Going back to the early '80s, the debate about the meaning of minor league stats has progressed significantly. Bill James' first article on this was called the Ken Phelps all-stars, where he looked at AAA players and said that a certain number of them, headlined by Phelps, had shown by their statistical production in the minors (mostly AAA), that they could be above-average players in the big leagues. This is from his original essay, I think in 1982 or so:
"See, on the one hand youâve got the Henry Cottos, and on the other hand, youâve got your Ken Phelpses. If Henry Cotto is a major league ballplayer, Iâm an airplane. Cotto is one of those guys who runs well and throws pretty decent, and one year he hit .270-something (in less than 150 at-bats, in Wrigley Field, with a secondary average of .164), so you get guys like Don Zimmer who will rave about this great young prospect and keep trading for him, so heâll get about eight chances to play in the major leagues before they figure out he canât hit. At first when he doesnât hit theyâll say he just needs more playing time, and then theyâll say he needs to stop wiggling his elbows while the pitcher is in motion, and then theyâll say he needs to point his lead foot and learn to keep his weight back, and then theyâll say he needs to be more aggressive at the plate, and then theyâll say he needs to go back to wiggling his elbows. They always figure that if you can run and throw theyâll teach you to hit. Of course they canât teach anybody to hit, but they always think they can, so they keep trying.
Then on the other hand youâve got your Ken Phelpses. Ken Phelps has been a major league ballplayer since at least 1980, when he hit .294 with 128 walks and a slugging percentage close to .600 at AAA Omaha, a tough park for a hitter. Through 1985 he had 567 at-bats in the major leaguesâone seasonâs worthâwith 40 home runs and 92 RBI. The Mariners still didnât want to let him play. See, the problem was that Chuck Cottier, in his day, was a Henry Cotto, a guy who could run and throw, but couldnât play baseball. Most major league managers were those kind of guys. Ken Phelps, on the other hand, canât run particularly well (though he isnât exceptionally slow, either) and doesnât throw well, and if youâre that kind of player and want to play major league ball youâd better go 7-for-20 in your first week in the majors, or theyâll decide itâs time to take another look at Henry Cotto. Ken Phelps in his first two shots at major league pitching went 3-for-26. Despite his limitations, the man is a major league player. Heâs a major league player because he plays good defense at first base and has a secondary average over .500, so that he can both drive in and score runs."
Phelps stats? 1,854 at-bats, 123 homers and 313 RBIs. He produced 10 WAR over seven semi-regular seasons, and would probably have doubled his number if he had been given more chances, especially early on. His wRC+ for four of his late-20s early 30s seasons was 143, 148, 148, 160. Seattle still did not give him ABs, and barely played him the year after his 143 wRC+ season.
The point I'm trying to make is that the argument that 'AAA stats (when adjusted for level and context) don't project to the major leagues just as well as MLB stats' is over. Several papers have been written showing the deviation is equal, on average, to performance predicted by MLB stats. That what the good projection models like ZIPS and Steamer have shown repeatedly. Cave and Clemens, IF GIVEN THE OPPORTUNITY, are projected by ZIPS to be about 1-1.5 WAR players in about 200-300 ABs. Dahl projects to 0 WAR, and Luplow to about 0.5 WAR.
The Phils see these players about the same way analytics sees them. I would rather have a Nick Maton or Matt Vierling type who might have an upside, but we are fresh out of those guys, barring a trade.
