One way would be to assign "RBI points" to each lineup as it entered the postseason. Something like this:
1 point = average hitter (Avila, career 105 OPS+)
2 points = plus RBI man (Kyle Seager type)
3 points = Victor Martinez on a tear
4 points = Miguel Cabrera, Mike Trout, Giancarlo Stanton, Babe Ruth
-1 points = 75 hitter
-2 points = 50 hitter
..........
That way, three Cabreras and Trouts are going to overweigh any six feebs you put with them. Well, y'know. There is an aspect to "lineup legitimizer" known to baseball people at field level.
It would be tough work and subjective at times, but I'll GUARANTEE you that for the same ERA+, the teams with high RBI points are going to blow away teams with low RBI points.
But I'm not suggesting we abandon Matt's paradigm. Just trying to figure out how to get it to better capture situations like the 2014 A's.
.
The Counselor sez,
Gappiness studies and gap indexes? This is some epic new baseball shtick.
Here's a question: What are the Gap splits on the Orcs during 2014? Did losing His Blingness increase gappiness to an unsustainable level, or did it cause an equally devastating collapse of spirit?
- See more at: http://seattlesportsinsider.com/comment/154940#comment-154940
.
That's the way to unner'stan something. Not by (1) running numbers alone, but by also (2) keeping the hypothesis (Hard RBI up and down the lineup, elite pitching SP's 1-3 and RP 1-3) in mind, and (3) applying it to reductio ad absurdum case studies.
See, the idea -- which would send college sophs screaming into the night -- is to take situations in which we "know" what happened, and bounce our testing off them. As James says, "if your defensive formula calls Johnny Bench a bad catcher, get a different formula."
.....
The Orcs, over the 2014 season in total.... I predict that our little experimental design will fail to capture the problem, even though the problem did* exist in the real world. But let's try it:
| Pos | PA | OPS+ | Black Hole Points |
| C | 661 | 128 | 0 |
| 1B | 657 | 81 | +27 |
| 2B | 606 | 70 | +30 |
| 3B | 716 | 121 | 0 |
| SS | 683 | 97 | +4 |
| LF | 684 | 108 | 0 |
| CF | 726 | 90 | +12 |
| RF | 659 | 106 | 0 |
| DH | 667 | 75 | +28 |
| SP1 | 0 | ||
| SP2 | 0 | ||
| SP3 | 0 | ||
| SP4 | 0 | ||
| SP5: JH, JL, DP, DS | 100- | 0 | |
| RP | 100- | 0 (I think) | |
| TOTAL | +101 |
.
Now, that's the whole year. At the trade deadline, they swapped out the 115 OPS+ Cespedes for pitching - which didn't improve the pitching "gappiness" much - and started using Craig Gentry, Sam Fuld, and Johnny Gomes there.
Per our "raw" first-cut formula, the A's would have gone from a +101 gap team to maybe a +105 or +110 gap team ... But, two interesting observations:
Per the Gap Theory, the 2014 A's were always "visibly" fragile. They had 3.5 gaping holes in their lineup BEFORE dealing Cespedes.
I believe that the Gap Theory should apply much more to offense than to pitching. Pitchers aren't really subject to being "exposed" by scouting reports.
If we had some way to zero in on "hard RBI" we would see the problem more clearly. For tough RBI when the pennant was on the line, the A's had RH Donaldson, LH Moss, and RH Cespedes. When you pulled 1 of their 3 RBI men out from under them, the embarrassment begun.
..........
First thing I would do with Matt's formula, is to more heavily weight gappiness on offense. Then I'd put in a qualifier that if a team has (say) 4 big time RBI men on offense, never mind. :- ) Then I'd weight RP much more heavily than SP 4 and SP 5.
But yeah. We're just trying to draw in the sides of the triangle here.
BABVA,
Dr D
Comments
Hmm. I wonder if there shouldn't almost be a "Regular season gappiness" and a "Post season gappiness". The 4/5 starters for the Giants, for instance, may have hurt their win total in the regular season but had no impact at all in their post season run. Contrast that with the 2001 M's - small gap in their 4/5 pitchers and racked up the regular season wins but that SP depth really didn't come into play in the post season. Interesting work here.
The typical bullpen gets 1800-2000 batters faced. That's THREE pitchers. For example, the 1997 Mariners had a 19 point gap between bullpen OPS+ allowed and average...which was worth 55 gap points. :)
Also for clarity's sake, the As were indeed always fragile and I warned as such many times this year. But I calculate the gappiness difference to be 13-15 points, not 5-8...and the core reason was that Moss was never an RBI man and their 2B was always awful.
That doesn't mean it's wrong...maybe the study I did "sort of" captured something that RBI points would do better at clarifying. But the theory of the gap index was to see whether it was better for a team to be balanced or unbalanced. Two teams that both won 95 games and made the post-season with 100 RDs. One team has 3 Cabreras and 4 giant line-up holes...the other team has 7 guys hitting 105. My theory is that the 7 105s is a better post-season club. The RBI points you propose would say the opposite. The data says it's better not to be gappy...but the question is...is that because gappy teams are also worse on average in total value...or because balanced teams are actually better?
The idea of a "balanced" team is another, important idea. Without a doubt, your 69 points vs 87 points finding was significant.
........
The actual problem faced by the 2014 Mariners -- most their lineup peed their pants against tough pitching when a win would put them in the playoffs -- was an overlapping one, but not the same problem.
What Dave Fleming had originally put forward was more your idea, the idea that "huge glaring flaws" hold back star-studded teams. He stated also that the 1997 Mariners were the first team he thinks of, in relation to this problem.
The 1997 M's had Junior, ARod, Buhner, and Edgar, so that would be a case against my RBI points system.... of course, that team had a weirdly bad bullpen which AGAIN is a confounding variable. A -1,000 points bullpen tends to confuse the findings. :- )
:: daps ::
It's hard to think of a way to design something that could capture this incredible implosion we just saw.
Definitely, your fine method shows that the A's went up in gappiness. And, as you say, you saw them as already "gappy," before the deal. But what might have warned us toward such a Jenga-tower catastrophe, it's hard to imagine...
...if I had to guess anyway. Teams with a big nasty flaw get weeded out...as do teams that seem to have competency in all areas but have players who really aren't good against top competition. If I were to try to get at the "hard RBI" thing, though, I wouldn't use OPS+ and scale things linearly...I'd probably do something like check RBI/runner on base and OPS+ against power pitching and then give non-linearly more credit to the guys that have skills in those areas. This is fun though...I'm all ears for other things I could check. :D
I think it should always be the case that a GM who is winning could be able to look at his roster, see where he is most vulnerable, and add to that vulnerability...not make his strength stronger. Synergy works both ways...you can create a defensive buzz-saw all you want but if 4 line-up slots are gimmes, it won't matter because the other team will just outlast you. See also: the 2010 Mariners.
Who BTW, profile thusly:
C) +38
1B) +41
2B) +23
3B) +48
SS) +35
LF) +20
CF) +20
RF) +6
DH) +41
SP1) 0
SP2) 0
SP3) +2
SP4) +51
SP5) 0
RP) +21
Total Gap Score: +346 (ow)
http://ken.arneson.name/2014/11/10-things-i-believe-about-baseball-witho...
Specifically
#8: Diversity is Good for Batting Lineups
&
#9: A lineup without holes scores runs exponentially, not linearly
But every point, well maybe except #10, is a worthy read.