Historical League Observations

[pstrickert]

I created a 1974 quickstart with these changes: I removed Rudy May from CAL and placed him on the NYY roster. I gave Bill Singer of CAL a CEI. I created three Mike Marshall "clones" and added them to the LAD roster. Then I simmed the season a few times. The changes hit CAL pretty hard. They no longer seem like much of a division title contender. LA improved its team ERA, predictably enough. But CIN still put up a fight for the division title each time. These changes made a noticeable difference.

Another observation: As I said before, the teams with the lowest ERA in real life still do not put up the same low ERA when I sim the '74 season. I noticed that each of these teams -- LA, ATL, OAK, and BAL -- also had the lowest DefEff ratings (see baseball-reference.com). So I'm guessing (again) that defense doesn't have enough influence on the outcome of games. Of course, this assumes that the OOTP engine assigns the proper defensive ratings to players. (Garlon's working on that problem, I think.)

[PhillieFever]

Instead of using ERA,try using RA and see what you get,sometimes errors skew what you're seeing,especially when your dealing with different personnel.The best thing to use would be something like Tango's FIP which takes fielding totally out of the equation and concentrates on just the pitchers

[pstrickert]

I'll definitely need some help figuring this one out.

In the most recent 1974 season sim, CAL finished ahead of OAK with the best record in baseball. (I left Rudy May and Bill Singer on the CAL roster, BTW.) Instead of looking at CAL (which I'm tired of!), I decided to look closer at OAK. IRL, OAK hit the second most HR in the AL and gave up the FEWEST HR in the AL. In this sim, OAK hit the third most HR, but gave up the MOST HR in the AL. Huh?! This could be why the OAK staff had a 3.66 ERA in the sim (compared to 2.95 IRL). (This, BTW, happens invariably when I sim the 1974 season -- the OAK team ERA is much higher than IRL.)

Here are some AL stats:

Sim (vs. Real)

BA: .257 (.258)
SLG: .378 (.371)
ERA: 3.73 (3.62)
BABIP: .284 (.284)
HR: 1529 (1369)
AB: 67090 (66044)

The LTM for HRs will need to be modified a little. Even so, how do you explain these individual results?

Sim (vs. Real)

Holtzman

IP: 318 (255)
HR: 19 (14)
BABIP: .278 (.293)
MOV: 158
GB%: 56

Hunter

IP: 312 (318)
HR: 42 (25)
BABIP: .270 (.238)
MOV: 124
GB%: 46

Blue

IP: 296 (282)
HR: 32 (17)
BABIP: .288 (.266)
MOV: 152
GB%: 53

Abbott

IP: 180 (96)
HR: 22 (4)
BABIP: .240 (.266)
MOV: 145
GB%: 49

Fingers

IP: 51 (119)
HR: 5 (5)
BABIP: .308 (.285)
MOV: 157
GB%: 64

NOTE: Hamilton, one of the OAK starters, was traded early in the season in the sim.

[pstrickert]

I ran another quick sim (i.e., Sim 2), this time lowering HR totals by 10%. California (this time without May or Singer) took the AL West by two games over OAK. The OAK staff still had an ERA a half-run higher than IRL. This time, the HRA numbers were better -- though not the fewest in the AL as IRL. Take a look at Catfish Hunter:

Sim 1 - Sim 2 - (vs. Real)

IP: 312 - 306 - (318)
HR: 42 - 25 - (25)
BABIP: .270 - .298 - (.238)
MOV: 124 - 140
GB%: 46 - 60

Should his MOV rating be that much different from sim to sim? How about that GB%? Is that cosmetic? Or does it help explain the difference in HR totals from sim to sim? I'm kind of disappointed in these results.

[pstrickert]

One final remark for tonight. Catfish Hunter won the Cy Young Award in 1974. He had the stats to back it up. In nearly every '74 sim, however, Hunter is nowhere close to replicating his real life stats. ERA, WHIP, BABIP -- not even close!

[swampdragon]

Originally Posted by pstrickert

One final remark for tonight. Catfish Hunter won the Cy Young Award in 1974. He had the stats to back it up. In nearly every '74 sim, however, Hunter is nowhere close to replicating his real life stats. ERA, WHIP, BABIP -- not even close!

The DIPS theory is that BABIP is mostly a combination of luck and the defense behind the pitcher. If you follow that theory, and you look at the BABIP of the other A's pitchers you listed, you come up with two reasons for the discrepancy:

1. Hunter was lucky in real life that year. If you look around your replay, you'll find other pitchers well below their BABIP totals. Those are the ones who got lucky in your replay.

2. All of the A's have high BABIPs, so the game is probably underrating one or more defenders.

[PhillieFever]

Originally Posted by swampdragon

The DIPS theory is that BABIP is mostly a combination of luck and the defense behind the pitcher. If you follow that theory, and you look at the BABIP of the other A's pitchers you listed, you come up with two reasons for the discrepancy:

1. Hunter was lucky in real life that year. If you look around your replay, you'll find other pitchers well below their BABIP totals. Those are the ones who got lucky in your replay.

2. All of the A's have high BABIPs, so the game is probably underrating one or more defenders.

This is what I was getting at with my post earlier,DIPS are a much better indicator of how a pitcher is performing because it takes out all of the other factors including luck,fielding and park effects.According to a study done by Tom Tango,Erik Allen,and Arvis Hsu,luck is the #1 determiner of the fate of a ball in play,hjere's the percentages:

1.Luck 44%
2.Pitcher 28%
3.Defense 17%
4.Park 11%

I'd be very interested to see how Clay Dreslough's DICE formula or Tango's FIP formula would match up.

FIP=13(HR)+3(BB)-2(K's)/IP+3.20
DICE= 3.00+13(HR)+3(BB+HBP)-2K/IP

[PhillieFever]

Here's the real life DICE statistics for Hunter,Holtzman and Blue

Catfish Hunter-3.59
Ken Holtzman-3.43
Vida Blue-3.60

[pstrickert]

Originally Posted by swampdragon

The DIPS theory is that BABIP is mostly a combination of luck and the defense behind the pitcher. If you follow that theory, and you look at the BABIP of the other A's pitchers you listed, you come up with two reasons for the discrepancy:

1. Hunter was lucky in real life that year. If you look around your replay, you'll find other pitchers well below their BABIP totals. Those are the ones who got lucky in your replay.

2. All of the A's have high BABIPs, so the game is probably underrating one or more defenders.

As I've said, I've simmed the 1974 season dozens of times. How can Hunter have bad luck every time? (If it wasn't for bad luck, he'd have no luck at all.) The A's staff, too, simply does not perform up to expectations. (And what about the fluctuating GB% from one import to the next?) I'm inclined to think that #2 (above) is the primary problem.

Thanks, guys. I appreciate the replies.

[swampdragon]

Originally Posted by pstrickert

As I've said, I've simmed the 1974 season dozens of times. How can Hunter have bad luck every time? (If it wasn't for bad luck, he'd have no luck at all.) The A's staff, too, simply does not perform up to expectations. (And what about the fluctuating GB% from one import to the next?) I'm inclined to think that #2 (above) is the primary problem.

Thanks, guys. I appreciate the replies.

Is the AI doing something funky like playing a player at an unrated position? That's one of its favorite tricks, and it doesn't help pitching stats much. I remember having a very hard time convincing the AI that Larry Lajoie should be playing 2B instead of SS the last time I replayed 1901.

[Garlon]

If you use the neutralized stats, GB% will not be random for pitchers. We calculated estimated groundouts and flyouts against each pitcher each season and added it to the DB.

Defense and defensiv ratings are a bit imperfect in OOTP9, so that can throw things off a bit. We should be looking better with OOTP10 in this area.

[PhillieFever]

Originally Posted by pstrickert

No matter what I do, the CAL and LA problems persist. Some more observations:

Mike Marshall pitched for more than 200 innings in relief for LA in 1974. When I sim, he's barely able to reach 50 IP. So, in my most recent sim, the Dodgers compiled a team ERA about 1 run per game higher than IRL. Marshall had an ERA of 2.55 in 40+ IP. Unfortunately, the rest of the relief corps did not pick up the slack. The team relief ERA in the sim was nearly 5.00. Yikes! No wonder LA has trouble replicating their success when I sim the '74 season. One more thing: The relief pitchers -- aside from Marshall -- converted only 4 of 20-something save opportunities.

Here's another thing that may help explain the problems you were having in 74',when you use a one year recalc,the modifiers for that particular season will not work correctly.Maybe if you tried a 2 year recalc you would get different results?