Tom Glavine’s first 72 starts yeilded 4 K’s per 9 innings. His career average of 5.3 is well below the league average. He was pretty good.

Interesting read, Evan. Thanks for the morning coffee entertainment! I have a lot of respect for Bill James, primarily because he actually backs up his arguments with research and numbers. He challenges the old fashioned way of doing things whereby one just “knows” or “feels it in his gut”.

However, I do have one problem with James’ argument here on K/9. Now, as a disclaimer, I should note that I did not yet read the article and am just basing my thoughts on your summary.

My problem is that basically there is no “theory” here. In other words, there is no attempt to explain WHY a K/9 rate below 4.5 equals failure. Why should that be the case? Strike outs are just one way to get outs, so it doesn’t logically follow that one must strike out a lot of batters in order to be successful.

But the numbers don’t lie. We have without a doubt an enormously strong correlation, but as probably anyone remembers from school, correlation doesn’t necessarily equal causation. Historically, as women’s skirts rose (got shorter), the US stock market also rose. Of course, no one thinks one of those things caused the other.

I am very open to believing that James’ idea, but would like to read his reasoning behind why the correlation exists. Often, that’s what is missing from the so-called sabermetric approach. It is entirely possible that there is another factor (or numerous other factors) that have more to do with long-term success than K/9, despite the correlation between the two.

Thanks for the really good research and writing Bob. Does one end up sober or pessimistic when you read something like this? Hard to manage sometimes, but impossible to argue with the conclusions. Thanks for provoking the thoughts!

I like the comparison and think the similarities should not be discounted. Is it possible that Feldman>Drese simply because Drese was backed by an infield of Soriano, Young @ SS, and Blalock @ 3B? Meanwhile, Feldman is backed by Kinsler (much improved defensively in 09), Andrus (a SS with range!), and Young (a better 3B than SS).

@beerguy: As I understand it, K’s are viewed so highly by sabermatricians because they are the only way for a pitcher to record an out that is completely under his control. Any batted ball has a chance to fall in for a hit (i.e. – see Cabrera’s ridiculous hits in his first 2 AB’s last night). Pitcher’s with higher K/9 rates reduce the variability and risk of balls falling in. With a low K/9 rate, James’s argument is that eventually all of those batted balls start finding holes, and thus the pitcher’s success drops off.

Here’s hoping Feldman is the exception, because he sure has been fun to watch the last 2 years.

There’s a phrase for what Feldman is doing – it’s called “done with mirrors.” And I really don’t mean that as an insult to him. It’s more that Feldman relies heavily on being crafty rather than overpowering. You can say the same about Maddux, Glavine, etc.

It also means he has much less room for error than other pitchers with “stuff.” He may very well not be a part of this rotation in 2010 and beyond. But you ride him as long as he produces these results.

Interesting read, great analysis Bob! And a good point by SDM as well. I’d add that Feldman is clearly still developing as a starter. He’s added a cutter this year, which has been a tough pitch on lefties and he’s starting to flash signs of a solid breaking ball. When he gets his K’s, its rarely with his cutter or sinker, but with that developing slider. It would follow that as his slider progresses, his K/9 would increase, Just an optimistic theory…

Thanks Clint and JRB, that makes sense to me. While I do recognize the importance of isolating stats that show a single player’s talent/value (as far as possible anyway, even called K’s are arguably influenced to some degree by the framing of the catcher), I also wonder about the ultimate utility of this. Assuming a solid defense behind them, why couldn’t a pitcher with a low K/9 rate succeed? Well, he could, although perhaps not over the long haul (the iron law of averages?). Maybe all this says is that these kinds of pronouncements work best over the sum total of a career.

Check out Joe Niekro, 221 career wins, lifetime strikeout rate of about 4.3/9. And when he started it was around 4. I think what this correlation says is more about the belief that pitchers have that to be a good pitcher you have to strike guys out and that’s what they focus improving upon. And as far as Feldman not being overpowering, I’d say there’s a huge pile of broken bats somewhere that would disagree with you…

Great thought-provoking stuff — and yet another strong reason to pull for Feldman, the pitching staff’s biggest underdog. Thanks, Sports Sturm!

@mike:

Great point.

If you’ve ever tried to hit a knuckle (or catch one), you’d know that it’s not hard to hit, but rather ridiculously difficult to get solid contact. It’s not a stikeout pitch, but it’s super effective none-the-less. Cutters and sinkers, especially in the same repertoire, do the same thing. Rivera strikes out a lot of guys because of amazing control – to a righty, he starts it on the outside and forces guys to swing a miss, and to a lefty he makes forces guys to hit it with the skinny part of bat.

Feldman will likely get even better, once he truly commands that cutter – and will be very effective if hitters know that a slider or two-seamer might come at them.

funny to bring up Mariano … Feldman’s cutter is pretty darn good, not in the same leagues as Marianos (who’s is) … but it continues to get better and better …

his breaks just enough to cause a ground ball while Mariano’s breaks enough to cause a swing and a miss … both pitches cause troubles for hitters …

I like what Feldman brings to the table … he is 27 and the dude knows how to pitch …

This is when I think you can overanalyze stats. To me, you have to watch the games. Might he flame out? Sure, but I don’t have grand ideas that Feldman is a top of the rotation guy. I think he is a good 3-4 starter, and he’s showing to be that.

I think that small point about Holland is a very good case why he should not be traded.
Also id have to ask Mr. James if he has seen a pitcher produce as many weak hits as Scooter.

Always fun to see someone start talking about Bill James and then go on to cite wins to talk about a pitcher’s success

Derek Holland’s SO/9 may be 7.0. But his SO% is only 17.2%, which is actually below league average.

Alright, there are a couple of important facts to bring to the table about Scott Feldman and the issue of strike outs.

First, there are many pitchers currently in baseball striking out a comparable amount of hitters, yet having pretty serious success.

Joel Pineiro
Jason Marquis
Aaron Cook
Nick Blackburn
Joe Saunders
Zach Duke
John Lannan

These guys are having success while having K rates below 5.

Second, I think that that commonality you see amongst successful pitchers with K/9 rates below 5 is that they all have one really good, plus pitch, like a sinker or cutter. These plus pitches are designed not to get swings and misses, but to get weak contact. FIP and tRA and other stats are really useful, but they have some noteworthy pitfalls, such as they too often value all contact as equivalent. Obviously, if you have watched Feldman this year, it is obvious that he has done a tremendous job getting the cutter in on the hands and breaking bats, and running the sinker low and away to induce ground balls. Even though he allows a very high rate of contact, he nonetheless should be expected to allow a BABIP that is significantly lower than an average pitcher.

Third, people cite the significant drop in BABIP for Feldman as solid proof that he is due for a major regression. But looking at fangraphs.com, you can clearly see that Feldman has dramatically changed his pitch selection this season; namely, he has added a cutter that he now throws almost 1/3 of the time. Further, that cutter is rated as a well-above average pitch, close to 15 runs above average this season, according to fan graphs. I don’t think that a .240 or .250 BABIP is completely sustainable. Sure, Feldman has been helped substantially by his defense, but lets also give some credit to fact that he has found a new pitch that induces a ridiculous amount of weak contact and results in a lot of easy outs.

Fourth, I think people who religiously follow most advanced statistics (like myself) have a tendency to discount a players ability to improve with experience. Look at Edwin Jackson. I would have been the first to say that Jackson was not a very good pitcher last season. His FIP and tRA were very high, and he didn’t have very strong peripheral numbers like K/BB. But this season, he has taken a major step forward, which no baseball statistic would have indicated possible. He had enormous talent and has really started to control all of his pitches. I see no reason why Feldman, a guy who has only been in the rotation for 1.5 years professionally, cannot continue to improve on his breaking pitches and refine his cutter. And I think he has every ability to increase his K numbers and lower the walks as the years go by.

Feldman is a power guy. Power sinker, power cutter. Those pitches are designed for contact. But just because he isn’t getting the KOs, doesn’t mean he isn’t “power”.

Nothing wrong with pitching to contact if you are having success. I’m pretty sure it’s been documented for the Ranger pitchers to do just that.

Lamar – the big issue is that K/9 rates seldom actually go up for big league pitchers. Many, if not most, actually see the K/9 rate at its highest in the first 3-4 years of their career. Then they slowly decline. So, if your base rate is at the bottom of the ladder, then you generally have bigger issues when you start to fade. Does that make sense? Glavine may exist, but there are not many who “figure out” how to strike out ML batters later in their career. That is generally a myth.

[...] The highly anticipated double-A debut for Martin Perez was one to forget. The 18-year-old struggled from the start, allowing eight hits, a walk and seven runs (six earned) in just 2.2 innings of work. He also had two baserunners steal second on him and fired a wild pitch that allowed a man to move to third. The lone positive to come from Perez’s outing were the three strikeouts he recorded, preliminarily establishing his Frisco K/9 ratio at 10.1. And we all know what Bill James thinks about high K rates. [...]

Here’s a problem with the numbers as they apply to Feldman. This is a guy that is just 2 years removed from throwing sidearm. He’s just now getting comfortable with his secondary pitches. Obviously, Bill James stats do not take that into account. Sometimes you have to think just a little bit more about what you’re writting before your write it. Feldman’s secondary offerings actually look to be legit major league pitches, but he’s just gaining a comfort level to throw them in any count.

His strikeout rate will rise above the mythical 4.5 k/9 inning mark once he gains that confidence in his secondary offerings. And let’s remember, this is a guy that touches 93-94 on the radar gun. It’s not like he’s throwing an 87-88 mph sinker up there with the majority of his pitches.

[...] curious point of Feldman is his K/9 rate in August. Given my obsession on the topic on July 29, it is amazing that he doubled his K rate in August. Is it a anamoly? Is it proof Feldman reads [...]