"What is the evidence, and what does it mean?" Bill JamesThese articles were mailed to the Blue Jays' mailing list in June, 1994.
To: jays@hivnet.ubc.ca Subject: Olerud, Ward, Cox Date: Tue, 21 Jun 1994 16:58:06 -0400 From: Stephen Tomlinson
A lot of you won't believe this, but it's true anyway. John Olerud's decline in performance this year has actually cost the Jays more than losing Ward or Cox (though not the sum of the two).
Also, the overall decline in the Jays' hitting has cost the Jays' one more win than losing both Ward and Cox:
With those 9 wins, the Jays would be first in the East by one game, and just 0.5 behind Cleveland.
For the bullpen, I'm comparing Ward and Cox to the *worst* Jays' relievers (as measured by total runs (i.e. earned runs + unearned runs) allowed per nine innings) summed up until I get the number of innings pitched projected for Ward and Cox. I inflate Ward's and Cox's runs_allowed by 15% to allow for the increased hitting in the league this year (5.46 vs 4.75 runs scored per nine innings), though it wouldn't matter much if I left that adjustment out. It works out that Ward and Cox would allow roughly 3 runs per nine innings instead of the 9(!) runs per nine innings allowed by the worst Jays' relievers. It works out that Ward would have saved 21 runs and Cox 19 runs, for a total of 40 runs, or roughly 4 wins:
Toronto'93 W L ERA TRA RPG RUN IP BA OBP SLG HR K/9 IP/G
Cox 7 6 3.12 3.33 3.68 31 83.2 .234 .299 .360 7? 9.0 1.9
DWard 2 3 2.11 2.11 2.45 17 72.2 .191 .263 .294 5? 12.0 1.0
Toronto'94 W L ERA TRA RPG RUN IP BA OBP SLG HR K/9 IP/G
0.5*Brow 6.75 ~11 14.0
Timlin 0 0 8.02 8.44 8.61 20 21.1 .330 .430 .538 4 10.1 1.0
Cadaret 0 1 5.85 6.75 7.91 15 20.0 .296 .418 .510 4 6.8 1.0
Spoljaric 0 1 30.86 38.57 31.14 10 2.1 .429 .678 1.24 3 7.7 1.2
StClaire 0 0 18.00 18.00 11.00 4 2.0 .412 .512 .522 0 9.0 1.0
Small 0 0 9.00 9.00 17.84 2 2.0 .467 .551 .846 1 0.0 2.0
9.05 62 61.2
Ward and Cox through 63 games (i.e. multiply by 63/162):
TRA RUN IP
Cox 3.33 ~12 32.2
Ward 2.11 ~6 28.1
Cox+Ward 2.66 18 61.0
adjust +15% (5.46 vs 4.75 runs scored per nine innings in AL'94 vs AL'93)
Cox 3.83 ~14 32.2
Ward 2.42 ~8 28.1
3.25 22 61.0
The Jays have scored 321 runs and allowed 339 through 63 games this year. That suggests a winning percentage of 321^2/(321^2+339^2)=.473, which is close to their actual winning percentage of .476.
Saving 40 runs:
team record changes from 321/339=>.473=>30-33 (same W-L as actual)
to 321/299=>.535=>34-29 (1 game out of wildcard)
Ward by himself: 321/318=>.505=>32-31
Cox by himself: 321/320=>.502=>32-31
Overall the Jays' would be allowing 4.81 runs per nine innings, which would be 5th in the league, same ranking as last year.
Some people will say that they *know* that Ward would have saved more than just two more games, and that Cox would have mattered in more than just two games. Some people will say the method is flawed because it doesn't take into account that Ward and Cox would be used at the most important times, while the worst Jays' relievers were used at the least important times. I don't agree though. Keep in mind that Ward and Cox actually *lost* 9 games last year, which projects to 3.5 losses by this time this year, and they probably would have lost more this year because the hitting is worse. Ward and Cox never were perfect. Also, even when you have good relievers, there are still times when you can't use them in important situations, and the other relievers would have blown some of those same leads that we remember. And the relievers we've had actually have saved lots of leads. So it probably really is true that having Ward and Cox would only have meant two more wins each.
"Losing" Olerud has actually cost us slightly more than Ward or Cox. I've projected his performance of last year through the same number of plate appearances as last year, then added 15% to estimate this year's run production:
Toronto'94 BA OBP SLG RPG ERP RUN RBI PA 2B 3B HR BB/9 K/9
Olerud .290 .379 .442 5.94 36 20 33 256 19 0 5 5.2 4.8
Toronto'93 BA OBP SLG RPG ERP RUN RBI PA 2B 3B HR BB/9 K/9
256/665 Olerud .363 .472 .599 10.28 53 42 41 256 21 1 9 8.4 4.8
+15% Olerud 11.82 61 256
Overall, Olerud has produced 25 runs fewer than expected. With last year's Olerud, we'd expect to have scored 346 runs this year, and with the same 339 runs allowed, that works out to a .510 winning percentage, a 32-31 record, and just 3 games out of the wildcard spot. The Jays' would have scored 5.56 runs per nine innings, which would rank 7th in the league, slightly above average, but still worse than their 2nd place ranking of last year.
The overall lack of hitting has cost us more than Ward and Cox combined. Last year's Jays scored 847 runs, which would be 329 through 63 games this year, and 379 after adding 15%. So last year's Jays would have scored 58 runs more than this year's. That alone would be good enough for .556 pct, a 35-28 record, and a tie for the wildcard spot. Last year's offense would be producing 6.09 per nine innings, 2nd in the league, just like last year.
If we assume last year's hitting, plus Ward, plus Cox, we get 379 runs scored and 299 runs allowed, for .616 pct, 39-24 record, one game ahead of the Yanks and just 0.5 behind Cleveland.
If we just assume that we had Olerud, Ward and Cox of last year, we get a .572 pct, 36-27, one game ahead of the present wildcard teams.
Summary:
offense'93+ward'93+cox'93: 0.616, +9.0, 39-24 (up 1 on yanks) | 90-72
olerud'93+ward'93+cox'93: 0.572, +6.2, 36-27 (up 1 on wildcards) | 85-77
offense'93: 0.556, +5.2, 35-28 (tied with wildcards) | 84-78
ward'93+cox'93: 0.535, +3.9, 34-29 (1 out of wildcard) | 82-80
olerud'93: 0.510, +2.3, 32-31 (3 out of wildcard) | 79-83
ward'93: 0.505, +2.0, 32-31 | 79-83
cox'93: 0.502, +1.8, 32-31 | 79-83
reality'94: 0.473, +0.0, 30-33 (5 out of wildcard) | 76-86
(Stats through games of June 15)
The last column above is the discouraging one. It takes today's actual record (31-36, not the 30-33 I stuck with above for consistency), and looks at what the Jays' record would be if the various players of last year suddenly regained last year's form. For example, if the Jays got back last year's hitting and Ward and Cox, they'd play at .616 the rest of the year, good enough for 90 wins and a wildcard spot (on current trends, the wildcard team will win 89 games). Anything less than that ideal, though, and the Jays come up short of a wildcard spot (we'd be 4 out of the wildcard spot even if all three of Olerud, Ward and Cox returned). Getting back Ward and Cox (today) would probably just mean an 82-80 record at the end of the year, unless Olerud and the rest of the offense came back. Just getting back Ward would probably leave the Jays under .500, like lots of other teams with star relievers. Of course, it's possible that the Jays could just get hot and be in first place a couple of weeks from now. :-)
-Stephen
To: jays@hivnet.ubc.ca Subject: Re: Olerud, Ward, Cox Date: Wed, 22 Jun 1994 15:59:13 -0400 From: Stephen Tomlinson
Q: Is run production really up 15% this year?
It's the actual percentage. As of last week, the average AL team scored 5.46 runs per nine innings, as opposed to 4.75 last year, a 14.9% increase. (With the up-to-date data, it's now 5.42 runs per nine innings, a 14.1% increase over last year.)
Q: Don't teams with good bullpens do better than their pythagorean projection?
Maybe. Last year the Jays' 847/742 suggested a 92-70 record, which they overshot by 3 with the best bullpen ERA in the league. However, that would only project to another 1.3 wins so far this year over what the formula would predict given Ward'93 and Cox'93. i.e. it would suggest the bullpen decline has cost us 5 wins instead of 4.
In 1987, the Jays' had a good bullpen (Henke, Eichhorn, Musselman), but their 845/655 implied a 101-61 record, which they missed by 5 games (quite unfortunately). So a good bullpen certainly doesn't guarantee outdoing the formula. Luck is probably a bigger factor overall.
Remember that other teams typically have a good pitcher or two in their bullpen which they can use in important situations against the Jays, just as the Jays try to do to other teams. So having a better bullpen is probably just a small edge over what the formula would predict, and probably a smaller factor than luck.
Q: Ward blew just 6 saves in all of '93, but this year's pen has more blown saves already.
Last year the Jays had 50 saves and *19* blown saves, for a .725 pct (source: Bill Mazeroski's Baseball'94, p. 34). The .725 pct was still 3rd best in the league. I think we tend to forget how often even a good bullpen blows saves. It's not a new phenomenon this year!
Last year's pen would project to 17 saves and 6 blown this year, after taking into account 15% fewer save opportunities from the hitting being worse. Since the Jays actually have 10 saves, maybe the pen has blown 7 more saves than last year's would have. A couple of those 7 blown saves might still have ended up as Jays' wins though, for a bullpen difference of 5 wins over last year.
Last year's bullpen had an ERA of 3.30, which projects to a TRA of 3.63, or 4.14 with the 14.1% inflation this year. So with the Jays' 5.00 runs scored per nine this year, that's a .593 bullpen, or 8 wins, 6 losses. This year's pen has roughly 4 wins, 10 losses, or 4 wins fewer than you would expect last year's to have.
Whatever methodology I use, whether it's pythagorean, looking directly at blown saves, or looking directly at bullpen W-L records, I get roughly a 4-5 win difference over last year. That's all. I really think that's all it is. Maybe 2-3 wins if we had Ward and 2 more if we had Cox. 4-5 more wins would give us a 35-33 or 36-32 record. The lack of hitting (from 2nd in the league last year to 12th(!) now) would explain the other missing 5 wins.
-Stephen
To: jays@hivnet.ubc.ca Subject: Jays' Mediocre Hitting Explained Date: Sat, 11 Jun 1994 12:20:55 -0400 From: Stephen Tomlinson
I was curious as to what the source of the Jays' hitting problems were, so I took a look. Here are the results. Keep in mind that that an average team is producing 5.49 runs per nine innings this year in the American League, compared to 4.75 last year.
The Jays' hitting has fallen off a lot compared to the rest of the league. Most of the drop is because Olerud is just an average first-baseman this year. We're also worse at shortstop, catcher, and third-base, though some of that has been made up by the improvement in left-field and by Carter:
RPG vsAL (vs93) | RPG vsAL
DH: Molitor'94: 8.1, +2.6 (-0.4) | 1B: Olerud'93: 10.7, +5.9
2B: Alomar'94: 7.7, +2.2 (-0.7) | DH: Molitor'93: 7.8, +3.0
RF: Carter'94: 7.7, +2.2 (+1.5) | 2B: Alomar'93: 7.7, +2.9
1B: Olerud'94: 6.6, +1.1 (-4.8) | CF: White'93: 5.6, +0.8
LF: Delgado/Huff'94: 6.5, +1.0 (+2.3) | RF: Carter'93: 5.4, +0.7
CF: White'94: 6.5, +1.0 (+0.2) | 3B: Sprague'93: 4.1, -0.7
3B: Sprague'94: 3.8, -1.7 (-1.0) | SS: Fern/Schof'93: 4.0, -0.7
C: Borders/Knorr'94: 2.8, -2.7 (-1.7) | C: Borders/Knorr'93: 3.8, -1.0
SS: Schof/Gonz'94: 2.8, -2.7 (-2.0) | LF: Assorted'93: 3.5, -1.3
--- ---- ---- | --- ----
Hitters'94: 5.8 +0.3 (-0.8) | Hitters'93: 5.8 +1.1
Relative to the league, the starting pitching actually looks to be *better* than last year (hard to believe). Stottlemyre is much better than last year, Leiter is better than Morris/Leiter of last year, and Hentgen has improved too. Stewart is average again, and just Guzman is worse (though by a lot):
RPG vsAL (vs93) | RPG vsAL
Hentgen'94: 4.2, -1.3 (-0.9) | Hentgen'93: 4.3, -0.4
Stottlemyre'94: 4.3, -1.2 (-1.9) | Guzman'93: 4.4, -0.4
Leiter'94: 5.2, -0.3 (-1.5) | Stewart'93: 4.8, 0.0
Stewart'94: 5.4, -0.1 (-0.1) | Stottlemyre'93: 5.5, +0.7
Guzman'94: 6.9, +1.4 (+1.8) | Morris/Leiter'93: 6.0, +1.2
--- ---- ---- | --- ----
Starters'94: 5.2 -0.3 (-0.5) | Starters'93: 5.0 +0.2
Yes, the bullpen is worse than last year. Hall and Brow haven't been as good as Ward and Cox. Timlin is worse than last year. Cadaret was much worse than Eichhorn. Only Williams and Castillo have been better than last year:
RPG vsAL (vs93) | RPG vsAL
Castillo'94: 3.9, -1.6 (-0.7) | Ward'93: 2.4, -2.4
Hall'94: 4.0, -1.5 (+0.9) | Cox'93: 3.7, -1.1
Williams'94: 5.0, -0.5 (-1.4) | Castillo'93: 3.9, -0.9
Brow'94: 5.3, -0.2 (+0.9) | Eichhorn'93: 4.6, -0.2
Timlin'94: 7.4, +1.9 (+1.0) | Timlin'93: 5.7, +0.9
Cadaret'94: 8.0, +2.5 (+2.7) | Williams'93: 5.7, +0.9
--- ---- ---- | --- ----
Bullpen'94: 5.6, +0.1 (+0.6) | Bullpen'93: 4.3, -0.5
Notes: I've left out some of the minor players, which probably makes the Jays' averages look better than they are. I weighted pitchers equally, not by innings pitched. For starters, RPG is their earned run average but with unearned runs included. For relievers, RPG is based on estimating the on-base average and slugging percentage against them from stats like hits, walks, home runs (actually for 93, I don't even have home runs). So take the numbers as being just rough estimates. They're based on stats through June 8/94.
-Stephen
Access count for this page:
Last Updated: 1997 Mar 31
Comments are welcome at comments@stephent.com.