Username:	Password:

Remember me

Forgot your username/password?|Not a member? Join now $3/Month

Print Email

Home>Hey Bill

Hey!

Welcome to Hey Bill, where Bill answers questions from his subscribers almost every day. Visitors can read the most recent Hey Bill's on this page. Subscribers can ask Bill a question directly and also view our archive of questions and answers.

The fifteen most recent questions are listed here and will change almost every day.

Hey Bill

FAQ Categories

15 Most Recent Questions

In your recent articles about starting pitchers over time I saw a brief reference to Dick Ellsworth 1963. I remember when you first published Win Shares I was startled to see that his 1963 season ranked even with or even just above Koufax 1963. I know that park effects was a big factor in the ranking, but to me that season is still a stunner. Have you ever written about that one in a similar vein to your discussion of Willie Mays/Duke Snider (195-) or Will Clark/Kevin Mitchell (1989), or as a fluke season a la Norm Cash 1961?

Asked by: joeashp

Answered: 9/24/2021

Oh, I've written about that season many times; it is kind of a hobby horse for me. Here is the line of thought. . .I have traced all this before, but. ..

Koufax was 25-5 with a 1.88 ERA; Ellsworth was 22-10 with a 2.10 ERA, I think, so we start with the assumption that Koufax was better, although Ellsworth season is still atually above the level of a typical Cy Young Award winner from the years 1956-1961. But Koufax is better.

But then we remember: Koufax was on a 99-win team, 99-63, Ellsworth was on an 82-win team, 82-80. Ellsworth won a larger percentage of his team's games than Koufax did, and. . .in the modern world we don't really trust won-lost records anyway, so is that actually a point for Koufax, or not?

And the ERA; Koufax had a 1.88 ERA, but in a park with a Park Factor of 84, whereas Ellsworth is in a park with a Park Factor of 109. Adjusting the ERA for the park, Ellsworth is. . . I'm not sure if he is ahead or behind, but he's not more than a point or two behind, if he is. So is Koufax REALLY better.

The argument is that as we get more sophisiticated in evaluating the stats, Ellsworth pulls up to even with Koufax.

But then we continue to get more sophisiticated, and then Koufax starts to move back ahead. Ellsworth gave up more UN-earned runs; they count, too. Koufax has 120 more strikeouts or something, and he has fewer walks. Ellsworth's advantage is on balls in play. The results of balls in play are not really in the control of the pitcher; they are things that happen TO him more than they are things that he DOES. So then Koufax is back ahead.

Or is he? Koufax gave up 18 home runs; Ellsworth, in a far, far, far better home run park, gave up only 14 (Park Home Run factors, 63 and 118.) Koufax got only 9 double plays; Ellsworth got 21. You can't legitimately write that off as just something that happened to them.

Koufax was supported by 4.30 runs per start; Ellswoth, in a much higher-run context, was supported by only 4.00. Ellsworth, although he hit only .096, was still a better hitter than Koufax.

And then there is this: that Koufax was REALLY that good; that is who he was. Ellsworth was a guy having a good year, which is not EXACTLY the same thing. You're having a good year, it's luck; if you can do it every year, it's skill.

You can never REALLY reach the end of it. There is always something else that could be put on the scales, and when the scales are so near to balance, anything that you add on could tip the scales. I have been fascinated by this since I was 13 years old, as I have been with Norm Cash and stuff, and I mmore or less think that Koufax was greater and deserved the Cy Young. But they are on the same general level.

For your readers interested in developing a pitch count estimator, I laid the foundation for such a model right here. As you've said in the past, "I can't do all this by myself". This is a great little project for the aspiring saberist: http://www.tangotiger.net/pitchCountEstimator.html

Asked by: tangotiger

Answered: 9/24/2021

Thanks.

HeyBill, to the extent BAbip speaks to bad fortune, Bellinger was at .305 in 2019, and this year is clipping along at a merry .188. If he qualified for the leaders, he would be last behind Suarez at .204....

Asked by: OBS2.0

Answered: 9/24/2021

Thanks. But it isn't ALL bad fortune. If you don't hit the ball hard, you get a low batting average on balls in play. Voros demonstrated that it is more or less random from the pitcher's point of view. It's kind of sort of random from the hitter's point of view.

Your comments (and stats) on the strong teams dominating September baseball can be found in the Orioles team comment of the 1984 Abstract.

Asked by: wovenstrap

Answered: 9/24/2021

Thanks.

I saw on the Hot Batters page that Juan Soto is at 130 degrees, much higher than anybody else, and higher than I can remember seeing. Is there a record for Hot Batters temperature, and what would be considered truly remarkable?

Asked by: pablo

Answered: 9/24/2021

130 is pretty remarkable. I think Frank Howard when he got to. . .what was, 10 homers in 6 games or something?. . . I think he would have been about 150. 130 is REALLY hot.

Through 153 games this year the Red Sox are 88-65 and the Blue Jays are 85-68 and this puts Toronto is three games back in the Wild Card race. The Red Sox and Blue Jays have scored almost the exact same number of runs so far (BOS 794, TOR 796) but Toronto has given up far fewer runs per game (BOS 710, TOR 630). As such Toronto has a run differential of +166 to the Red Sox run differential of +84. My questions are in terms of run differential what level should a team be at when the season ends to make the playoffs and why does Toronto's record seem to not reflect the fact that they have scored a whole lot more than they have given up. I figure Toronto will still sneak into the Wild Card game, but historically have their ever been any teams that have had a plus +150 differential and failed to make the playoffs?

Asked by: BrianFleming

Answered: 9/24/2021

I'm sorry; I don't have any of those facts in my head, nor do I know where I could look them up any more quickly than you could. Maybe a better question for reader posts? Sorry.

Did you ever get a satisfactory answer to your Cody Bellinger question on Twitter? It seems to me the best answer I saw in the replies (Twitterbacks?, Tweetbacks?) was George Scott. Of course he eventually rebounded, at least somewhat, from his fall in 1968, but he did lose nearly 100 points in his OPS+ that year. Bellinger will hopefully have a rebound too.

Asked by: bhalbleib

Answered: 9/23/2021

I am now certain that there is no answer, there is no precedent. Scott or Grady Sizemore were probably the best answers selected, but honestly, they're not remotely comparable. So. . .there just isn't any answer. It's unprecedented.

Regarding pitches thrown by Cy Young or Nolan Ryan: The Ryan Express faced 22,575 batters in his career. Take that times 3.6 pitches per plate appearance or so, and you've got more than 80,000 pitches. But Cy Young faced 29,565 batters in his career -- a huge advantage of about 7,000. Even if batters were taking fewer pitches during that era, ol' Cy probably threw close to 90,000 pitches.

Context: 40-year-old Adam Wainwright hasn't even thrown to 10,000 batters yet.

Asked by: rwarn17588

Answered: 9/23/2021

At the rate he is going, he will pass Nolan Ryan when he is 68.

I did a little research here, and it seems clear to me that not only is it POSSIBLE that Ryan threw more pitches than Denton True Young, it is enormously likely that he did.

First of all, the pitches per batter average for this season, 2021, is 3.90 through September 23, not 3.60. It may have been 3.60 several years ago. This relates to the earlier discussion here about batters now fouling off more two-strike pitches, thus extending at bats.

But within the data for this seasons, pitchers throwing 600 or more pitches, the range of pitchers per batter goes from 3.10 (Richard Bleier) to 4.52 (Heath Hembree). That's a range of 46%, top to bottom.

Ryan faced 22,575 batters in his career, Young 29, 565, your data, I didn't double-check it. That's only a difference of 31%. That's with the range of variation by pitchers in a season, if we use the 600-pitches cutoff.

If we use a higher cutoff for repreentative pitchers, 150 innings, then the range in 2021 is from 3.48 (Cole Irvin) to 4.21 (Tyler Mahle), and that is only a 21% range. But we are dealing with the combined effects of two types of deviation--deviations between pitchers of a different type, and deviations over time. Over time, it is overwhelmingly likely that the number of pitches per batter has increased steadily. This seems obvious from:

1) The increasing length of games,

2) The increasing numbers of strikeouts, and

3) The fact that this HAS increased quite significantly just in the 30 to 40 years that we have been tracking it.

From the beginning of Cy Young's career to the end of Ryan's is more than 100 years. We would have to assume that there was AT LEAST a 20% increase in pitches per batter over those 100 years, more probably 40% or 50%. But if there was a 20% increase in the level for the average pitcher, then Ryan would only have to exceed Young by 9% relative to the league to have a higher total. That's nothing.

But Ryan was a VERY extreme power-type pitcher, probably the most extreme pitcher of that type in baseball history, while Cy Young was a very extreme control-type pitcher. Cy Young led his league in fewest walks per nine innings as a rookie in 1890, and subsquently led his league in 1893, 1894, 1895, 1896, 1897, 1898, 1899, 1900, 1901, 1903, 1904, 1905, and 1906, whereas he was never among the league leaders in strikeouts per nine inning until, incongrously, he broke into the league leaders in 1905, when he was 38 years old.

So a 9% difference in them, relative to the league, seems impossibly small. THus, I would have to conclude that it was overwhelmingly likely that Ryan did, in fact, throw more pitches than Cy Young did.

My thanks to Joe Rosales, Sports Info Solutions, for help in the research.

Not a question but a follow up. Mitchell Page and his wonderful 1977 season were discussed in your recent reply to a question about the 1978 Oakland A's. I remember Page as a minor leaguer. He went to college, so his first full pro season was with Salem (Carolina League) at age 22 in 1974. Early in the year, Page was struggling. The manager (John Lipon) and other people in the Pirates' organization couldn't figure out why this strong guy with the beautiful swing could only hit consistently in batting practice. It turns out he needed glasses. Once he got the glasses, he really hit very well. His slash line that year of .296/.400/.494 was probably fifty points higher than that across the board after he got the glasses.

Asked by: evanecurb

Answered: 9/23/2021

Thanks. I thought I was discussing the season the reader had asked about, but those things often befuddle me.

Michael P asked who you thought threw more pitches in his MLB career, Cy Young or Nolan Ryan. Several years back the great Tom Tango came up with this formula to estimate number of pitches thrown when pitch count data is unavailable:

3.3 PA + 1.5 SO + 2.2 BB, where PA = 3 IP + H + BB

Using this estimator (and hoping my math is correct) we get 107,126 thrown by Cy, 90,211 by Nolan.

I'm not sure how well the formula holds up across eras, but it seems like a pretty good back-of-the-envelope estimate.

Asked by: briangunn

Answered: 9/23/2021

See later answer.

You had a bit in one of the Abstracts about counterintuitive winning percentages after September 1 or something. Your finding was that in the AL, only 4 teams had a winning record but 8 teams had a losing record or something of that nature. Likely related to the issue you described as "normalcy" in a recent answer.

Asked by: wovenstrap

Answered: 9/23/2021

Can't say that I remember that. I didn't realize that teams tend to separate as the season rolls on until about 2004 or 2005. As the season goes on, bad teams get worse, which causes the winning percentage of the good teams to improve. I didn't learn that until years after finishing the Abstracts.

"Players get hurt during the season; players find their confidence and improve their skills during the season."

I understand that streaks and slumps during the season are considered random events that even out. You always hear players and coaches say that so-and-so is really seeing the ball well or something like that. My question is do players' mechanics change during the season so that they really are better at given times and worse at others? Does that have anything to do with streaks and slumps?

Asked by: Marc Schneider

Answered: 9/23/2021

Probably not, but you can't absolutely prove that it doesn't. What we can demonstrate is that slumps and hot streaks have near-zero predictive significance. A player can be 0-for-20 in his last 20 at bats, or 11-for-20; his chance of getting a hit in the next at bat is the same. It seems likely that if variations in performance levels during the season had a actual cause, rather than being random, that they would persist enough to have predictive significance.

Or, looked at this way. . ..if you create a statistical image of a player and then run him through 600 randomly generated imitation at bats, the player will get hot and cold to (essentially) the exact same extent that an ACTUAL player with the same skills will get hot and cold. This also suggests that it is not a real phemonon, but merely a trick of perception used to explain what otherwise seems without explanation.

But we should be cautious in drawing conclusions from the data. We know that hitters skills change over time, otherwise Frank Thomas would still be hitting .325 with a 1.000 OPS or whatever it was, and we know that players fighting small injuries would certainly lose effectiveness. We have to conclude, then, that THOSE changes are able to hide in highly randomized data so as to be not reliably detectable, and from that, we should conclude that it is possible for changes in the level of performance to hide undetected in the data.

THe batter performs in a context in which there are dozens or hundreds of variables which are outside of his control, many of which have near-random sequences. Certainly and obviously a player who is no better this week than last week can go 3-for-20 last week and 11-for-20 this week based on changes in variables outside of his control. But to assert that ALL changes in player performance levels during a season are a result of variables outside of his control is a step too far, and would leave you asserting things that you don't actually know to be true.

Today I did a deep dive on the 1978 Oakland A's, who, after finishing last in 1977, started the season 19-5 and were still in contention (60-54, 2 1/2 games out) as late as August 7th before a disastrous finish that sunk them to 69-93 and sixth place. Against the five AL teams that won 90+ games, those A's were 13-43. Against the six teams that won less than 75 games, they were 40-26. I know you have a method to calculate expected W-L records in matchups that is based on the teams' overall W-L records. Are those A's results against the good teams and the bad teams outliers, or are they within the range of expectations?

Asked by: evanecurb

Answered: 9/22/2021

I'd have to study the question; I don't know. Generally you can assume that a 70-92 team is a 70-92 team from the beginning of the year to the end, but that (a) is never ABSOLUTELY true, and (b) is completely untrue in some cases. That certainly sounds like an odd case.

I mean. . .team's level of competitive quality changes, right? Those changes are not sudden; they usually are gradual, but sometimes they SEEM sudden. We tend to ASSUME, as an analytical convenience, that those changes occur between seasons rather than within seasons, but logically, this cannot be an absolute truth. Players get hurt during the season; players find their confidence and improve their skills during the season.

*****************************

Here's something about that team that has always fascinated me. Mitchell Page that year, a rookie, was truly a fantastic offensive player. He hit over .300 with 20+ homers, but he also drew 78 walks, giving him an on base percentage over .400, and also, he was 42-for-47 as a base stealer. It truly is one of the greatest rookie seasons of all time. When you remember that this is the Oakland Coliseum, rather than Fenway Park, the season is not that far from Fred Lynn's rookie season in '75.

He never had another season like that, although he was still a good player in 1978, but never close to that MVP-candidate level of 1977. But here is what has always fascinated me most. Mitchell Page was a high school teammate and close friend of Al Cowens. The two of them were both born in LA in October, 1951.

Cowens ALSO had a fantastic season that year, and a season quite siimilar to Page's. ..about the same average and home runs. Cowens' season is probably a little less impresive than Page's--and Cowens was second in the MVP voting that season.

But, like Page, he never really did anything in the rest of his career. I know I have written about this before, but it has just always fascinated me.

You've said that you don't believe that teams tank. But isn't it also true that when contenders play against non-contenders this time of year, the contenders win more often than they do ordinarily? It sure seems like they do. If it's not tanking, what is it?

Asked by: manhattanhi

Answered: 9/22/2021

Normalcy. Go back to data from 1950, 1953, 1960, 1965, 1937. . .you'll find the same phenomenon. It has always happened; it doesn't have anything to do with decisions made by the front office or the manager. Once a team is eliminated, they start to fall apart a little bit. They lose hope, lose focus; it's just normal life. The same syndrome happens in your life; it happens in any business, in a marriage, whatever. When you lose hope, you lose energy, you lose focus, you start to drift. It's not "tanking"; it is human nature. It's life.

Bill,

I'm not sure if you are aware of this, but I am reading a novel called "The Index of Self-Destructive Acts" in which you are essentially the hero of one of the main characters:

"He'd discovered Bill James, the Society for American Baseball Research, and the way of thinking that James called sabermetrics. His favorite baseball book was now James's Historical Baseball Abstract."

Character speaking: "Bill James demonstrated the relationship between wins and run differential almost thirty years ago."

Asked by: Marc Schneider

Answered: 9/22/2021

Oh, it's a very good novel. They sent me a copy of it pre-publication, and I planned to do an event in this area with the author, but then the pandemic screwed everything up and I wound up doing a more minor event by Zoom and never met him face to face.

But the book is GREAT. It really is a terrific novel; I was very proud to be a small part of it. The character you mention is based on Nate Silver.