The first assumption is that baseball talent is normally distributed. It is Gaussian, that is, you get a bell curve when you graph it. Now, baseball talent is not measured directly, like height or weight for example, but our experience of baseball makes this intuitively obvious. When I played Little League or pick-up ball as a kid it was clear there was a great range of talent. Most of us were ordinary, some of us were really bad, and some of us were really good. By the time we were in high school the ballplayers had been "sorted" from the non-ballplayers. Again, the talent distribution was obvious: some good, some bad, most just average. This sorting continues through college and the minors and you wind up with a tiny percentage of the world's population that can play professional baseball.
A "replacement player" is a fringe major leaguer or an accomplished minor-leaguer. These are players that are readily available on the waiver wire, have been DFA'd, or who are good enough to get an invite to Spring Training. They are abundant (relatively) and inexpensive.
A replacement player is NOT an "average" major league player. Such a fellow is NOT cheap or readily available. A guy like Jeff Samardzija is a great example of an average major-leaguer. These are valuable players. Good teams are stuffed with guys like this. There are very few real outliers in the bigs--there are not enough Mookie Betts-types to go around. Average players make up the bulk of the rosters.
In theory a team of replacement players would win about 30% of the time. In fact, the definition of a replacement player is a matter of debate, but it is generally agreed that a win percentage of .250 to .350 would be a reasonable outcome for a team made up of all replacement players. That's 40 to 57 wins. A .300 percentage is a 49-113 record. This is another key assumption behind WAR.
Last year the Giants produced about 24 WAR. Add that to 49 and you get 73 wins. They won 77, but the WAR framework gives us a nice predictor for win totals. The Diamondbacks produced 35 WAR which says (49+35) 84 wins. They won 85. The Dodgers? 51 WAR which is 100 wins (they got 106). The Rockies and Padres rate about 28 WAR and that predicts 77 wins (they got 71 and 70). It's not perfect, but it gives us a thumbnail sketch.
So, what's it all about? A good team needs about 40 WAR (for 89 projected wins), an average team needs 30+ WAR. How many WAR do you think the Giants are good for in 2020?
--M.C.
4 comments:
From FZ:
“I would still expect us to add one or two players to the outfield mix before we get to camp,” Zaidi said. “Whether by trade or free agency, whether it’s the NRI (non-roster invitee) route or the major-league roster, there are still guys out there we think will be good additions for us. That would still be our expectation.”
Rumor has it the Giants are talking to Hunter Pence, who is a free agent. Pence had a .910 OPS (.377 wOBA) in a half season with Texas in 2019. That's about as good as he's ever hit in the bigs.
Here's Baggs (from The Athletic):
This will not be a clubhouse for cranks, malingerers or malcontents. Kapler assembled a mostly unconventional 13-person coaching staff, including several people who never played a game of professional baseball, majors or minors. The Giants plan to experiment and take risks. They’ll need players who approach the game and their careers with an open mind and a positive attitude to embrace change. What better conduit to that clubhouse than Sandoval, a player who is eternally upbeat and always down to do anything?
Not many guys can pull off a "second act" in baseball, but Pablo seems determined to do so. He played well last season. He is one of the few Giants hitters in the histroy of the park who has better numbers at home. We'll see if he can recover from his surgery and be able to play.
Your curve gives us a big picture of possibly the baseball playing universe. The part that is Minor league and Major league continues on the downslope. But, if you perhaps just graphed major league players, you could probably also fit them into a similar type curve. So it is all in how one presents the statistics, obviously as you well know. A good point is made when comparing talent levels from zero to 100. But a curve comparing only the upper echelon players that have made the high minor leagues and the majors, would look quite different than the downslope at the far right of your curve. I think we get it, that even bad major league players, and so called 4A, players are exceptionally talented compared with the general public. But for me, I like to see the curve representing the few hundred skillful players who have made it to the majors. Not a criticism, just an observation.
An interesting observation--I'm glad you picked up on it. There is a lot of talk amongst the saber-istas about such a curve. Most think it will look like the extreme right end of the big bell (the universe, as you say). That is, talent is NOT normally distributed. You have a large number of major-league competent players and an even larger number of marginal/replacement players and a very small cluster of outlier/superstar players. There is no "left-hand side" of the curve, in other words. I don't really follow the arguments, so I can't say much. Bill James, for example, is convinced the curve is not a normal distribution.
I think the key question is how do we measure the talent? I'd be willing to bet that the physical differences between established major leaguers and AAAA-types are very small, perhaps negligible. They can all throw the ball really hard. They can all run and catch hard-hit balls. They can all crush a pitch. They ALL have talent. Maybe a few freakish players will have unusual eyesight or reaction time or some other measurable physical quality, but I'd bet most of the guys on that extreme right-side of the big bell curve are much closer to each other in "raw" ability.
Post a Comment