I got started pondering a point I’d made in passing last week, how well Jack Fisher might have done if he’d been allowed to remain a San Francisco Giant instead of being airlifted cross-country and forced to serve a four-year sentence as a New York Met from 1964 to 1967.
Let’s settle some of the real-world issues right out of the box: Fisher got moved from one coast to the other in a special draft after the 1963 season that appears to have been a "Feeling Sorry" draft. The Powers That Be’ed, I assume, looked at the Mets’ performance in 1962 and 1963 and decided, "Oh, wow! We pissed on the rug big time, making a team in New York as sucky as this team is, we really gave these New York dudes a royal hosing. Let’s have a do-over, or a second shot at drafting players, or something, because this aggression cannot stand!" (I don’t know why I’m imagining these stodgy MLB execs to be talking like they were in The Big Lebowski, but stick with me—I’m going somewhere with this.) So, quite arbitrarily, it seems to me, they decided to pluck Fisher off the Giants roster and put him on the Mets’.
As a 10-year-old Mets fan, I was way cool with this move, but if I were a Giant exec, I might have asked, "Why us? Why Fisher? Why the Mets? If not us, then who? If not now, when?" I mean, the arbitrariness of the "Special" draft seems weird to me. Why should the Giants in particular have to pay the price for the Mets sucking quite as hard they sucked? (For the record, I decided to look up who, if anyone, got placed on Houston’s roster in 1963 via a Special Draft: turns out that’s how Claude Raymond became a Colt .45 and then an Astro. If you like, you can trace out a similar process for the Braves, who lost Raymond this way, but it’s bound to be dicier since Raymond was almost exclusively a reliever in his 12-year MLB career, as well as being one of Paul McCartney’s lesser-known pseudonyms.) Seems pretty ineffective as a remedy for the lousy players made available in late 1961 to the expansion teams, but it also seems very unfair to penalize the Giants and the Braves alone—much better IMO if the NL would have let the Mets and Colt .45s pluck one player off of four NL teams’ roster apiece. Four decent players like Fisher and Raymond could have made a real improvement for each expansion team; one, not so much.
My speculation concerned the potential usefulness of Fisher to the back of the Giants’ rotation in the mid-1960s. In 1964, for example, the Giants’ rotation was a dog’s breakfast, aside from Marichal. Gaylord Perry wasn’t quite Gaylord Perry yet (12-11, mostly in relief), Bob Bolin was okay, but not especially effective (6-9 in 175 IP), Jack Sanford, Billy Pierce, and Billy O’Dell had gotten old (a collective 16-14 in 200-odd IP), Bob Hendley and Ron Herbel were young, collectively below-.500 starters and would remain so, and for some reason they were using veteran starter Bob Shaw as their closer that year. By himself, Fisher wouldn’t have made up the difference between the Giants’ 90 wins and the champion Cardinals’ 93 wins. Or maybe he would have?
As noted in my earlier piece, Fisher was mostly useful to the Mets as an innings-eater from 1964 through 1967. He stayed healthy the whole time, and pitched over 220 innings every season he was on the team. His stats look pretty bad (one year, he lost 24 games, which I think is the last time any pitcher lost that many games in a season) but then of course, he was pitching a lot for the New York Mets when they were a terrible team, offensively and defensively. If he had remained on the Giants, and pitched about the same amount of innings as their fourth or fifth starter, he certainly wouldn’t have put up W-L records like 8-24, and since he also wouldn’t have had the Mets’ terrible defense behind him, he wouldn’t have had ERAs quite as high as he did, either.
Given those obvious points, it’s unarguable that Fisher could have contributed something to the Giants’ pitching staff, which like most teams’ pitching staffs, used untested rookies, washed-up starters, relievers pressed into emergency-starting roles, etc. at the back of their rotation. Without even attempting a serious analysis of what exactly Fisher’s contribution might have been had MLB not plucked him off the Giants’ roster, it’s plain that he would have added something positive to the Giants’ starting rotation.
The intriguing part of this concerns how the Giants did in the years 1964-1967. As I remarked in passing at the end of my "Piece of Piffle" article https://www.billjamesonline.com/a_piece_of_piffle/ "if Fisher had stayed with the Giants and their fabulous offense, some of his 200+ IP seasons with the Mets from 1964-7 would have looked pretty good…I’d like to have seen as him a back-of-the-rotation guy for the mid-1960s Giants, and so would he have." The Giants were known in those years as some kind of legendary team—I wouldn’t be surprised if they led the NL in total wins from 1964-67, or the 1960s as a whole—but they consistently "spent the decade finishing in second place to somebody after somebody," as I put it. This isn’t literally true, but for our purposes here is close enough for jazz. So, I wonder, would Jack Fisher’s competent 220+IP have been enough to push them over the top in all those years? In any of those years?
Remember, this wasn’t a trade, wasn’t a sale, wasn’t even a regular draft. It was basically MLB deciding to steal Fisher from the Giants and give him to the Mets as an act of charity. There’s no reason on earth, other than pity, to rationalize the special draft of Jack Fisher on October 10, 1963.
It’s not as if the Giants had gotten Fisher for free: they’d acquired him ten months earlier in a deal with the Orioles, trading Mike McCormick (1967’s Cy Young winner), Stu Miller (who would finish 19th, 7th, and 11th in the AL’s MVP voting in 1963, 1965 and 1966, respectively) and John Orsino, for Jimmie Coker, Billy Hoeft, and Fisher. Fat Jack was a fairly valuable piece of horseflesh, 24 years old when the Mets got him and already the winner of 36 big league games with Orioles and Giants.
The Giants finished a total of 17.5 games behind between 1964-7, or a little over 4 games behind year, on average. This doesn’t mean that winning five more games per season would have secured four more pennants for them, of course, because that figure fluctuated quite a bit annually, the worst being 1967 when they finished comme d’habitude in second place but 10 ½ games behind the Cardinals—adding another Willie Mays and Juan Marichal might not have helped them edge out the Cardinals in 1967. (They almost did add another Mays and Marichal—they picked up prodigal son and eventual Cy Young winner Mike McCormick from the Senators for approximately nothing, and the next season added rookie Bobby Bonds to their roster.) But that also means that in the other three years of Jack Fisher’s Mets career, 1964-1966, the Giants finished under 2.5 games behind, so without doing any figuring, you’ve got to think that Fisher would have helped significantly in those three seasons.
For example, let’s examine the back of the Giants’ rotation in Fisher’s first year as a Met:
They gave 36 starts to pitchers other than their top five starters (Marichal, Hendley, Bolin, Herbel, Perry) in 1964. Mainly these were relievers who got just a spot start or three, but as starters
Jack Sanford (17 starts) went 5-7 with 39 ER in 106 IP
Billy O’Dell (8 starts) went 0-4 with 32 ER in 33.7 IP
Dick Estelle (6 starts) went 1-2 with 14 ER in 41.7 IP
Jim Duffalo (3 starts) went 1-0 with 5 ER in 18.7 IP
Billy Pierce (1 start) went 1-0 with 1 ER in 7.7 IP
and
Bob Shaw (1 start) went 0-1 with 4 ER in 6 IP.
The total for the back of the SF rotation, then, is 36 starts (8-14, .364) in 219.7 IP. Their collective ERA is 3.89.
This is not too far off Jack Fisher’s actual 1964 stats with the Mets: 34 starts (10-17, .370) in 227.7 IP, 4.23.
The question is "How to translate the stats Fisher produced with a last-place team to the stats he would have produced with a strong contender?" There’s no question that Fisher suffered from having Jim Hickman, Larry Elliot, Rod Kanehl, Joe Christopher, Ed Kranepool (oh, yes, they did!), John Stephenson, and Dick Smith in CF that year instead of Willie Mays, right? In addition to accounting for the difference in the fielding behind him, we can also try to estimate other factors: the difference in the Giants’ potent offense and the Mets’ anemic one, the difference in the home ballparks, maybe the difference between the Mets’ inept bullpen and the Giants’ deep and strong one.
Allowing just one game for each of those four factors gets us up to four extra Fisher victories right there, 14-13 instead of 10-17, and one victory apiece seems conservative. More precise calibrations than this back-of-the-envelope kind of thing is called for, but I’m wondering about the most appropriate methods.
For example, is the ballpark factor really relevant? If Fisher would give up more runs in Candlestick than he did in Shea because of the ballpark, he’s also gaining a few more runs scored by his own offense, isn’t he? I devised a quick’n’dirty way to estimate ballpark factors without doing an extensive study, just using bbref’s park factors. They list four ways of viewing the park’s effect: 1964 pitchers, 1964 hitters, multi-year pitchers, multi-year hitters. Presumably, if we’re interested in studying only 1964 pitchers, such as Fisher, the sample size might be too small to be anything like definitive, so it seems safer to me to take the average of all four factors. In this case of Candlestick in 1964, that comes to about 101, a 1% increase in offense. In the case of Shea Stadium in 1964, that figure is about 97%, or a 3% decrease in offense. If it’s off, it’s not very far off.
Is it necessary? I don’t know. But let’s look at the other factors.
The other factors do seem more directly relevant. The Giants scored 656 runs as opposed to the Mets’ puny 569 runs, a difference of 87 runs. If we knock that down by the 4% difference in home parks, that’s about 3 runs, so let’s call it 84, or a hair over a half a run per game.
Is that an appropriate reduction? After all, teams play only half their games in their home park, so maybe I should cut that 4% in half? Whether it’s half a run per game or quarter of a run per game, it seems to help on about the same order, so we can come back to this one later. We’re just figuring orders of magnitude now.
Let’s look at defense for a minute. Again, having Willie Mays in the center of your outfield seems as relevant as having him in the center of your lineup, to say nothing of having Hal Lanier at 2b and Jim Davenport at 3b and Jose Pagan at ss. (I don’t remember Pagan particularly being a defensive wizard, as I do Lanier and Davenport, though he did have a long career as an MLB shortstop, including playing shortstop for 164 games on the Giants’ 1962 NL championship team and finishing 11th in MVP voting that season while batting .259 with 7 HRs and 47 BB. Lifetime 79 OPS+ in over 4000 plate appearances—I think I’ll go with "defensive standout" until shown otherwise.) It’s safe to say that the Giants’ defense was a YOOOOGE improvement on the Mets’ but let’s try to quantify that a little bit:
The 1964 Mets gave up 776 runs and the Giants allowed 569, a big difference, though most of "defense" is pitching. Shall we estimate that defense is two-thirds pitching, and one-third fielding, which is what we’re looking for here? That general assumption might be safe, but on the individual team-level, not so much. The 207 run difference between the teams might be 69 runs in fielding and 138 runs of pitching, but maybe not. It’s even possible that the Mets had a GREAT-fielding team and that ALL 207 runs are attributable to the pitching alone. Looking at the actual players on both teams, starting with that super-human guy in CF for SF, we can say safely not. So let’s guesstimate the fielding to be worth 69 runs per 162 games, a little less than the offensive difference between the Giants and the Mets. Together they’re worth about a run per game.
Finally, the difference between the bullpens is a factor I’d like to account for. Subjectively, I’ve felt frustrated for many, many Mets starting pitchers over the decade, not least for Jacob DeGrom last year and beginning long before Jack Fisher, at the phenomenon of pitching a strong game and having the bullpen blow it all out of proportion.
The worse the team in general, I think, the worse the bullpen is likely to be. The bullpen difference may be the largest of the four chief areas on teams’ rosters—starting rotation, starting lineup, bench, and bullpen—since the better teams view having a strong bullpen as essential to their success and the weaker teams view it as "Hell, by the eighth inning, we’re too often too far behind for a good pen to matter much." So the better teams tend to glom onto available relievers and the worser teams tend to deal them off in trades. Also, the stronger teams have better pitchers to choose among on the back side of their rotations to press into service as relievers if need be.
That said, the Mets’ bullpen in 1964 wasn’t half bad: I’ve written at length about Bill Wakefield’s spectacular year in the Mets’ pen in 1964, Larry Bearnarth had a fairly decent year, and I’m hereby putting you on notice that I’m working on another article on a third young Mets’ reliever of 1964 named Ron Locke who also had a surprisingly good season. (I’ve spoken to Locke, who has a gig at age 78 collecting tickets for the Red Sox’ spring training site, and arranged for a sit-down interview.) The 1964 Mets had 27 save opportunities and saved 15 of them, blew 12 of them.
The Giants had almost twice as many save opportunities, 50, and converted a slightly higher percentage into saves, 31, with 19 blown saves. The Giants’ bullpen also allowed 28% of all inherited runners to score, while the 1964 Mets allowed 38% of inherited runners to score. Clearly, the Giants were better positioned to preserve late starters’ leads, though not a huge advantage.
So, my initial questions remain: what would Jack Fisher’s 1964 have looked like if he’d been able to play for the Giants as a back-of-the-rotation starter instead of for the Mets as their ace, and would the Giants have won the 1964 pennant with Fisher on their roster?
Since the fourth-place Giants finished 3 games out of the pennant, it seems entirely possible to me that Fisher could have made up the difference.
Rather than attempt to compute this using the figures I’ve compiled, though, I’d like to crowd-source this one. Personally, I’m speculating that if you lop about a half run per game off Fisher’s 4.23 ERA and you give him the average Giants’ runs-per-game of about 4 runs per game, the crudest of measures, that works out to a Pythagorean W-L advantage of about one game over .500. Fisher would have won about 4 more games and lost 4 fewer with the Giants, I figure, and his W-L record in about the 36 starts that the six fill-ins got for the Giants would have been more like 14-13 than the collective 8-14 they pulled.
I’m really asking you what you think the difference would have been. I began with the assumption that the numbers, when crunched properly, would yield the result that, yeah, Fisher woulda had better stats with the Giants. I haven’t crunched them properly yet, because I’m still trying to figure out what the proper values are, whether ballpark factors really matter here, how to quantify the Giants having a better bullpen, how to make sure I’m not double-counting numbers, how to estimate accurately the Giants’ fielding advantage, so for now I’ll go with my original estimate and say that the 1964 pennant race might well have gone to the Giants if they’d been allowed to keep Jack Fisher.
I’m wondering what you think. How would you crunch these numbers? Are there other numbers that you’d want to include? Or to exclude? Would you, for example, reduce the number of IP that Fisher had pitching for the Mets? They had some lousy pitchers, after all, and probably left Fisher in the game sometimes because they didn’t have a lot of better options than a tired, overweight, struggling pitcher who’d already thrown 130 pitches, and the Giants did. Should I somehow allow for the fact that Fisher got some very weird run support from the 1964 Mets? He did, you know—in three of his starts, the Mets scored 46 runs in support of him. In other words, they scored about 8% of their total runs for the season in three of Fisher’s starts. (He got 1 win and 2 NDs out of those three games where the Mets scored 12, 15, and 19 runs for him.) Do you think it’s fair that I’m using the Mets’ total runs for the season to get the figure for their average runs per game, or do you think I should use instead the number of runs they actually scored in Fisher’s starts?
Once we get a methodology worked out, we might move on to 1965 and 1966, when the Giants finished 2 games and 1 ½ games behind the pennant-winning Dodgers. Also when Fisher pitched the most innings of his MLB career, 253.7 and 230 IP, respectively. So I’m asking you for two somewhat different pieces of input here: give me your best eyeball estimate of the number of wins you think Fisher would have added to the Giants’ record in 1964 if he’d been allowed to remain a Giant, based on the numbers I’ve provided or on anything else your heart desires and your mind thinks relevant, and also give me your suggestions for including or excluding both bits of data and factors in devising a methodology to compute accurately any given player’s hypothetical results pitching for a team other than his actual team.
It may well be that I’ve overcomplicated the issue here. Or that I haven’t complicated it enough. I haven’t even tried to account for luck in Fisher’s W-L record. Seems to me that you need some pretty bad luck to lose 24 games, as Fisher did in 1965, but is luck relevant here? Fisher’s stats from 1964 through 1967 seem to me remarkably consistent: he started 34, 36, 33, and 30 games each year, relieved in 3, 4, 4,and 3 others, pitched 227.7, 253.7, 230, and 220.1 innings, gave up 23, 22, 26, and 21 HRs, walked 56, 68, 54, and 64 batters, struck out 115, 116, 127, and 117, and gave up 124, 121, 108, and 121 runs. His W-L records on the other hand ranged from a horrendous 8-24 in 1965 to a pretty fair 11-14 in 1966, pitching just about the same in the two years. So would it be a better methodology to rate Fisher in each season by his effectiveness in the four years overall, or by his results in each of them one by one?
My back-of-the-envelope thinking, that Fisher would have been worth a handful of wins to the Giants, took me about 5 minutes (or five decades, depending on how you look at it) to come up with, and I’m not sure if it’s going to be very far off from the results of a complicated, multi-step, formula worthy of Bill James. Is this just a rationale for laziness, or can you tell me which steps we’d need to take to devise a more precise method?