moscow25
In summary, instead of giving each new start a 3% value in starting pitcher ratings, you could use a sliding scale based on total starts in the past calendar year. Interpolate the values between 3% [25+ starts] and 6% [0 starts]. Strasbourg would still earn his rating, but he could do so up to twice as fast since he was inactive for a long time.

Similarly, John Smoltz would be able to re-establish himself faster as a top notch starter after he returned from the bullpen. Hideo Nomo and Jose Contreras were really good out of the gate when they were fooling everyone with decent speed and weird deliveries. I don't see a reason not to let them rise quickly in their first few starts.

I don't see this infecting the top 10-20 spots with bad pitchers who were laid off and got lucky when they got back. Their initial scores would be far too low to rise that high before the normal rating adjustment and their mediocrity kicked in. And having Jose Rijo earn a cameo in the top 20 after coming back strong from surgery wouldn't be so bad, would it?
4:23 AM Apr 20th

bjames
With regard to Strasburg. ...earn his bones or earn his stripes is a pretty good expression. I want to see him earn his posiiton among the top pitchers, rather than have it handed to him on the basis of scouting reports.

The more serious issue, to me, is Johan Santana, which is more analogous to the Serena Williams issue. I did think about whether there was a way to make a pitching coming back from an injury, like Santana or Wainwright or Josh Johnson, bounce BACK more rapidly than a pitcher who has no history of being good. I wouldn't say that I've ruled that out; I just haven't really figured out how to do it yet.

7:51 PM Apr 16th

bjames
Also, I would point out that there is NO problem with Strasburg or Darvish; you are merely imagining that there might be one.
8:50 AM Apr 16th

moscow25
As to Tango's point about ratings for new pitchers... in chess, there are two changes made to make it easier for fast-improving players to rise in the ratings. First, any unrated players gets 20 games or so rated as "provisional" where rating changes are essentially doubled. Secondly, lower rated players have larger changes to their rating as result of a single game, than do higher rated players do.

The first point would certainly help here. If players get a 5% adujstmen (instead of your current 3%) as a result of their first 30 starts, the Darvish and Strasbourg problems would be resolved. It's slightly inelegant, but not really. Bill's numbers are tuned for adjusting values of sustained performance. It's reasonable to make each start matter closer to 5% when the pitcher has very few starts under his belt.

It doesn't seem unreasonable that Gooden peaked at #11 by the end of his rookie year, but him being #70 in August sure seems a little low. The same will be true with Darvish if he is what he's supposed to be.
3:44 AM Apr 16th

bjames
1) Regarding Mark Fidrych. . ..Fidrych never came close to making the top 10 on the list. What most people probably don't remember. . .Fidrych did make a start in 1976 until May 15, one month after today's date. By the end of May Fidrych had climbed to 105th in the rankings, among 142 ranked pitchers; by the end of June, he was 81st among 149. By the end of August he was 39th of 165 ranked pitchers, based on having pitched brilliant baseball for 3 and half months.

Fidrych was shelled in his first start in September, however, and posted a 3.83 ERA in September, well above the league norm of 3.52; he thus dropped in the rankings from 39th to 52nd. He opened 1977 injured, and fell further in the rankings, but pitched well in June of 1977, and recovered to 40th on the list by the end of June, 1977.

2) Fernando opened 1981 like a house afire, and by May 1, 1981, was in 87th place among 156 ranked pitchers. By June 1 he was in 73rd place, and, by the time the strike the season in mid-June, he was 67th among 157 ranked pitchers.

He held there, of course, during the strike, resumed moving after the strike, and by the end of the season he was in 33rd place; he opened 1982 in 33rd place among 120 ranked pitchers. By May 1 he was in 28th place, by June 1 in 12th place. When he beat Cincinnati 2-1 on June 19, 1982, with a 5-hitter, that put him in the top 10. So really, it took him just over one year to reach the top ten. It took him a year and almost three months, but the 1981 strike wiped out almost two months of that, so. . .just over a year to reach the top ten.

I don't believe that Fernando ever reached the top spot; I wouldn't swear to that, but I don't think so. He did reach second place, as we will see. . ..

3) Dwight Gooden, the same; it took him just over a year to reach the top ten. Gooden, of course, opened the 1984 season unranked. He moved to 124th in the rankings by April 30, 98th by May 31, 68th by June 30, 53rd by July 31, 38th by August 31, and 11th place by the end of the regular season. Since the Mets did not make the playoffs he treaded water, and opened the 1985 season in 11th place.

He moved into the top ten in mid-April, 1985, and by the end of April he was in 7th place. By the end of May he was 4th. By the end of June he was second, behind Dave Stieb.

Stieb, I think, had held the #1 spot during almost all of 1984 and the first half of 1985. Stieb was a good pitcher, and there was a real shortage of GREAT pitchers in the early 1980s; many of the Cy Young Awards were won by relievers, and two of the others were won by Pete Vuckovich and John Denny, who were hardly great pitchers. Stieb was the best there was in that era; Fernando and Mario Soto took turns in the 2-3 spots until Gooden passed them in June, 1985, and then mostly held the 3-4 spots after then.

Stieb held off Gooden until August 4, 1985. On that date Stieb pitched poorly and Gooden pitched a 5-hitter, beating Chicago 4-1, which pushed Gooden suddenly almost ten points ahead of Stieb into first place. By the end of the 1985 season Gooden was far in front, with John Tudor second and Dave Stieb in 3rd.

4) While I was tracking Gooden, I also tracked Bret Saberhagen, who, you may remember, also came up at the same time as Gooden, was just one year older than Gooden, and also won the Cy Young Award in 1985, as did Gooden.

Saberhagen in 1984 moved to 125th place in the rankings by May 31, but then went BACKWARD most of the rest of 1984; by August 31 he was in 139th place. He pitched well in September, and finished the season in 117th place.

In 1985, his Cy Young season, he moved up to 112th place by April 30, 94th by May 31, 80th by June 30, 48th by July 31, and by the end of the season was in 20th place. So in that case the Cy Young Award winner wound up the season still ranked as the 20th best starting pitcher in baseball.

4) By the end of the 1986 season, Mike Scott had emerged as the #1 starting pitcher in baseball, dropping Gooden to 2nd, Jack Morris 3rd. Roger Clemens, 24-4 that season (as Gooden was in 1984), wound up the season ranked as the #7 starting pitcher in baseball.
3:15 PM Apr 15th

tangotiger
Right, I agree with your basic sentiment. I look at it like Wins Above Replacement. Let's say the replacement level for Game Score is 40 (I don't know what it is, just an illustration). You have one guy with a 70 game score on average, on 210 IP, and you have another guy with a 55 game score on average, over two years, on 420 IP. So, I think, we want those two guys to come out close to even. (Of course, things get more complicated because 210 IP for the first guy is under 30 starts, and 420 IP for the second guy is closer to 65 starts.)
12:38 AM Apr 15th

tangotiger
Bill, when did Fidrych break the top 10 in his rookie season, and where did he end up?
9:17 AM Apr 14th

bjames
1) In 1971 Vida Blue--essentially a rookie--was in 7th place by the end of the season, against very strong competition (Seaver, Gaylord Perry, Bob Gibson, Jim Palmer, Catfish Hunter, Ferguson Jenkins, Mike Cuellar all in their prime.) Of course, that's unusual, and I wouldn't expect to see that again in my lifetime.

2) Regarding Lilly. . ..yes, but there are other reasons why he is better than you are giving him credit for. You said that Kuroda pitched more innings with a better ERA than Lilly each of the last two years, but Lilly actually outpitched Kuroda in 2010. Kuroda pitched more innings by a margin of 196 to 194, not a real factor. Kuroda gave up 4 fewer earned runs but 4 MORE total runs. Lilly had more strikeouts and fewer walks. Lilly was pitching in Wrigley Field (Park Factor 108); Kuroda, in Dodger Stadium (Park Factor 98). Lilly was far ahead of Kuroda through 2009, was better than Kuroda in 2010, and was better than Kuroda in the last third of 2011. Kuroda was better in the first 2/3 of 2011, but that's not enough to say that he should rate higher.

3) Comparing Lilly and Hudson, again, Lilly is better than you're giving him credit for (in addition to having been significantly better in the last third of the 2011 season.) Somebody said that Hudson was FAR better than Lilly over the last two years, but that's not true. He was SLIGHTLY better, perhaps. In 2010 Lilly had 27 more strikeouts than Hudson, and 30 fewer walks. In 2011 they had the same number of strikeouts, but Lilly had 5 fewer walks. Lilly's ERA in road games was better than Hudson's (3.74 to 3.84) over the two seasons, despite the fact that Lilly had started 4 games in Colorado to Hudson's 2.
5:37 PM Apr 13th

enamee
I had forgotten how strongly Lilly ended 2011. That said, compare his last 11 starts to Kuroda's last 11 starts (and Tim Hudson's, since I brought him up too):

Lilly: 69 IP, 64 SO, 23 walks, 5 homers allowed, 2.09 ERA

Kuroda: 69 IP, 58 SO, 13 walks, 10 homers allowed, 3.00 ERA

Hudson: 73.2 IP, 54 SO, 21 walks, 5 homers allowed, 3.05 ERA

All things considered, those are very similar lines. Lilly had more strikeouts, Kuroda fewer walks but more homers allowed. Lilly was the better pitcher, sure, but he wasn't THAT much better than Kuroda and Hudson -- certainly not so much better that I'd expect him to vault ahead of them. Especially given that both Kuroda and Hudson were very good over the course of 2011, whereas Lilly was quite bad until those last 11 starts.

Anyway, all this amounts to nitpicking. I just find the Lilly ranking to be very curious, so I thought I'd dig into it a little further.
11:44 AM Apr 13th

bjames
Responding to Moscow, I'm a little afraid of presenting the reader with more information than he will find interesting. If I bury the rankings in a blizzard of other numbers, I'm afraid it would be hard to focus on the information that is the main point of the chart. Maybe we could do that when people get used to reading the chart, and can assimilate it more quickly.
3:01 PM Apr 12th

moscow25
Really enjoying these weekly updates. Reminds me of reading about monthly updates to the world chess ratings with Kasparov and Anand then later Karlson and Nakamura. Is there any way to include more info in the columns, like the number of starts since the last update, game scores since last update, or a link for each pitcher to recent game performances?
2:28 PM Apr 12th

enamee
Also, Lilly ranks ahead of Tim Hudson, who has been vastly better over the past two years. Any idea why the rankings like Lilly so much?

Matthew
9:54 AM Apr 12th