No, the goal of a pitcher is to prevent runs from scoring. Striking men out is one way to do that.
NOT WALKING THEM is another tool in not allowing runs to score. Ryan wasn’t so good at that.
511 wins. 749 complete games. 7300 innings pitched. 1 vote.
I was tempted to vote for Christy or Walter, but it’s really really hard comparing those guys to the modern era. I’ll go with Pedro, because he’s the Sandy Koufax of our generation, but probably even better. Perhaps the most useful rubric is comparing pitchers to their peers, and Pedro was throwing up 60’s-style numbers during the height of the steroid era. His #1 all time career ERA+ (for a starter) is a testament to that, and his 2000 season was simply ludicrous - an ERA of 1.74 when the average ERA was 4.92!
So I guess I’m leaning on the strong peak here, but even going by the full career, Pedro’s rate stats are still incredible. And if you buy into the je ne sais quoi reputation standard, anyone with an ounce of baseball sense knew that Pedro was the absolute king of the mound after Maddux’s mid-90’s run was over.
I also thought of Old Hoss, but with much the opposite point. The contemporary tendency to evaluate pitchers almost solely on pitching skills (counting pickoffs as “pitching” not “defense”) is not the standard used for most of baseball history. While most pitchers were mediocre or worse batsmen, they did appear in the batting rotation, and a pitcher who could hit as well as pitch was desirable. While you would always choose a pitching ace with execrable batsmanship over someone mediocre at both pitching and batting, when choosing between two journeymen pitchers to round out your rotation or provide you with a spare ‘bench’ pitcher, as bewtween two men with roughly equivalent pitching skills you would prefer the one who bats .170 over the one who bats .083.
By 1884, Old Hoss’s memorable season, it was no longer the slow-pitch-softball-style deliver-the-ball-to-the-plate type of pitching of the earlier 19th century. His job was no different than Cy Young’s or Tom Seaver’s: avoid giving up hits, try to get either strikeouts or balls hit where they can easily be turned into fielding outs. Skills and techniques may have evolved and improved since then, but the object – the logical basis for comparison – has not. And he was the Greys’ sole starting pitcher, pitching nearly every game for the latter 2/3 of the season to get his 59 wins to take the NL championship. Then in post-season play against the AA pennant winner – the only reason it was not termed the World Series is that no marketing genius had yet come up with the term – in a best-3-of-5 series he pitched and won all three games. While we do not and cannot know how he would have performed with a modern pitcher’s repertoire of split-finger fastballs, sliders, change-ups, etc., by the one standard that has not changed – that his task was to deliver the ball in such a manner as to maximize outs and minimize hits – his performance was truly outstanding. Since the OP called for “all-time,” not “modern era,” he deserves at least honorable mention in any “best of” list.
However, looking at the candidates proposed in the poll and post-1900 play, and bearing in mind that batting ability did play a role for most of baseball history, even if secondary to pitching skills, I’d have to point to Walter Johnson as the best. Not only did he get the lackluster Senators wins with his exceptional pitching skills, but he batted at nearly a .250 average, with better than 1% of his at-bats being home runs. There were more than a few occasions when not only did he “win” games in the modern sense of being the pitcher who holds the opposing team to a lower score, but he also “won” the game in the other sense – he was the batter/runner who either scored or drove home the winning run for his own team. His only superiors in the category of pitchers-who-hit-well after 1900 were both converted to outfielder in order to use their batting skills daily – G.H. Ruth and Smokey Joe Wood.
One final amusing note, totally off topic for “all-time greatest starting pitcher,” belongs here. Some years back the Angels made it to the World Series with a spectacular relief pitcher who, though a good batsman in Little League and high school ball, had come up through the minors after the DH rule was implemented, and of course played for an AL team. The World Series that year was under NL rules – no DH – and, having been brought in in relief, he was due at bat. So his first at-bat as a professional baseball player, at any level, was in the World Series. And it gets better – he connected and hit it into the stands for a home run, setting what has to be the ultimate in unbreakable baseball superlatives: not only did he have a 1.000 batting average and 4.000 slugging average, but his only professional at-bat ever was a home run in the World Series. It sounds like the sort of thing they’d have happen in one of those wish-fulfillment baseball movies – but it actually happened. 
Which is precisely why I am a bit skeptical of these huge peaks we see in several 90’s/00’s pitchers-Pedro’s peak was stupendous sure, but right behind him were Maddux, Clemens, and Johnson. In terms of ERA+ (essentially league ERA divided by pitcher ERA, with adjustments for park etc.) these were peaks which have been historically rare, yet we get 4 of them all within a few years of each other. 
:dubious:
My hypothesis is that it is easier to exceed the league average like that in a big hitting era than it is in a big pitching era; in other words, if you put Koufax in that era, his raw stats won’t be all that much worse-he’s already pitching at or near the maximum limit possible-at that limit what the rest of the league is doing is pretty irrelevant (tho it still constitutes significant value-in these discussions I am more interested in ability however). Likewise drop Pedro into 1968, and he won’t be all that much better, simply because he too was at or near the limit beyond which it is simply impossible to improve any further.
In other words, if you assume that ERA+, at the limit with the best pitchers ever, scales across eras in a purely linear fashion, you are likely incorrect.
Does it not follow that not allowing balls put in play is the best way to prevent runs from scoring?
Looking at Walter Johnson’s stats, it’s pretty interesting that his ERA from 1920 (the end of the dead ball era) to the end of his career was probably a touch over 3. Nolan’s career ERA was 3.19, Maddux was 3.16 and Clemens was 3.12. So, not a lot of difference in allowing runs. I think Nolan Ryan is a perfectly justifiable pick for greatest of all time.
ETA: Johnson’s ERA in the live ball era was 3.33.
If Babe Ruth had continued pitching, we wouldn’t be talking about him at all. He was a really good pitcher, but he wasn’t great. He had two really good seasons in a five year pitching career.
He also got quite a bit worse at preventing hits when there were men on base. And he was horrible at holding runners on.
Of course, but now why don’t we just look at how good pitchers were at, you know, not allowing runs to score?
If you examine that, Ryan is not the best pitcher of all time. Forget comparing him to pitchers from different times; Ryan was not nearly as great a pitcher as Tom Seaver. Nobody at the time thought Ryan was as good as Seaver and there’s no reason, in hindsight, to think they were wrong; Seaver was better as preventing runs and winning games, full stop. If you adjust his ERA for the time in which he pitched he isn’t even in the same class as Maddux or Clemens. He’s not remotely close to Lefty Grove, Walter Johnson, Bob Gibson, or dozens of other pitchers who had admittedly much shorter careers so don’t get named here.
One of my favorite baseball trivia questions when younger was, “What all-time record held by Babe Ruth was broken by which New York Yankee in 1961 in a shorter time?”
Whitey Ford broke Ruth’s record for most consecutive scoreless innings pitched during the World Series.
Fair enough, though note that Pedro’s the only one to reach the top career-wise. Clemens is below Grove, Walter, and Walsh, while Maddux is well behind that. But I think I do agree with your premise, even though Pedro still stands out in this era, as his contemporaries only had ridiculous ERA+ over a shorter span.
If we look at WHIP, I think Martinez comes out looking even better. He’s right at the very top (#5 all-time, and some of those ahead are relievers), and beats Walter and Christy on pure WHIP already! And this is during the juicy steroid era, where teams are giving up 4.5 earned runs a game. By comparison, fellow same-era great Roger Clemens is at #91 all-time. Baserunners allowed is a pretty fair metric, and adjusted for era, Pedro can’t really be beat on that regard, no?
Sandy Koufax.
I don’t think he is overrated. His numbers represent what he was able to do, while he was able to do it. He might have dropped off if he played longer, but he might not have. You just have to judge based on the numbers you have.
He threw as many perfect games in 12 years as Young did in 22, Halladay did in 14, and Johnson did in 22. He won the Cy Young Award 3 times against the whole MLB, not just the NL or AL, and is the only person to win more than two unanimously. Only Nolan Ryan has pitched more no-hitters, and he had more than double Sandy’s amount of seasons. Koufax averaged 200 strikeouts per season. Of the top 40 career strikeout leaders, only Ryan and Johnson had more. He is also the only pitcher outside of the dead-ball era to win three Triple Crowns.
It did not happen, at least as described. The Angels have been in the World Series only once, in 2002, and no pitcher hit a home run for them in that Series.
The last AL pitcher to hit a World Series home run, and so far as I am aware the only AL pitcher to homer since the DH rule was introduced, was Ken Holtzman, who homered for the A’s against the Dodgers in the 1974 World Series. Holtzman did not have an at bat *during the 1974 regular season *but he had hundreds of at bats during his career, having come up long before the DH rule was created (and he came up in the NL, anyway) and had previously hit two home runs during the regular season. He was also the starting (and winning) pitcher in that game, not a reliever, hitting his home run in the third inning off Andy Messersmith.
I’m pretty sure you are mis-remembering some aspect of this story. The Angels made the World Series only once, in 2002, and none of their pitchers hit a home run in the three games played in the Giant’s park.
Edit: Maybe I should read the thread before responding.
You can get on base with a hit or a walk. Ryan is 271st all time in WHIP. 271st. He let people get on base. He is tied for 278th in adjusted ERA. He let them score. He was often (usually?) not even the best pitcher on his own team, and ended up with a mediocre win loss record even compared to teammates. He is clearly not a top ten starting pitcher of all time.
Pedro Martinez, Walter Johnson, and Christy Mathewson are all in the top 10 in WHIP. Pedro has the best adjusted ERA for a starter.
Strikeouts? Pedro actually has a better strikeout rate than Ryan, even beating him at his own game.
Baserunners are one of the major components of ERA (and ERA+), so my point would still stand. During their peaks, the baserunners/9 of Pedro and Sandy were:
Koufax 62-66 .926
Martinez 97-03 .940
Certainly Pedro’s mark here is more impressive, but what I am saying you can’t simply make 2 simple ratios (pitcher baserunners per 9 / league baserunners per 9) and then extrapolate that into another era. Picking two years out of those peaks (Koufax '64 and Martinez 2000, when both had identical 1.74 ERA’s), I get
Koufax .928
64 NL 1.269
Martinez .737
00 AL 1.49
If you were to extrapolate Pedro into 1964 (he allowed walks and hits at almost 50% of the league) you would get .628 WHIP. That isn’t going to happen-oh it might be .710 or something, but there’s really no way for him to drive that down much lower, no matter when or where he pitches (within reason). [Koufax in '00 comes out as 1.09, which likewise seems to high-Randy Johnson beat that 6 times during his 8 year peak.]
Hell being a Red Sox fan I was definitely psyched to watch the man pitch, so don’t get me wrong. But the stats were so wacky during the recent live-ball era that you have to be careful making cross-era comparisons, and simple math like the above probably doesn’t reflect the reality.
There’s also a certain law of diminishing returns.
Look at it this way; Pedro in his best ERA+ season, in 2000, started just 29 games. No matter how you slice it, it isn’t logically possible for a man who starts 29 games to help his team win any more than those 29 games. Pedro Martinez’s job is to help his team win games; we use ERA and WHIP and DIPS as a way of determining how much he helps his team win the games they did, but the fact remains he only pitched 29 games and so his contribution cannot, even in theory, go beyond that (and he didn’t pitch 29 complete games, and of cpourse wasn’t scoring the runs for his team or playing the defensive positions, so it’s really less than that.) Since Boston in fact lost at least six of them, he can’t have helped them win more than 23 discrete ballgames, and Pedro was not personally responsible for all of those wins. WAR awards Pedro 10.1 wins for 2000, which is quite amazing, and which to me seems logical given the other numbers.
By way of comparison, Sandy Koufax in 1966 with a comparatively lower ERA+ pitched 41 games, and in fact completed most of them. He pitched 50% more innings than Pedro did, so despite being by the ERA+ measure not as dominant, he was helping his team win games more. The WAR method says he was slightly more valuable. That makes perfect sense to me. You cannot discount the innings pitched. Pedro Martinez’s 2000 is NOT more impressive, in any logical sense, than Sandy Koufax’s 1966 (or, even more so, Steve Carlton’s 1972) unless one is prepared to say that the number of innings pitched isn’t relevant - and, again, if you say that you have to conclude that Tom Henke was a greater pitcher than Tom Seaver and all manner of other silly conclusions.
As amazing as Pedro was WAR places his 2000 season as not the best, or second best, or even third best season of the live ball era, and it’s simply because he didn’t pitch as many innings as the other guys, and logically that HAS to count, for the same reason they didn’t give Shane Spencer the 1998 MVP Award, despite his 236 OPS+.
In case anyone cares, the best seasons, by Baseball Reference WAR, of the post-WWII era by a starting pitcher:
Steve Carlton, 1972 - 12.2
Bob Gibson, 1968 - 11.9
Dwight Gooden, 1985 - 11.7
Bob Gibson, 1969 - 11.0
Sandy Koufax, 1963 - 10.8
Sandy Koufax, 1966 - 10.8
Wilbur Wood, 1971 - 10.7
Gaylord Perry, 1972 - 10.5
Roger Clemens, 1997 - 10.3
Dick Ellsworth, 1963 - 10.3
Bob Feller, 1946 - 10.1
Pedro, 2000 - 10.1
Yeah, I’m suspicious about so many of them being clustered around 1963-1972, too. But still; it’s hard to make a case Pedro’s peak was higher than a lot of guys.
Watch me! 
I’m quite aware that a) I’m probably being simplistic in general and b) I’m not really using counting stats in my valuations, including WAR, as long as they’ve pitched some minimum arbitrary amount. When I think of the term “best starting pitcher” I think of the guy who, if I could only get him at a random point in his career, I would most want out on the mound in, say, a playoff series. For me, the career rate stats in the context of that player’s era tell that story best.
This is distinct from most valuable starting pitcher ever, or perhaps starting pitcher who’s contributed most to the success of his team. There, playing more definitely bumps you up (as long as it’s a positive effect).
Kind of a joke Johnson is atop this poll. I couldn’t seriously consider anyone who played in the dead-ball and/or the segregated era. Its not the individual players fault, but they were competing in a diluted and inferior league, compared to the modern game.
Poly is notoriously unreliable on the subject of angels and wish-fulfillment.