Finding the right comps for Joe Shlabotnik

Doc…Delgado is NOTHING like Teixeira.

. :- ) :shaking head:

=== Slow Down A Minnit, Johnny ===

When comparing two hitters, there are any number of criteria you can use to match them:

1 RH/LH 2 Walk rate 3 HR rate 4 Phenotype 5 AVG/OBP/SLG 6 Speed score 7 etc etc etc

Which of these you believe to be most important is a matter of judgment. Your judgment can be tested; we'll talk about that in a moment. ................ In comp'ing Teixeira, I'm looking first and foremost for these things:

1 Lefthandedness 2 Extreme talent (130-140 OPS+ demonstrated level of performance) 3 High OBP (.400) 4 High HR's (30+ at least) 5 Good BB/K ratio 6 Big guy with natural power

The rest is negotiable. But the MOST important thing is ALWAYS how GOOD the comp is. Aging depends first and foremost on whether you're good at age 30. Players are skiing down a slope. How long they stay at high levels, depends on how high they started! You don't want to start with a hitter's G/F ratio and use that before you use OPS+. You'll wind up comparing Brad Wilkinson to Jim Thome. Start by asking, "who was at this hitter's level?". Give me every ballplayer who was truly MVP-caliber at the plate, and that pool of players all by itself will be an excellent comp pool for many purposes. Show me everybody who had three straight years of 140+ production, and that group will have a LOT of things in common... ................... Now, my criteria could be right or they could be wrong, but to say that Mark Teixeira has "NOTHING!!" in common with Carlos Delgado is .... weird, and it doesn't take the discussion in a constructive direction. I've listed what he has in common with Delgado. Where do we take the discussion, if you're going to claim that Mark Teixeira does not have 30+ homers a year in common with Carlos Delgado? Suppose you say that he doesn't share the attribute of having two nostrils, and I post internet pics, and you still disagree? What do I do then? :- ) Tex has ALL of my six main criteria in common with Delgado: he's a lefty masher, with natural power, who controls the strike zone beautifully. Of course you can find things that don't match. You might then say, "Personally, I feel that the most important attributes are X, Y, and Z, and therefore don't accept the Delgado comp." . === Let's Check It Out Dept. === Are my criteria the right ones? Here's how you tell, in a first cut at it. Let's acknowledge that there DOES EXIST such a thing as a pool of players with high HR's, high OBP, lefty swings, 210 lbs. plus, have proven 130 OPS+ production, etc etc. ...whether this is the RIGHT pool isn't the point. That pool of players exists. And you can measure them. And Mark Teixeira is a member of that pool, by definition. I'm claiming that if you go measure this pool of players, they don't fall off the table at no blinkin' 32 years of age. Go check it. .............. SABRMatt can, and no doubt will :- ) propose his own set of criteria: a certain G/F rate, and switch hitter, and white rather than black, and whatever. Mark Teixeira will be a member of this pool, also, by definition. Matt can then go measure this group, but if he's not careful, it will include a bunch of guys like Pete Rose and Julio Cruz, etc etc. But he can measure this group and see how it ages. ............. Both groups will be (semi-)accurate representations of a guy like Mark Teixeira. If Teixeira is a member of all three groups, Group Yellow, Group Orange, and Group Red .... and EVERY member of group yellow achieved at least a 100 OPS+, and EVERY orange got at least a 110 OPS+, and EVERY Red got a 120 OPS+, which one forms Teixeira's lower bound, seeing he's a member of all three? . === No Cooking the Books, Now, Dept. === Caveat: it is true that you don't want to set the lower bounds of Teixeira's comp-group right at Teixeria's own performance. In other words, you don't want to take a guy with a 140 OPS+, and comp him to ten guys with OPS+ of 140 to 170. The broadcasters do this: "Dante Bichette is a member of a very select group with .300 AVG's and 100 RBI a year." Bichette winds up getting comp'ed to a whole lot of guys who are better than he is. ................ But we're not doing that with Teixeira. Teixeira's OPS+ level is 150, hard on the barrelhead. We comped him to guys 130-160. Teixeira's OBP the last two years has been .400 and .410. Carlos Delgado's was .383. etc etc. You compare Mark Teixeira to a group of players, half of which are a little better and half of which are a little worse, and you're going to have quite a group of stars, amigos. Go out and find all the big lefty 30-homer guys who ran .400 OBP's and who were as good as Teixeira, kiddies. You're not going to find them collapsing at age 32. Cheers, Dr D

Printer-friendly version

Categories:

Other

Replies

But I'm not comping Tex to

But I'm not comping Tex to Wilkerson. I'm choosing between all of those players who have similar skills to Tex...I'm choosing the ones who are of similar batting type. Delgado isn't one of them. He's similarly skilled and left handed...but after that, they don't look alike at all.

A word on player comps

A word on player comps ...

There is an inherent WEAKNESS in loading up on multiple similarities between players. (huh?!?)

The weakness is that it SHRINKS the available pool of potential matches. If you can find 20 guys that match a given player across 8-10 metrics, maybe you've got enough to make an "informed" decision on.

If you take one stat (OPS+), and start there - and come up with 80 players ...
And then take another stat - (leftyness) - and that gets you down to 30 players ...
And then you take another stat - (K/BB) - and that gets you down to 10 players ...

Going any further COULD actually make your assessment LESS accurate.

I cannot say how many players are "necessary" to have a "good" comp -- but most player-to-player comps are more illustrative than predictive. This is because it is possible to generate very similar numbers, but still have some major underlying differences.

My assessment of the age-based decline is *NOT* based on attempting to find a large pool of similar players. It is based on the general knowledge that *ALL*, except a VERY small pool of elite players tend to decline at about that time. This allows me to go and find similar players (stat-wise) with very little effort. Mind you, if I use the general rule, and then have difficulty finding comps - THEN, I'll raise an eyebrow and ask myself, "What about this type of player makes him less *HUMAN* than everyone else?" Because the reason a player like McGriff or Mo Vaughn drops off in production soon after they turn 30 is *NOT* because they are lefty or righty beefy sluggers who like mangoes. It is becaue they are HUMAN -- and humans in practically every sport ever played have shown this tendency.

Now, if someone WANTS to do a study to find 50 similar players and create a composite career arc trend for Tex-types, I'm all ears. But, I would then also want to hear the theory as to why this type of hitter manages to avoid the genetic fate of the rest of mankind.

++ There is an inherent

++
There is an inherent WEAKNESS in loading up on multiple similarities between players. (huh?!?)

The weakness is that it SHRINKS the available pool of potential matches. If you can find 20 guys that match a given player across 8-10 metrics, maybe you’ve got enough to make an “informed” decision on.
++

Agree 100% amigo. That's the catch-22.

If you could find 25 guys you believed were really good matches to the 28-year-old Teixeira, that would be awesome, but usually you (or at least, I) can't.

.................

The theory as to why this type of hitter manages to avoid the genetic fate of the rest of mankind? Is that this type of hitter is a lot better than the rest of mankind. :- )

All hitters age, of course. The better they are at their peak, the longer it takes them to fall off, generally speaking.

FWIW, I try to come up with

FWIW, I try to come up with my own comps, and try to make the criteria as few and as fundamental as possible. But you've got to have at least a few things like these:

Is the guy "natural"? How well does the phenotype match?

Is he lefthanded or righthanded?

Does he control the strike zone?

How good is he -- 100 OPS+ or 140 OPS+? Does his arc take a similar shape?

Does his AVG/OBP/SLG profile look at least similar? Does he have "old player skills" or young player skills?

.................

Oddly, when you look at b-ref.com's comps, usually about 50% of them are VERY good matches as to the above. When you look at PECOTA's, they don't seem to be as good.

You guys see Teixeira's comps as being relatively modest. Personally I don't think we should undersell this guy. He's been tremendous -- playing in small cities, for losers.

.................

The whole thing is moot. Come to find out that Tex is looking at $24-27M a year, and that's not Mariner territory.

This is true Sandy...but

This is true Sandy...but that's why you shouldn't build you player comps as either "yes he is one" or "no he is not"...you should create a statistical measure of similarity (one that is context adjusted unlike B_Ref's lame attempt)...and weight your total sample of guys who are at all similar by how similar each one is. I do this with the Chi Squared goodness of fit test.

Bottom line. Carlos Delgado will be on that list of guys who gets a weight...but he's not going to be leading the comp pool...specifically BECAUSE he's a different TYPE of hitter. Mark Teixeira's OBP is much more heavily forced by BA (the first thing to go with power hitters as they age) - Delgado is more of a stalker and his OBP is more walk-driven (lasts longer with old guys). Teixeira's PX can't compare to Delgado's. Tex's SLG is also forced by a higher BA...Delgado has more raw power. Not that Tex is a weakling...just that he's not naturally INCREDIBLY powerful...he generates his power with a good quick level swing...not with massive opposite-field explosive power. You need to be able to see that difference to understand why I think McGriff is a better comp for Tex...and why I think looking at Delgado is a mistake. Delgado's performances were driven by things that take longer to lose with age...his swing is quick, but it's also tremendously powerful and long. That doesn't disappear with age. Tex's power demands that he continue to be insanely quick and maintain perfect mechanics. As soon as he loses a couple milliseconds in reaction time and through-stroke speed drops even a smidge, and his power is going to drop.

I think you have to group

I think you have to group player in relevance to skill sets.

Find out batter's with similar production, Contact rates%, power, BB/K, LD%, FB%, etc., etc. and find their best comps.

Intuitively it doesn't seem like Tex is one of those guys thats going to age particularly well. His contact rates are average, power isn't "overwhelming", neither is eye ratio, and much of his value is in his defense (which goes away quickly for 1B).

Agreed Taro...that's all I've

Agreed Taro...that's all I've been trying to say.

Tex has good power but not great.

Tex has a solid LD% and BABIP, but hardly amazing.

Tex is a good fielder but not a top notch gold glover. Just very solid.

Tex has a decent K/BB, but it's not like Thome's.

Tex walks some, but not like Giambi (who BTW was a way better player than Tex at this age and was half as good three years later).

Face it...Tex's value is based on the assumption that he can continue to be good at everything, great at nothing. He's maximizing his skills. The instant something goes wrong...the instant he loses a tick on his bat speed...the instant he starts to get beat by fastballs up...he is going to lose half of his offensive value just like Fred McGriff.

I often take a gander at the

I often take a gander at the b-ref comps. Of course, MANY of their comps make me just laugh. But, the very fact that on the majority of lists, the guys that lead the comp list at age 24, 25 are almost never the guys that match comp-wise at 30, 31, demonstrates their limited usefulness as a "predictive" tool. Individual player careers are far too dynamic.

Mind you, I use comps -- and consider them "suggestive", not predictive. But my first instinct is to trust the "deep" data pools more than the shallow ones.

Seattle Sports Insider

You are here