that Ichiro is not an elite player (and probably not anywhere close to being one). In general, despite a lot of overblown talk about who the “most well-rounded player” is, fans tend to get impressed by players who do one thing very well. People are impressed because Ichiro gets a ton of hits. He’s been in the top 2 every season of his career. That’s a result of a combination of a high batting average (very important to a lot of fans), playing virtually every day (ditto), and not walking much. He also steals a lot of bases, appears to be a very good defender, is Japanese, and goes by his first name.

One point of contention - according to UZR, Ichiro is merely an average rightfielder. But, according to Dewan’s system, he was the best RF in baseball from 2004-2006, worth 59 plays above average. I’m not taking any position about which system is correct, but that makes a pretty big difference in his overall value (even if he’d still be short of “superstar” with an extra win on defense).

Seems reasonable.  Hard to believe the defensive numbers but that only because we hear non-stop about how great he is.  I actually haven’t seen him play much. 

His arm is very strong though, right?  Do you have a link where I can read more about the arm methodology?  Is that based on actual runners thrown out/extra bases prevented?

Anyway, I’ll probably get myself in more trouble with this (like in the MVP thread) but to me superstar doesn’t have much to do with ability.  Surely they’re correlated, but it has a lot more to do with fame.

I would call Ichiro a superstar.  Jeter too.  However, take someone better than them who just isn’t as famous.  I don’t know, like Markakis - better player, but not a superstar.  Know what I mean?

Basically I think endorsement dollars correlate better with being a superstar than on-field performance does.

Regarding his arm, my analysis at Hardball Times (click on my name) finds that he’s solidly above average in arm, but not the top of the heap as his reputation would suggest.

I guess he’s around +5 runs above average per season, something like that.  If you go to Ichiro’s stat page at THT, you’ll see his arm results for 2004-2007.

Dewan’s +/- in CF for 06-08 has him right around average.  In RF, from 03-08, he’s number 1, which looks like he’s a +10 or +12 run player according to Dewan’s numbers.

These two results are consistent, since there’s a 10-run gap between CF and RF.

Let’s also give him a bit of a bump for his arm and for being able to cut down on extra base hits with his fantastic speed.  Just +0.2 wins or so.

***

MGL’s “average RF is 1.0 to 1.5 wins” is the thing I was railing against a little earlier.  The adjustment for a RF is -0.75 wins, and not -1.0 to -1.5 wins (it’s a -0.25 wins adjustment as a CF).  It is irrelevant how much talent is at RF, just like it is irrelevant what the mean talent is at SS in high school ball.

***

Ichiro was a monster as a hitter (for him) in 2004, which MGL doesn’t show.  It was 4 years ago which drops the relevancy, but I wanted to point that out.

His WPA/LI, which accounts for whatever performance he has had with men on base, his GB outs moving runners over, etc, etc, is +3.3 wins over the last 4 years.  Dude averages 740 PA per year (!!!!), and so, per 150 G (i.e., per 650 PA), is +0.7 wins.  This includes his basestealing, but not his baserunning.  With some tiny accounting for his missing 2004 season, and his baserunning, let’s make him +1 win as an offensive player.

***

Let’s see what we have so far:
offense: +1.0 wins
defense: +0.5 wins (that’s for glove, arm, cutting extra base hits, and positional adjustment)

So, he’s a +1.5 win above average player.

That was based on 2005-08, meaning that we need to apply some age adjustment, even though he looks like he won’t age at all.

He is now a +1 win player, per 150 G.

Since his rookie year, the fewest number of games he has played is 157 games.  His fewest number of PA is 728.  This guy is a machine.  He seems to not break down and not age.  His SB/CS numbers in the last two years is 80 SB and 12 CS.  That’s incredible right?

I normally don’t go over 90% playing time for any player, but I can certainly see an exception here.  Let’s leave him at 90%. 

So, he’s +1 win per 150 G.

The replacement level, in the AL, is +2.3 wins per 150 G.  (It’s 1.85 in the NL)

Add it up and we have a +3.3 WAR player making him a star player.  Perhaps not superstar, but the level below that.

I think MGL was overly harsh on two fronts:
1. relying on his fielding without regressing those sample stats to players of his skillset

2. the level of talent to compare against

I am going to have to continue to side against MGL on this issue.

And if I used the Fans Scouting Report, he’d be a 4 WAR player.

Where you rank Ichiro pretty much depends on how you value speed and defense (and of course just where you make the star/superstar distinction). Speed and defense are two areas that we do not measure with the same confidence level as hitting.

I believe that for a few seasons Ichiro was a superstar, meaning (for me) an MVP-caliber player. Not that he *should* have been MVP, but that he was a serious candidate. He is past his prime now and not as valuable.

Let me ask you, was Pete Rose a superstar? Ichiro as a hitter is very similar- high BA, not really a HR threat, doesn’t walk enough to be the perfect leadoff man, but pretty darn good in the role. Ichiro has better fielding skills that Rose did, and is faster than Pete was. Rose was more valuable as a hitter, but the difference isn’t huge.

Basically, UZR hates Ichiro - always has, always will most likely.  Why? No idea.  Is it right? I don’t think so, but it could be.  Since +/- disagrees, it’s probably just a data thing, since UZR and +/- are similar in structure. 

But here’s the thing - the people who think Ichiro is a star are basing it on the belief that a huge portion of his value comes from his defense.  So, really, the entirety of any disagreement about his value will come down to that portion.  If you think he’s a below average defender, then yes, he’s an average player.  If you think he’s a really good defender, then yes, he’s an all-star. 

There’s evidence on both sides.  It really just depends on how much you trust UZR.

The star/superstar line is really irrelevant since MGL and I have each provided quantified measurements.

There really isn’t much disagreement between MGL and I in terms of evaluating Ichiro’s numbers.  But, we have a disagreement as to how evaluate the position/league context for him.

To say that a RF is up to +1.5 wins above average, implying that an average RF would get paid the same amount as an average 2B, is the thing I rail against weekly here.

If the average DH was +0.5 wins as a hitter (and zero as a “fielder"), is that the baseline you compare him against?

You might as well compare a SP against the average starter and a RP against an average reliever.  It means nothing until you can align the average starter and average reliever the same.

Speed and defense are two areas that we do not measure with the same confidence level as hitting.

Actually speed is quite easy to measure and put a number to.  For one thing, we don’t have much noise in evaluating a player’s speed.  Watch him run a couple of times and you know how fast he is.  For another, baserunning value is easily measured (how often he advances or gets thrown out on hits and other batted balls).  And of course basestealing value is even easier. How many runs or wins a basestealer adds to his team’s RE or WE when he steals a base and how much he costs his team when he gets thrown out. Pretty simple.  Most analysts, including myself, have concluded (with some degree of certainty) that a basestealing threat (on first base) adds nothing to the batter, despite what you hear from the talking heads.

Arm is also easy to measure.  When a ball is hit to an outfielder, given the location in which the ball was fielded and the park, how often does the runner advance the extra base or not or get thrown out.  Simple.

I agree with Tango about the “value” thing at each position.  Tango, the average salaries at each position (adjusted for age or service time I guess) match up with your relative values?  I have it stuck in my head that an average player at each position is worth the same.

Yes, of course, the overall value for Ichiro depends on what defensive metric you believe.  I go with UZR for obvious reasons, but I give him the benefit of the doubt (quite a bit of benefit actually) based on the other metrics and the Fan Scouting Reports.

Again, I am talking about a current projection, not what he has done in 2004 (+27 lwts per 150) or any other year.  While he may have a different aging curve than the average player, I don’t believe for a second that he is not aging.  And yes, his projected playing time makes him more valuable of course.

I think the biggest discrepancy between the public and the analysts, in terms of Ichiro’s value, is his offense.  There is NO doubt that his current offensive projection is a little above average.  That alone would get snickers and stares from most average fans and the media (what about his BA, number of hits, his “consistency,” the pressure he puts on the defense, etc.?).

The batting runs for Ichiro for 2005-07 are lower than I see from other sources.  Total of +22 runs.

B-Ref has +31 runs, +6, +22, +3.

I have +32, +20, +18, though mine include sb/cs.  Back that out and I get +29, +14, +16.  These are park adjusted, custom baseruns where the total will add up exactly to the team run total.

Anyway, here are my numbers for Ichiro’s 2007 season:

Bat (inc SB/cs) +31.5
bsr +3.0
avoid DP +3.2
ROE -2.9
TZ defense -4.5
arm 6.2
Position +4.2

Total 40.7 above average

He was +30 for 2005 and 2006.

I’ll have to double check my offense numbers.  There shouldn’t be that much difference since I think we’re doing the same thing (LWts).

Be careful,or Dave at USS Mariner will
claim you said that because he is Japanese. lol
By the way,terrific data;kudos!

I suspect the difference between MGL and B-Ref is the lessor value of infield singles.  For most players it probably won’t make a difference, but if he hits 60 and they are worth .05 runs less (for example, not sure what the right figures are) that’s 3 runs per year.

Between mine and B-ref, it’s probably the difference between using custom LW vs static run values, and a lower negative value for strikeouts.

An IFS is worth around .08 runs less than a non-IFS. I think Ichiro gets around 20 more per 500 PA than the average batter, for a difference of only around 1.6 runs per year, not a whole lot. It might be in the park factors, I don’t know.  My numbers are also adjusted for the quality of the pitching that Ichiro faces.  I don’t know how much effect that has.