Congratulations Randy Johnson
Posted by Andy on June 5, 2009
In case you've been under a rock, Randy Johnson became the 24th pitcher to win 300 career games.
Among that group, he's second in career strikeouts:
Cnt Player **SO** W From To Ages G GS CG SHO GF L W-L% SV IP H R ER BB ERA ERA+ HR BF IBB HBP BK WP Teams +----+-----------------+--------+---+----+----+-----+----+---+---+---+---+---+-----+---+------+----+----+----+----+------+----+---+-----+---+---+---+---+-----------+ 1 Nolan Ryan 5714 324 1966 1993 19-46 807 773 222 61 13 292 .526 3 5386 3923 2178 1911 2795 3.19 111 321 22575 78 158 33 277 NYM-CAL-HOU-TEX 2 Randy Johnson 4845 300 1988 2009 24-45 607 597 100 37 5 164 .647 2 4097.1 3307 1682 1494 1487 3.28 136 402 16902 37 188 33 106 MON-TOT-SEA-ARI-NYY-SFG 3 Roger Clemens 4672 354 1984 2007 21-44 709 707 118 46 0 184 .658 0 4916.2 4185 1885 1707 1580 3.12 143 363 20240 63 159 20 143 BOS-TOR-NYY-HOU 4 Steve Carlton 4136 329 1965 1988 20-43 741 709 254 55 13 244 .574 2 5217.1 4672 2130 1864 1833 3.22 115 414 21683 150 53 90 183 STL-PHI-TOT-MIN 5 Tom Seaver 3640 311 1967 1986 22-41 656 647 231 61 6 205 .603 1 4782.2 3971 1674 1521 1390 2.86 127 380 19369 116 76 8 126 NYM-TOT-CIN-NYM-CHW 6 Don Sutton 3574 324 1966 1988 21-43 774 756 178 58 12 256 .559 5 5282.1 4692 2104 1914 1343 3.26 108 472 21631 102 82 21 112 LAD-HOU-TOT-MIL-CAL-LAD 7 Gaylord Perry 3534 314 1962 1983 23-44 777 690 303 53 33 265 .542 11 5350.1 4938 2128 1846 1379 3.11 117 399 21953 164 108 6 160 SFG-CLE-TOT-TEX-SDP-TOT-ATL-SEA 8 Walter Johnson 3509 417 1907 1927 19-39 802 666 531 110 129 279 .599 34 5914.2 4913 1902 1424 1363 2.17 147 97 23749 203 4 155 WSH 9 Greg Maddux 3371 355 1986 2008 20-42 744 740 109 35 3 227 .610 0 5008.1 4726 1981 1756 999 3.16 132 353 20421 177 137 28 70 CHC-ATL-TOT-SDP 10 Phil Niekro 3342 318 1964 1987 25-48 864 716 245 45 83 274 .537 29 5404.1 5044 2337 2012 1809 3.35 115 482 22677 86 123 42 226 MLN-ATL-NYY-CLE-TOT 11 Cy Young 2803 511 1890 1911 23-44 906 815 749 76 84 316 .618 17 7354.2 7092 3167 2147 1217 2.63 138 138 30058 163 3 156 CLV-STL-BOS-CLE-TOT 12 Tom Glavine 2607 305 1987 2008 21-42 682 682 56 25 0 203 .600 0 4413.1 4298 1900 1734 1500 3.54 118 356 18604 145 66 7 65 ATL-NYM 13 Warren Spahn 2583 363 1942 1965 21-44 750 665 382 63 68 245 .597 29 5243.2 4830 2016 1798 1434 3.09 118 434 21547 60 42 5 81 BSN-MLN-TOT 14 Tim Keefe 2562 342 1880 1893 23-36 600 594 554 39 7 225 .603 2 5047.2 4439 2468 1472 1220 2.62 127 81 20975 96 0 233 TRO-NYP-NYG-NYI-TOT-PHI 15 Christy Mathewson 2502 373 1900 1916 19-35 635 551 434 79 74 188 .665 28 4780.2 4218 1616 1133 844 2.13 135 91 19136 59 6 114 NYG-TOT 16 Early Wynn 2334 300 1939 1963 19-43 691 612 290 49 66 244 .551 15 4564 4291 2037 1796 1775 3.54 107 338 19408 36 64 2 51 WSH-CLE-CHW 17 Lefty Grove 2266 300 1925 1941 25-41 616 457 298 35 123 141 .680 55 3940.2 3849 1594 1339 1187 3.06 148 162 16633 42 1 51 PHA-BOS 18 Eddie Plank 2246 326 1901 1917 25-41 623 529 410 69 74 194 .627 23 4495.2 3958 1569 1174 1072 2.35 122 41 18225 196 6 87 PHA-SLM-SLB 19 Pete Alexander 2198 373 1911 1930 24-43 696 599 437 90 80 208 .642 32 5190 4868 1851 1476 951 2.56 135 164 20928 70 1 39 PHI-CHC-TOT-STL-PHI 20 John Clarkson 1978 328 1882 1894 20-32 531 518 485 37 12 178 .648 5 4536.1 4295 2376 1417 1191 2.81 134 161 11954 35 0 182 WOR-CHC-BSN-TOT-CLV 21 Kid Nichols 1868 361 1890 1906 20-36 620 561 531 48 56 208 .634 17 5056.1 4912 2477 1660 1268 2.95 140 156 21243 133 1 169 BSN-STL-TOT-PHI 22 Mickey Welch 1850 307 1880 1892 20-32 565 549 525 41 15 210 .594 4 4802 4587 2556 1447 1297 2.71 114 106 18186 29 1 274 TRO-NYG 23 Charley Radbourn 1830 309 1881 1891 26-36 528 503 489 35 24 195 .613 2 4535.1 4335 2275 1348 875 2.67 119 117 18910 54 0 217 PRO-BSN-BOS-CIN 24 Pud Galvin 1806 364 1875 1892 18-35 705 689 646 57 17 310 .540 2 6003.1 6405 3355 1910 745 2.86 107 122 25234 57 1 223 STL-BUF-TOT-PIT-PBB-TOT
I give him less than a 5% chance of catching Nolan Ryan in that department, although RJ is a far better pitcher than Ryan.
Among the 300-game winners, top K/9 rates:
Cnt Player **SO/9** W From To Ages G GS CG SHO GF L W-L% SV IP H R ER BB SO ERA ERA+ HR BF IBB HBP BK WP Teams +----+-----------------+---------+---+----+----+-----+----+---+---+---+---+---+-----+---+------+----+----+----+----+----+------+----+---+-----+---+---+---+---+-----------+ 1 Randy Johnson 10.64 300 1988 2009 24-45 607 597 100 37 5 164 .647 2 4097.1 3307 1682 1494 1487 4845 3.28 136 402 16902 37 188 33 106 MON-TOT-SEA-ARI-NYY-SFG 2 Nolan Ryan 9.55 324 1966 1993 19-46 807 773 222 61 13 292 .526 3 5386 3923 2178 1911 2795 5714 3.19 111 321 22575 78 158 33 277 NYM-CAL-HOU-TEX 3 Roger Clemens 8.55 354 1984 2007 21-44 709 707 118 46 0 184 .658 0 4916.2 4185 1885 1707 1580 4672 3.12 143 363 20240 63 159 20 143 BOS-TOR-NYY-HOU 4 Steve Carlton 7.13 329 1965 1988 20-43 741 709 254 55 13 244 .574 2 5217.1 4672 2130 1864 1833 4136 3.22 115 414 21683 150 53 90 183 STL-PHI-TOT-MIN 5 Tom Seaver 6.85 311 1967 1986 22-41 656 647 231 61 6 205 .603 1 4782.2 3971 1674 1521 1390 3640 2.86 127 380 19369 116 76 8 126 NYM-TOT-CIN-NYM-CHW
Fewest hits per 9:
Cnt Player **H/9** W From To Ages G GS CG SHO GF L W-L% SV IP H R ER BB SO ERA ERA+ HR BF IBB HBP BK WP Teams +----+-----------------+---------+---+----+----+-----+----+---+---+---+---+---+-----+---+------+----+----+----+----+----+------+----+---+-----+---+---+---+---+-----------+ 1 Nolan Ryan 6.56 324 1966 1993 19-46 807 773 222 61 13 292 .526 3 5386 3923 2178 1911 2795 5714 3.19 111 321 22575 78 158 33 277 NYM-CAL-HOU-TEX 2 Randy Johnson 7.26 300 1988 2009 24-45 607 597 100 37 5 164 .647 2 4097.1 3307 1682 1494 1487 4845 3.28 136 402 16902 37 188 33 106 MON-TOT-SEA-ARI-NYY-SFG 3 Tom Seaver 7.47 311 1967 1986 22-41 656 647 231 61 6 205 .603 1 4782.2 3971 1674 1521 1390 3640 2.86 127 380 19369 116 76 8 126 NYM-TOT-CIN-NYM-CHW 4 Walter Johnson 7.48 417 1907 1927 19-39 802 666 531 110 129 279 .599 34 5914.2 4913 1902 1424 1363 3509 2.17 147 97 23749 203 4 155 WSH 5 Roger Clemens 7.66 354 1984 2007 21-44 709 707 118 46 0 184 .658 0 4916.2 4185 1885 1707 1580 4672 3.12 143 363 20240 63 159 20 143 BOS-TOR-NYY-HOU
Nolan gets the edge there.
Best W-L%:
Cnt Player **W-L%** W From To Ages G GS CG SHO GF L SV IP H R ER BB SO ERA ERA+ HR BF IBB HBP BK WP Teams +----+-----------------+---------+---+----+----+-----+----+---+---+---+---+---+---+------+----+----+----+----+----+------+----+---+-----+---+---+---+---+-----------+ 1 Lefty Grove .680 300 1925 1941 25-41 616 457 298 35 123 141 55 3940.2 3849 1594 1339 1187 2266 3.06 148 162 16633 42 1 51 PHA-BOS 2 Christy Mathewson .665 373 1900 1916 19-35 635 551 434 79 74 188 28 4780.2 4218 1616 1133 844 2502 2.13 135 91 19136 59 6 114 NYG-TOT 3 Roger Clemens .658 354 1984 2007 21-44 709 707 118 46 0 184 0 4916.2 4185 1885 1707 1580 4672 3.12 143 363 20240 63 159 20 143 BOS-TOR-NYY-HOU 4 John Clarkson .648 328 1882 1894 20-32 531 518 485 37 12 178 5 4536.1 4295 2376 1417 1191 1978 2.81 134 161 11954 35 0 182 WOR-CHC-BSN-TOT-CLV 5 Randy Johnson .647 300 1988 2009 24-45 607 597 100 37 5 164 2 4097.1 3307 1682 1494 1487 4845 3.28 136 402 16902 37 188 33 106 MON-TOT-SEA-ARI-NYY-SFG
Guess who's last out of 24? (Yep, Ryan.)
And finally, top ERA+ values:
Cnt Player **ERA+** W From To Ages G GS CG SHO GF L W-L% SV IP H R ER BB SO ERA HR BF IBB HBP BK WP Teams +----+-----------------+--------+---+----+----+-----+----+---+---+---+---+---+-----+---+------+----+----+----+----+----+------+---+-----+---+---+---+---+-----------+ 1 Lefty Grove 148 300 1925 1941 25-41 616 457 298 35 123 141 .680 55 3940.2 3849 1594 1339 1187 2266 3.06 162 16633 42 1 51 PHA-BOS 2 Walter Johnson 147 417 1907 1927 19-39 802 666 531 110 129 279 .599 34 5914.2 4913 1902 1424 1363 3509 2.17 97 23749 203 4 155 WSH 3 Roger Clemens 143 354 1984 2007 21-44 709 707 118 46 0 184 .658 0 4916.2 4185 1885 1707 1580 4672 3.12 363 20240 63 159 20 143 BOS-TOR-NYY-HOU 4 Kid Nichols 140 361 1890 1906 20-36 620 561 531 48 56 208 .634 17 5056.1 4912 2477 1660 1268 1868 2.95 156 21243 133 1 169 BSN-STL-TOT-PHI 5 Cy Young 138 511 1890 1911 23-44 906 815 749 76 84 316 .618 17 7354.2 7092 3167 2147 1217 2803 2.63 138 30058 163 3 156 CLV-STL-BOS-CLE-TOT 6 Randy Johnson 136 300 1988 2009 24-45 607 597 100 37 5 164 .647 2 4097.1 3307 1682 1494 1487 4845 3.28 402 16902 37 188 33 106 MON-TOT-SEA-ARI-NYY-SFG
Somehow, it almost seems like Randy Johnson is underrated.
Next stop: Cooperstown
June 5th, 2009 at 10:43 am
Underrated maybe.. more like underexposed. This guy played in a time when there were twice as many teams, and he played in relatively small markets, seattle, montreal, and his best years for arizona. yes he played 2 years for the yankees but they were well past his prime.
also, not sure if u noticed but his 10.7 k/9 is the best ever, not just for 300 winners
June 5th, 2009 at 10:53 am
amazing work after the age of 30. but i must say, it bugs me to hear just about every commentator say "This may never happen again" or "This probably won't happen again in our lifetime."
Oswalt could do it by 2019, Buerle and Sabathia by 2020, Santana by 2021 and Beckett by 2022.
June 5th, 2009 at 10:57 am
Randy Johnson got overshadowed for much of his career by fellow 300-game-winners Roger Clemens and Greg Maddux. While Randy's five Cy Young awards are more than anyone had ever won prior to the 1990’s, they came in the midst of Rocket’s winning seven Cy Youngs. And though The Big Unit won four Cy Youngs in a row from 1999 to 2002, that feat had already been done by the Mad Dog just a few years prior, from 1992 to 1995.
Clemens was known for his toughness (except in the post season), and as an established winner; he posted his fourth 20-game-winning season the year Johnson posted his first (1997), even though Roger is only about a year older than Randy. Even in that year, Clemens out-won Johnson 21 to 20. Maddux, about 2½ years younger than Johnson, established himself as one of the elite pitchers of the era by winning those four consecutive Cy Youngs and posting ERAs under 2.75 for seven consecutive years (including ERA+s of 271 and 262 in back-to-back years in 1994 and 1995). Maddux earned a reputation as one of the smartest pitchers ever to pitch in the big leagues.
Johnson, on the other hand, was known mostly as a curiosity – the tallest player (at the time) to ever make the big leagues. He hadn’t really established himself as a winner and a staff ace until his final seasons in Seattle, as he was going into his 30’s. He didn’t pitch 250 innings in a season until 1993 (also his first season winning more than 14 games), at age 29. His three 20-game-winning seasons all occurred in his thirties (the first at age 33), as did his four ERA titles (the first occurring at age 31). Though he was clearly establishing himself as a dominant pitcher (and one of the most feared by batters – just ask John Kruk), most did not give him a serious shot at 300 wins because of his “slow start.”
But that slow start – pitching less than 220 innings and facing less than 1000 batters per season until age 29 – may ultimately have ensured his long-term success and his membership in the 300-win club. There are countless flame throwers who light up the radar guns in their twenties, only to burn out and retire by their mid-thirties. On the other hand, hard-throwing pitchers who face more limited use early in their careers (more common with lefties, since they have to master predominantly right-handed lineups) tend to have fewer arm problems and pitch longer – in cases like Johnson’s, well into their forties. Unit’s left arm might well have been saved some of the wear-and-tear of his nasty fastball and wicked slider while still in its formative years thanks to that conservative use.
All that said, we, as fans, might have overlooked Randy Johnson until this year because we were too busy going gaga over Clemens (until he fell from grace) and Maddux, but you can bet big league hitters over the last twenty-plus years haven’t overlooked him!
One final note: Johnson has more seasons of 16+ wins after his 36th birthday than he did before it! (We'll even count 1999 - when he won his 15th game of the season on his 36th birthday - in the "before" category.)
June 5th, 2009 at 10:59 am
Does anybody have Favorite Toy updates for active pitchers getting to 300 wins?
June 5th, 2009 at 11:58 am
Here's what I came up with from the start of the 2009 season:
Sabathia 19%
Webb 10%
Halladay 8.6%
Zambrano 7.6%
Oswalt 5%
Santana 5%
Beckett 0.9%
Pettitte 0.6%
Clearly Webb is seriously hurting his chances with his DL stint and Halladay will increase his, if he stays healthy and wins 20 again.
June 5th, 2009 at 12:02 pm
Yeah, those numbers make a lot of sense to me. No offense kingturtle, but anybody who thinks that ANY active pitcher has a good shot at making 300 wins is out of their mind. All those guys taken together have less than a 50% chance of any ONE achieving it.
Honestly the active pitcher with the best shot might be Jamie Moyer, since he's said that intends to pitch at least until age 50. That's 4 more seasons past this one, and he'd have to average 12 wins per year or so. A very tall task for someone so old by baseball metrics, but I think a better bet than anybody on Zim's list.
June 5th, 2009 at 12:06 pm
I just ran Halladay projecting 20+ wins for this season and he jumps to the following:
20 -- 14.9%
21 -- 17%
22 -- 19.2%
23 -- 21.3%
24 -- 23.6%
25 -- 25.8%
So he's certainly getting in the mix of having an oustide shot.
June 5th, 2009 at 12:21 pm
PS it would be pretty cool if there were a Favorite Toy calculator linked to player pages, perhaps as a frivolity.
PPS Grienke and Lincecum can get "on the board" with 19 and 17 wins respectively.
June 5th, 2009 at 12:28 pm
As to other currently established players reaching 300 wins, while possible, I don’t think it is likely. CC Sabathia has perhaps the best chance (22% by the “favorite toy” method), but as Tim Kurkjian pointed out last night on Baseball Tonight, if CC pitched to age 40, he’d need to average 14.1 wins a year for the rest of his career. In his eight seasons prior to this year, he’s only averaging 14.6 wins – and these are supposed to be his prime years. For the other pitchers Kingturtle mentioned, if they pitched to age 40 they would need to average the following number of wins per year (current average wins per year in parenthesis, followed by “favorite toy” probability estimate):
Oswalt 17.1 W/yr (16.1, 12%)
Buehrle 16.2 W/yr (14.8*, 9%)
Santana 17.4 W/yr (15.1**, 12%)
Beckett 17.6 W/yr (12.4***, 12%)
* average does not include 2000 (used mostly as a reliever)
** average does not include 2000 & 2001 (used mostly as a reliever)
*** average does not include “cup of coffee” in 2001
Three others that Kingturtle left off his list are:
Webb 19.4 W/yr (14.5, 15%)
Zambrano 15.7 W/yr (15.2*, 15%)
Halladay 18.8 W/yr (14.8**, 13%)
* average does not include 2000 & 2001 (used mostly as a reliever)
** average does not include 1998 - 2000 (used mostly as a reliever)
Each of these guys would have to significantly step up their production (as Johnson did in his thirties), and/or pitch into their forties (as all of the current group of 300 winners did).
That said, I don’t think that no one will reach 300 wins again. After all, when the last group reached 300 wins (Perry and Carlton in 1982, Niekro and Seaver in 1985, Sutton in 1986, and, dragging behind, Ryan in 1990), the current group were just starting their careers (Clemens started in 1984, Maddux in 1986, Glavine in 1987, and Johnson in 1988) and they were not on anyone’s “300-win radar” yet. By 1986, when most thought that the then-current wave of 300-game-winners was over (most did not expect Ryan to keep chugging away and make it), Clemens had 40 wins and got everyone’s attention with a 24-win season, Maddux had 2 wins, and Glavine and Johnson were still in the minors. If any of the young studs of 1986 were 300-win candidates, it would have been 21-year-old Dwight Gooden, who had compiled 58 wins and Rookie of the Year and Cy Young awards in just three major league seasons. Even by the time Ryan entered the 300-win club in 1990, of the current group only Clemens had over 60 wins (he had 116). Gooden was still considered the front-runner with 119 wins and only one season with less than 15 wins (plus he was 2 years younger than Clemens).
So will we see another 300-game winner in the next 20-or-so years? Probably. But will it be one of the currently established pitchers with 80+ wins already? Probably not. Who knows, it might just be Derek Holland of the Texas Rangers…
June 5th, 2009 at 12:44 pm
Oops! My error - the probabilities I quoted in the last post were based on the Pythagorean method, not the "favorite toy" method. Here's what I have as the probabilities for both methods for the guys mentioned above going into the 2009 season (Favorite toy first, then Pythagorean method):
Sabathia 16.2%, 21.9%
Oswalt 0.0%, 12.2%
Buehrle 0.0%, 9.4%
Santana 0.8%, 12.9%
Beckett 0.0%, 11.5%
Webb 5.4%, 15.3%
Zambrano 5.3%, 15.3%
Halladay 1.1%, 13.1%
June 5th, 2009 at 2:48 pm
It's great to see the 300-club comparisons, and how well RJ stacks up. Consider the ultimate goals for a pitcher, and he's clearly in the top group ever, possibly the top 10:
1) Pitch a lot: It's one thing to be great, but all the talent in the world doesn't do any good on the DL. Johnson has nearly 600 starts and over 4,000 IP, among the top 40 in both categories. He has 14 seasons of 200 or more IP, leading the league twice.
2) Pitch well: Stacking up numbers is one thing; being great is another. RJ won all those Cy Youngs, led the league in ERA four times, Ks nine, WHIP three and ERA+ six. His rate stats, including the greatest ever Ks/9, are elite. His ERA+ is 136.
3) Pitch well when it counts: You might take offense to his playoff record, but don't be fooled. RJ's 3.50 ERA isn't exactly bad, and his 1.14 WHIP speaks for itself. He went 2-1 with a firt-time playoff team in 1995, and 5-1 for the champs in 2001. He was the biggest thorn in the Yankee dynasty's side, going 5-0 combined between 1995 and 2001. But the Yankee dynasty started in 1996, you might say. That's BECAUSE of RJ. The Yanks had the AL's best record in 1994 and were a serious threat in '95. And look how he finished that '95 season. His team won his last 10 starts, including a three-hitter in a one-game playoff.
4) Win: Although it's hard to measure greatness strictly by wins or wins and losses, there's no doubt about it--a pitcher takes to the mound to win. The object of the game isn't to be dominant, but to get the W. RJ can claim 300 Ws, among the highest ever, and averages 17 wins per 34 starts. His .647 WP is among the greatest for a pitcher with so many wins. His team is 367-230 in his starts, including a three-year period in which it went 55-12, and seasons of 23-11, 27-3, 21-8, 21-13, 23-12, 22-13, 24-10, 29-6, 22-12 and 21-12, the last two after age 40. From 1993 through 2006, a 14-year period, his teams went 279-138 in his starts.
In the mid to late 90s I dismissed Johnson as a frail fluke. I was dead wrong. He is among the greatest ever, with Walter Johnson, Christy Mathewson, Lefty Grove, Warren Spahn, Tom Seaver and Greg Maddux, and perhaps better any of them.
June 5th, 2009 at 3:09 pm
Tomepp, I see a discrepency between your favorite toy numbers and mine. It looks like you were including players 2009 wins in the calculations and their current ages, rather than those at the end of 2008.
The FT calculator I was using from ESPN also said to use age as of June 30th of the previous year.
June 5th, 2009 at 3:44 pm
Zimcity: I did not include 2009 wins in the calculations, as I actually did them back in January (my list still had Mussina with a 31%/33% chance of 300 wins). One possible reason for the discrepancy is that I use a player's age as of July 1 of the given season; 7/1/2009 in this case. I decided long ago (I've been playing with this - along with the 500 HR club - since the 1980's) that it was fairest to use the age a player was for the majority of the given season. Why should a player who turns a year older two days into the season be counted as the younger age? Or conversely, a player who advances a year in September be counted as a year older? If using an off-season cutoff date, like January 1 for example, why should a player born January 3 get the benefit of being a year younger than someone born December 30, when they're playing at the same age during the season? It seems to me that if you’re age 29 for more days of the playing season than you’re 30, you should be counted as 29; but if you’re 30 for more days than you’re 29, you should be counted as 30. Since determining the true midpoint of each season is needlessly tedious, I decided that 7/1 was a reasonable approximation, and quick to check. Thus, I’ve been using a player’s age as of July 1 in all my calculations over the past decades.
Not everyone agrees with this, but that’s okay – it’s just used to get approximations anyway. 🙂
June 5th, 2009 at 4:30 pm
I was looking at this on the SnakePit today, and one of things that stood out was that more pitchers are playing into their forties, and playing well too. Through the Play Index, I found that 35 forty-something pitchers have had 15 win seasons. But thirteen of those have come in just the past six years, compared to the mere eleven who reached the mark during the game's entire first century, from 1871-1970.
You can't expect everyone to pull a Jamie Moyer, but odds are that some of the candidates will work, and work effectively, past their 40th birthday. Johnson had only 246 wins at the end of his age 40 season,
June 5th, 2009 at 4:40 pm
When I do Favorite Toy calculations, I try to make the age as accurate as possible. I do the age as of Jan. 1 (middle of the off-season, assuming most calculations would be done through a completed season) and if the player is 31 years and 6 months old then, I take his age as 31.5.
Anyway, anal as I am in that regard, as we've discussed before I don't think the Toy is all that useful for predicting 300 wins. The pitchers who racked up the most wins in their early 20s flamed out. Guys who get to 300 do it with big performances in their 30s, which aren't really possible to project. Randy Johnson is just an extreme example of that. Still though, it is fun to try making the predictions.
Would be interesting to see what chances the recent 300-win crowd had when they were 30 or so. Johnson of course would appear to have no chance. I think Clemens might appear to have no chance after '96 (though a couple years later and a couple years earlier he would have looked better).
June 5th, 2009 at 6:27 pm
Thanks for the explanation Tomepp. I assumed incorrectly based on plugging in Sabathia's current numbers and coming up with the same percentage as you had.
I'm sure I have probably played around with exact age over the last couple decades too. I think my '84 Abstract is somewhere in my mom's basement. 🙂
June 5th, 2009 at 9:05 pm
Hey, don't worry; I take no offense. This is all fun and games. Speculation is part of the fun of baseball. Let's look at how Oswalt, Buehrle, Santana, Beckett and Sabathia stack up against a few other 300 game winners...
*Age 31
**Roy Oswalt, 131
**Early Wynn, 121
**Gaylord Perry, 118
**Randy Johnson, 99
*Age 30
**Mark Buehrle, 128
**Johan Santana, 116
**Early Wynn, 101
**Gaylord Perry, 95
**Randy Johnson, 81
*Age 29
**Josh Beckett, 95
**Early Wynn, 83
**Gaylord Perry, 76
**Randy Johnson, 68
*Age 28
**C.C. Sabathia, 122
**Don Sutton, 120
*Steve Carlton, 117
**Tom Glavine, 108
**Nolan Ryan, 105
**Carlos Zambrano, 99
**Early Wynn, 72
**Gaylord Perry, 60
**Randy Johnson, 49
June 5th, 2009 at 9:44 pm
By the way, my first vivid memory of Randy Johnson was the night he pitched his first no hitter in 1990. They played portions of his post-game press conference on the news (tv or radio, i forget now), and Johnson was nearly in tears, not exactly for pitching the no-no, but because (as he explained) he and his catcher (Scott Bradley) were in perfect sync all night. Johnson said that they were in tune with each other, like reading each other's minds. And I thought, wow, I love this guy! Sure I've seen baseball players cry (Mike Schmidt's retirement press conference, for example) but not for having a spiritual experience on the field. Can anyone dig up that press conference?
I was so impressed to see him refine his pitching through the years, from under 2 K/BB a season to over 4 and sometimes over 5 a season...climaxing at 6.59 at age 40! In my book, and quite obviously, he was a lock for the Hall of Fame by the end of 2002.
If you get a chance, listen to Johnson's postgame interview from last night, or from his interview today on ESPN radio. The guy loves baseball, and baseball history.
June 6th, 2009 at 1:44 am
Regarding speculation: I remember going to the Kickoff Classic football game between Syracuse and Wisconsin in 1997. Wisconsin's Ron Dayne was being looked at as a candidate to break the all-time career rushing record. He got hurt, and didn't do much in that game. His first game of his sophomore year, and already written off for the career mark! He bounced right back. Sure enough, he did break the record, and still holds it.
Tomepp, you're right--Gooden would have seemed the surest bet back then. And you can do that for a million different cases. It's like the stock market. "Past performance is no guarantee of future results." Especially when you talk about young guys, or peak guys who could either run into injuries and sudden breakdowns(Griffey, Thomas, Puckett, Mattingly and a million pitchers) or get a late surge(Tony Phillips, Dwight Evans, Charlie Hough).
June 6th, 2009 at 9:28 pm
To call RJ a better pitcher than Ryan is foolish... just because Ryan has a lower W-L % does not make him worse. His ERA is a better indicator, as a great pitcher could play for a horrible team, and lose. Nolan Ryan's statistics (that do not rely on his team's offense) show that he has done a better job.
June 7th, 2009 at 8:12 am
Nolan Ryan has a career ERA+ of 111 and Randy Johnson has a figure of 136. So Johnson is WAY ahead of Ryan in that area.
To suggest using ERA, and not ERA+, is itself foolish, and moreover, you are wrong.
June 8th, 2009 at 2:53 pm
JohnnyTwisto’s and other comments got me to thinking; there is no research that I’m aware of that tests the accuracy of the “favorite toy”, Pythagorean, or other methods we like to use for these types of predictions. If we took 100 arbitrary pitchers with a 10% “favorite toy” probability of 300 wins, then theoretically about 10 of them should have actually won 300 games in their career.
While perhaps the numbers don’t put an accurate percentage probability on the event, they do seem to be useful in comparing players. Based on how the formulae were developed, it seems rational that a player with a 30% favorite toy rating would have a better shot than a player with a 10% favorite toy rating. (Though even that has not been tested to my knowledge). But is he three times as likely? I’m not convinced that either favorite toy or Pythagorean method results are linear.
One problem with both of these methods is that the unadjusted “remaining years” for a player top out at age 40. A twenty-year-old player is projected to play to age 32 (or at least, produce the equivalent number of wins as 12 years at his current established performance level). Each additional year that the player is still in the majors adds 0.4 years to his “retirement age.” Thus at age 39, the player projects to play to age 39.6, and at age 40, the player projects to play to age 40 – no additional production. While some (myself included) have added “fudge factors” to account for players already playing past 40, they generally don’t kick in until the player has already reached 40 (or near 40). This would seem to be more problematic in predicting events – like reaching 300 wins – that typically occur at or past the age of 40. Seven of the last ten players to reach 300 wins did so on or past their “age 40” season (by my method). If we’re looking at whether a pitcher will reach, say, 150 wins, topping out at age 40 would not generrally be a consideration; but when we’re looking at the extremes, like 300 wins, then a formula that does not account for the ages when the majority of players who do reach the milestone accomplish it, that is a problem.
All that said, it’s still fun to play with these. I looked at the favorite toy and Pythagorean probabilities of last ten players to reach 300 wins for each year of their careers. I don’t know how to post a spreadsheet (or CSV file) on this forum, so I’ll give a synopsis.
Randy Johnson never had a chance at 300 wins. The favorite toy gave him 0%, and the Pythagorean method less than 10%, until age 41, when the “post-40 adjustment” kicked in. Even then, he didn’t break 20% until last year, and at that time he was just 5 wins away.
The Unit’s numbers were most similar to Phil Niekro’s, who also had 0% (toy) and <10% (Pythagorean) until his 40’s, and less than 20% until the year before he actually reached the milestone.
The guy who most consistently had the highest percentages by age was Greg Maddux. His toy percentage was in the teens from 1989 to 1991 (ages 23 – 25), then in the 20%’s and 30%’s until 1998 (age 33), then steadily climbed each year until he reached the mark in 2004. Of course, having 17 straight years of 15+ wins helps by keeping your established performance level from dropping radically.
Most of the rest had rises and dips, based on good and bad years affecting their established performance levels. Tom Seaver and Roger Clemens both looked promising early on (toy percentages in the 30%’s in their mid- to late-20’s), but both actually dropped to 0% before rebounding later in their careers. Tom Seaver actually reached 43.2% at age 34 before dropping to 0% three short years later. His percentage didn’t go back up until 1984, when he was just 12 wins away. Similarly, Clemens’ toy percentage reached 32.5% in 1992 at age 29 before falling to 0% just two years later. His toy value started climbing again in 1997 (age 34) when he put together back-to-back 20-win seasons.
Half of these ten players’ toy percentages did not go above 50% until the year before they reached 300 wins. Only Maddux (240), Carlton (249), and Perry (279) had fewer than 280 wins at the time their toy percentage rose above 50%. That means that most of the time, the player has to already have over 90% of the target value to have a 50-50 chance by the toy method. Here are the (age), win totals, and toy percentages two years before each reached the 300-win plateau:
Perry (41): 289 W, 96%
Carlton (36): 262 W, 61%
Seaver (38): 273 W, 0%
Niekro (44): 268 W, 8%
Sutton (39): 28 W, 44%
Ryan (41): 273 W, 9%
Clemens (38): 280 W, 75%
Maddux (36): 273 W, 93%
Glavine (39): 275 W, 20%
Johnson (43): 284 W, 48%
Finally, here is the answer to JohnnyTwisto’s query; each of their win totals, favorite toy, and Pythagorean probabilities after their “age 30” season (by my definition):
Seaver: 168 W, 31.1%, 32.9%
Maddux: 165 W, 23.3%, 26.9%
Sutton: 155 W, 21.7%, 25.7%
Ryan: 141 W, 16.0%, 21.8%
Clemens: 163 W, 13.5%, 20.2%
Carlton: 148W, 9.2%, 17.5%
Glavine: 139 W, 5.9%, 15.6%
Perry: 95 W, 0.7%, 12.9%
Niekro: 54 W, 0.0%, 9.6%
Johnson: 81 W, 0.0%, 8.3%
It would be interesting to compare these to a group of “Goodens” promising stars who faded out before reaching 300 wins.
June 8th, 2009 at 3:07 pm
Interesting post TomEpp. Re expected years remaining for older players, I recall that the Favorite Toy assumes any regular still has 1.5 years remaining, even if he is calculated to have less.
What is the Pythagorean projection method? Is there a link, or can you give an easy description on how it is calculated?
I've often thought about trying to a) determine whether the Toy actually has any accuracy and b) trying to make a better model. Have never gotten around to it though. I suppose it's not worth spending that much time to try improving it when there are people who have developed complex projection systems like PECOTA or ZIPS.
June 8th, 2009 at 4:01 pm
That's some great data and analysis, Tomepp. What it means to me is that a player with 20-30% who is still under, say, age 32, does have a decent shot.
What I would really love to know is how the FT has performed historically for ALL players--not just ones who did eventually make it to 300 wins.
Also, with regards to age, we went through a period in the late 90s and early 2000s where lots of guys routinely played until 40, and even up to say 43--a much higher number did that than usual. Now, with steroids use presumably well into decline, the number of guys doing that is shrinking again. The FT was developed a number of years ago and doesn't take into account, I don't think, how the average age of players changes through the years.
I would to know historical FT 300-win values for Blyleven, T John, and others in that general category.
June 8th, 2009 at 4:01 pm
I would **LOVE** to know.
June 9th, 2009 at 7:30 am
JohnnyTwisto: I do use the 1.5 years remaining adjustment (that’s the “fudge factor” that I mentioned), but that only kicks in once a player is 38 years old. And I’m not so sure that that is accurate either – surely at some point we must assume that even if a player plays more than one additional season, his production will continue to decline and his total remaining wins will be less than 1½ EPL-equivalent seasons. Ideally, the “fudge factor” itself should diminish as the player ages.
As to the Pythagorean method, it uses the same established performance level and projected remaining years and wins as the favorite toy, but uses the formula (Remaining wins)^2 / [2 * (Wins needed)^2] to predict the probability instead of (Remaining wins) / (Wins needed) – 0.5. It is based off of the Pythagorean method of win-loss projection. I don’t remember anymore whether I “created” this or whether I saw it somewhere else. If I did, indeed, “create” the Pythagorean method for projecting likelihood of achieving X widgets, let me unconditionally state that the idea was borrowed from other methodology – it was not an original idea of mine.
June 9th, 2009 at 7:50 am
One advantage of the Pythagorean method is that it never goes to 0% (unless the projected remaining wins goes to 0%, which only happens if the player has had no wins in the last 3 years). With the favorite toy, if the projected remaining wins falls to less than half the wins needed, the projected probability falls to 0%. And it theoretically becomes a negative percentage beyond that (you have to cap it with a MAX function). With the Pythagorean method, no matter how small (> 0) the projected remaining wins are and how many wins are needed, there is still some miniscule chance of success.
Both formulae have to be capped at the top end (with a MIN function) or they could produce percentages larger than 100%. With the favorite toy, this happens when the projected remaining wins is larger than 1.5 * the wins needed; with the Pythagorean method, it happens when the projected remaining wins is larger than sqrt(2) * the wins needed (~1.4 * wins needed). I use the 97% rule for both – the maximum percentage is 0.97^(projected years needed).