Seattle Sports Insider

M's +12 over .500 according to Pythag

So why are they Free Fallin'

Posted by

7/27/14

Q. I'm not a geek; I just like baseball. What is "Pythag"?

A. Bill James noticed, in like 1984, that if you ratio a team's runs scored to runs allowed, you can come up with an "expected winning percentage." To James, the algebra of it sort of looked like geometry's Pythagorean Theorem.

Then everybody else tried to claim rights to the formula by making little tweaks to it. In this section, (whoever wrote) the Wikipedia (entry) argues ferociously that the original formula Just. Wasn't. Good. Enough. For Real. Scientists, so you should get the current "expected" standings from them.

I can't imagine how James feels, to have 100+ major discoveries co-opted by people who took his INSIGHTS and made them a teeny bit more precise, and then ... sigh. Voros McCracken built a baseball life around ONE James-level insight, the idea of BABIP. Imagine if everybody tweaked that, and then forgot about McCracken. Were the Wright Brothers more important, or the average Boeing engineer?

But hey. James has been named one of the WORLD's 100 most influential people (?!?!), is inside baseball, etc., so it's not like he didn't reap benefits. :- )

Q. Wow. He invented Pythag in 1984? That's like before computers.

A. I dunno when it really was. Many of James' discoveries did come in the 1980's, even 1970's.

Actually, much of James' early revolution occurred before computers meant anything to the regular person. "Sabermetrics" was often defined by sportswriters as "the computerized use of baseball statistics." Personal computers were kind of cutting-edge in the 1980's.

(James himself says "I can't imagine a worse definition" and defines his invention --sabermetrics -- as "the search for objective knowledge about baseball." In his zeal for that, he'll listen to anything as evidence. Including public perception, sportwriter HOF balloting, advance scouting reports, etc.

He "gets it," that we are not world-class scientists here, writing for the Journal of the American Medical Association. We're just fans, trying to "see" baseball better. We're surprisingly good at it, but the fact that we use algebra --- > doesn't make us scientists.)

Q. Is Pythag reliable?

A. The larger the group of teams you apply it to, yes.

Nobody knows about single teams; a single team might deserve to be +10 or -10 games versus its runs expectation ... bad bullpen or whatever. (Of course, few sabermetricians will concede this point. We have a strong preference that our theories work consistently, like we have a strong preference that our tape measures work consistently.)

Q. The Mariners have lost way more games than Runs Scored & Allowed would expect?

A. Right. They "should" be sitting on 59 wins per b-ref.com. This would have them +3.5 games up for the Wild Card and only 4-5 games out of the AL West lead.

Q. So ... facing the M's lineup is not the inverse of facing the M's pitching staff?

A. Not by a long ways. We realize it feels that way.

They pointed out on TV that Seager has 43 of his 59 RBI in Mariner wins. It seems like if it isn't Seager, Cano or Zunino bailing us out, it's going to take a lucky "soft" RBI to win 2-1.

.......

But against most M's pitchers, things are virtually hopeless.

Forget Felix and WBC-san ... Charlie Furbush against a left hand hitter? Did you realize that Chris Young has the lowest opponent batting average EVER allowed by an M's starter in Safeco? Wilhelmsen and Farquhar as afterthoughts?

The M's ERA+ is 121; its offense has two Stars, and is 91. Theoretically the lineup is nowhere near as bad, as the staff is good.

Your eyes tell you differently, in July? Good for you. You're watchin' the games.

.........

Did you notice that Brandon Maurer has already been "promoted"? He's pitching in tie games and with leads, after only a handful of "reincarnated" game outings. That's quick.

The growing pains seem a hundred years ago :- )

Q. Okay. Does SSI think the M's "should" be -5 wins worse than expectation?

A. Actually -7 wins.

The bullpen is leveraged. James himself found that the teams with the best bullpens will tend to do better vs expectation.

Obviously the bases that are gained and lost -- tie game 7th inning -- count extra. Though none of the formulas account for this.

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

The M's should have SIXTY-ONE wins. According to the formulas we use to calculate these things. Rerun this season 1,000 times via computer simulation, and the M's would tend to be 61-44 right now.

Hey, if there's a PythagenPat (there is), why can't there be a PythagenDoc? Adjust for the bullpen too.

Q. Something sounds fishy about this team "deserving" a 61-44 record. Maybe the Runs Scored + Allowed are themselves lucky for the Mariners?

A. They have a "2nd-order" Pythag that does not use the basic Runs Scored & Allowed -- it adjusts those to "Runs Created" by using walks, doubles, etc. to also project how many RUNS the team SHOULD have gained / lost.

By this method, the M's are also +12 over .500. For example, they have hit for 1310 total bases and pitched for 1209 total bases. The 2014 M's simply gain more bases than they concede. A lot more bases!

Q. Don't tell me there's a "3rd-order" Pythagorean expectation. :: groan ::

A. There is. It adjusts for strength of schedule, in addition to everything else. I agree that 2nd-order is better than 1st-order, that 3rd-order is better than 2nd, and so on. We got a 4th and 5th for ya :- )

The M's are the same +12 over .500 according to this method.

....

Here are the best 3 teams in the American League, per 3rd-Order Wins:

Oakland = 1
LA = 2
Seattle = 3
Everybody else = worse than those three teams. Including Detroit

Let me read that section again.

Q. What's the key takeaway?

A. Bear in mind that Fangraphs, over the next 5-7 days, will be basing its Trade Analyses on who "should" try to compete for the pennant. Here's an awesome example of how this one factor completely reverses a conclusion.

A person analyzing the M's trades this week should be aware that --- > the M's have the #3 team in the American League. In theory.

Q. :: sigh :: Are there really 4th- and 5th-Order Pythagorean Wins.

A. Yes. That's to go ask SSI what is really going on.

;- ) That's an old Steve Mann joke. You do realize we're 70% tongue-in-cheek here.

NEXT --- >

Blog:

Sabermetrics

Comments

Coupla things

The Safeco Field effect is still not fully understood and appreciated, nor may it ever be. The ballpark itself may depress Pythagorean expectations, by demoralizing the home team offense over 81 games.
Also, the M's have been in a tremendously high number of 1-run ballgames, or so it seems. (When was the last time the M's were involved in a blowout?) If you accept that the winners of those games are determined basically by a roll of the die, then doing better or worse than pythag shouldn't surprise; they are basically leaving the outcomes of those games to chance.
Considering those things, I wouldn't expect the Pythag for the M's to improve much, or 'regress to the mean', at least not necessarily.

RBI's in wins

The Seager "RBI's in Wins" statistic isn't that outstanding for a good player on a team with a winning record. Basically it simply means that teams score more runs in wins than they do in losses. That's about it.
Seager has 43 of 59 RBI's in wins. 43 of 59 is 73%.
Trout has 57 of his 76 RBI's in wins. 75%
Miggy has 66 of his 81 in wins. 81%
Donaldson has 67 of his 73 in wins 92%
OK, those guy's teams have won a bit more than the M's, but you get the idea. The vast majority of a teams 1st or 2nd best player's RBI's will come in wins. Or any players, for the most part. When team scores 0 or 1 runs they don't win and there aren't many RBI's. When they score 7 or 8 runs they do win and there are lots of RBI's. There are more RBI's to go around in wins and your best players get their share. Almost all players have more RBI's in wins than in losses. Ichiro has 14 RBI's this year, 10 (71%) are in wins. Matt Dominguez leads the lowly Astros, who don't win much, with 43 RBI's. 30 (70%) are in wins.
It applies to almost all offensive statistics.
Over their careers:
Seager OPS's .816 in wins and .718 in losses.
Trout is 1.022 and .869
Miggy is 1.123 and .781
Donaldson is .899 and .649
Ichiro is .876 and .667
Russel Wilson has a better passing efficiency in wins. Jerry West, Michael Jordon and Walt Frazier shot a better percentage in wins. You get it......
And to give you an idea of how badly the M's overall offensive performance has been, since July 1 we've played 22 games. We've scored more than 3 runs in only 6 of them. We've scored exactly 3 runs 3 times. In only 41% of our games do we score more than 2 runs.
Which makes the lack of urgency to get a COF bat even more ridiculous.

.5 (point 5) order wins...

Sorry for the tongue-in-cheek subject title, but hey, when something occurs to you sometimes you just run with it...
I'm not sure that all 1-run games are the same. If you are an average or healthy offensive team, you are most often going to win a 1-run game with offensively healthy offensive play. Oh sure, sometimes you'll win with small ball, but you don't RELY on the ability to manufacture runs to win. You might win on a walk, a sac bunt, a wild pitch or stolen base, and a sac fly. But more likely you will win on a base hit and a double to the gap, or three singles, or a home run. Why? Because you CAN, and you often DO those things. But a poor offensive team does those things so infrequently that they simply CANNOT count on doing them with any regularity. They don't lose EVERY 1-run game, because if you try sometimes you just might find you get what you need (with apologies to the Rolling Stones). But failing the occasional runs scored by normal offense, you are FORCED to rely on small ball. Those teams that are accomplished in small ball can indeed win more than their share of 1-run games if they have lockdown pitching. But if you have lockdown pitching, and you are NOT accomplished in small ball, then by definition the vast majority of you losses WILL BE 1-run games. The nature of the beast demands it.
One other note about poor offensive teams. There will be a number of occasions when they actually get their tying or winning run only to see the game slip from their grasp with the opponent's next at-bat, assuming their opponent is a healthy offensive team. The converse is not true, simply because your team scores runs far less frequently than the average offensive team.
One final point. A poor offensive team is going to get shut out much more frequently. Even if you hold the other team to 1 run you lose. You CANNOT win a game if you don't score. You can pitch your be-heebers (don't ask) off for 15 innings, but if you don't score you will inevitably lose that game. The team that scores runs, while they may have more trouble preventing runs, almost never has to say that. If they score a single run, they have at least a chance to win.
I still contend the M's are not the kind of team that will win more than it's share of low-scoring 1-run games, because they simply do not have the chops to consistently (emphasis on this word) manufacture runs at a rate that will sustain their outstanding pitching. Neither do most teams these days. But most teams don't have to rely on this as much as the M's do.

Question is

Question is, can an offense drop below a certain level of competency and, by doing so, ruin Pythag?

An Illuminating Response, Griz.

You said in one sentence what it took me a mini-novel to say. Yup, that's about par for the course for me...

First point I agree with GL

I think that is part of what DID HAPPEN in 2011. But you stated it with beautiful economy. What, did you spend an hour working on that paragraph? ;- )

M's are 13-19 in 1-run games … same as Pythag deviation … but

In theory, the team with 20 great relievers is the team constructed to win 1-run games. Which is precisely why James says great bullpens should skew Pythag to the positive.
My question is whether the "lack of hard RBI," as it were, is trumping even that. Good stuff GL.

Point cheerfully conceded … with one asterisk...

You betcha Mo'. When we passed on that Seager stat, that was our first question -- how much is that true of, say, Josh Donaldson. Thanks for doing the research.
........
Problemo is, right now the M's have 2-3 players compiling those stats, as your last couple of paragraphs imply ...

Thumbs up

I knew you were just passing on some TV info that had value....or not.
When I heard that last night I almost followed up with some data. Your morning post just got me off of my duff....
Thanks...
And you're exactly rught. The point remains that we were 2 tough RBI bats short of a load. Miller has shown some ability to be .5 of a tought RBI..and Hart may have been at one point (no longer).
Morales is one. Now get the other Jack.
And the Skipper needs to quit batting Jones at the top of the lineup, too.

No Ichiros or Rickeys here

It has always been Dr. D's thesis that --- > Star leadoff men / table setters (Ichiro, Rickey, now Jose Altuve) grow in importance as the games tend towards pitchers' duels.
There is an analogy in chess: Bishop vs Knight -- the Knight is relatively important with 32 pieces on, the Bishop more important with 10 pieces on, and there are many such "continuums" in chess.
The Mariners are obviously the OPPOSITE of that, so your point has traction DaddyO. Good stuff.
..........
Nobody has ever agreed with Dr. D that Rickey can skew Pythag too, but you just did :- )

I would like to see that studied

Sabes would tell you No, but that's kind of what the original article was supposed to be about. :- )
Hard RBI ... manufactured runs ... hitting "pitchers' pitches" in 1-1 ballgames. The M's offense is terrible at all of that stuff.

Yes-indeed-ee

Ah, yes. As an OBSERVER, not a sabermetrician, I think a Rickey Henderson is a major chip in any run-scoring environment (his OBP), but a crucial chip in a low-scoring environment. It is not coincidental to me that the M's are OBP-challenged. The mid-60's Dodgers had Rickey-lite (Maury Wills, OBP about .330-ish) and other team speed as well (WIllie Davis (OBP challenged), but they also had some under-appreciated elements. Ron Fairly got on base a LOT, Wes Parker also tended to have very good OBP. Lou Johnson had speed. And most ALL of them were excellent bunters, I mean REALLY good. And most were adept at the hit and run even if, like Parker, they struck out a lot for non-power hitters.
All this is the long way around to say that the M's are not constructed to be a small-ball team like the mid-60's Dodgers. They just don't have the inherent skillsets. James Jones has the speed and basestealing, but not (yet?) the consistentcy at getting on base. The M's in general are TERRIBLE baserunners, not because they get thrown out sometimes (the Dodgers did OFTEN-- look at Maury WIlls' caught stealing percentage), but because they run into bonehead outs so often. The Dodgers would get thrown out one inning taking a legitimate chance, then do it again next inning and make it...and it would create the run that beat you.