The comment that Jim Leyland made about knuckleballer Tim Wakefield was "The problem with Tim Wakefield is that he's never mediocre."
And when Wakefield went to the Red Sox and worked on his craft some more he converted some of those disaster outings into mediocre outings and made himself a valuable pitcher.

There is a segment of MLB that simply doesn't see knuckleballers as legitimate. Torre was that way.

But I don't think it's any more true now than it ever was.

So, as a player, you have to constantly be proving that you can be a trustworthy piece on the chessboard.

Uh…yeah. I hate it when you have, like, an unreliable rook that you can’t trust not to start moving diagonally and stuff.

There is a segment of MLB that simply doesn't see knuckleballers as legitimate. Torre was that way

Scarred early by chasing Phil Niekro offerings to the backstop :)

I could see the value of somebody who can throw a bunch of innings as being very valuable. Extra innings, blowouts etc.

At a time when starting pitchers are being asked to cover fewer and fewer innings, having one pitcher on the staff who could throw all day if needed seems pretty awesome. You do what Dickey says in the excerpt, and find a pitcher who has a lot of the makeup of a successful major league teammate, maybe has expressed interest in trying the knuckleball out, and give them a few years to perfect it. If it takes, you let him be your swing starter/long man, and save your bullpen and rotation in blowouts.

I also think that a pitcher willing to do this would find, if they are moderately successful, that they'd be a prime candidate to be nabbed in the Rule V draft by a team willing to take a chance on stashing them in this sort of role for a year.

I mean, we all know about Wakefield, Candiotti, the Niekros, Wood, Hough, etc, but are there examples of knuckleball pitchers who just completely bombed in the big leagues? For that matter, are there very many who fail in the minors, and never get past AA or something? It doesn't seem like it. (The Red Sox did have one, Steven Wright, and he was actually pretty successful. They gave him one full year as a starter, and he went 13-6, 3.33 ERA, four complete games, 156 innings...then he struggled in early 2017, and although he pitched well again in 2018, he was pretty much off the radar screen for them.)

are there examples of knuckleball pitchers who just completely bombed in the big leagues?

Charlie Zink got lit up in his MLB debut in 2008,* and that was it for him. Not much of a sample size, of course. Mickey Jannis got lit up in 3.1 IP for the Orioles last year. But that reflects the reality that knuckleballers who bomb initially aren't around to bomb much longer.

*Edited to add that I remembered "that Charlie guy from a few years ago." Turns out it was 13. Sigh.

I'd extend the idea Dickey proposes to submarine guys too. I'd have a whole arm of development to work on knuckleball and sidearm/submarine pitching, and focus it on high-character guys who aren't going to make it otherwise. It would be a cheap development of potentially valuable arms

I mean, we all know about Wakefield, Candiotti, the Niekros, Wood, Hough, etc, but are there examples of knuckleball pitchers who just completely bombed in the big leagues?

Charlie Haeger wasn't very good

Jared Fernandez had a 5.05 ERA on 108.2.IP.

Steve Sparks had a few okayish seasons, but overall wasn't all that good (56-69, 4.88 ERA). 10.4 war in his career, almost half of it in the 2001 season, when he went 14-9, 3.65 for the Tigers.

#8 Craig Wright proposed this in The Diamond Appraised. Tom House gave what is basically the standard response. We can't do that because reasons.

Really what it boils down to is the the minor leagues are very much an up or out model and it's important to wash out loads of players every year.

Didn't make sense to me decades ago and makes less sense to me now.

Has anybody ever measured how well the starting pitcher does when he follows the knuckleball pitcher against the same team from the previous day in relation to his overall stats? There may be an indirect benefit which has never been calculated.

It's not clear "can eat up all your garbage innings" has much if any value. That's what the AAA shuttle is for and MLB teams don't seem to care whether those innings get spread across 2 or 4 or 10 relievers. While the number of relief innings keeps going up, the number of relievers who reach 60 innings in a season does not. It's been in the low 80s for nearly all of the last 15 years, there were more in the early 2000s. How do you throw 400 innings of relief (old days) to 600+ innings (these days) yet are lucky to even have 3 relievers reach 60 innings. The AAA shuttle has been going on for a long time.

In 2021, 43% of innings were thrown by "relievers" (a few of whom were starters following an opener but then some "starter" innings are really relief innings). It was an average of almost exactly 625 innings per team. Your top 3 (healthy) might cover 200 innings. That's 425 innings across the remaining 5 relief slots. Probably two of those guys are other "top" relievers for you throwing 100-110 between them, call it 300-320 innings across 3 relief slots. The mop-up garbage guys are there, there are just 10-20 of them ... having a knuckleballer on the roster all year just takes up one of those garbage slots which doesn't do anything but reduce the number of transactions, it's not clear it really helps you win.

The reduction in the size of the minors probably hurts the chances of a knuckleballer. Most knuckleballers are injured/failed prospects and teams should be less willing to burn a spot on a longshot experiment now. The next knuckleballer will probably have spent some time in the indy leagues unless maybe there's some genuine college fireballer who has been throwing a knuckler all his life out there.

That's a good point, there should be some indy ball knuckleballers. Anyone know of any?

Let's take a look at the White Sox. They had a pretty healthy and effective rotation and used only 548 relief innings.

193 -- Hendricks, Kopech, Ruiz (excludes 3 early-ish starts for Kopech covering 12 innings, look like legit starts)
110 -- Bummer, Crochet (their two lefties)
245 -- another 13 guys plus 3 position players (2 of those 13 were Kimbrel and Tepera)

Reynaldo Lopez seems to have been the modern equivalent of a swingman. He was hurt for the first half of the year, made 20 appearances, 9 of them starts (5 in Sept) but covering just 37 innings. He did OK giving up 18 ER (21 R) but 6/7 of those in one start. Looks like he had a 10th "start" following an opener (5 IP, 0 runs).

Give or take, you need a garbage specialist to take about 80 innings of relief and cover Lopez's 40 IP of starting to even save a roster spot. But if you've got a guy who can throw 120 innings effectively, that's now a #4 starter who might earn decent money in arb. (And if he's ineffective, what have you gained?)

The back end of the bullpen is just another example of the temp/gig economy -- minimum wage, high turnover.

What organization wants to spend years developing a knuckleball pitcher when they can get a fastball/slider reliever for minimum wage?

Oh, for the days of the 1945 Washington Senators, who almost won a pennant with four knuckleballers in the rotation...

#18 Umm Dutch Leonard played on the Crackers? That one won't be making a comeback me thinks.

I hate it when you have, like, an unreliable rook that you can’t trust not to start moving diagonally and stuff.

Such a drama queen.

So, as a player, you have to constantly be proving that you can be a trustworthy piece on the chessboard.

Uh…yeah. I hate it when you have, like, an unreliable rook that you can’t trust not to start moving diagonally and stuff.

Actually, this should be something that we can actually quantify.

Almost all of our rate stats are based on measures of central location (totals, averages, ratios of totals or averages). Should, especially for pitchers, we be looking at measures that balance average level against volatility?

We all agree having a pitcher who is consistently below replacement is unacceptable. But what about a pitcher like Leyland is implying Wakefield was (I don't know if this is true from the data or not and don't feel like looking right now, but just go with me): a pitcher who was amazing sometimes (let's say P% of the time) and a disaster sometimes (1-P%). Given levels for amazing and disaster, at what point is the value of that player higher than what we think of as replacement level (which, let's be clear, is mostly based on an average)? Is it P=2/3? Or is it 1/3?

It seems like win probability should help us to answer that question. At a certain point, will a pitcher who gives you half their games with 90% win probability performance and half their games with 0% win probability performance might be better than a pitcher who gives you the same 45% win probability peformance every time out? This may depend on the quality of the rest of the team! Someone must have asked and tried to answer this question already ,right?

I hate it when you have, like, an unreliable rook that you can’t trust not to start moving diagonally and stuff.
Such a drama queen.
--
he says in passing, this could go on all (k)night, mate.

While I'm no Beth Harmon, I'm pretty sure having a rook that could suddenly move diagonally would be really beneficial.

en levels for amazing and disaster, at what point is the value of that player higher than what we think of as replacement level (which, let's be clear, is mostly based on an average)? Is it P=2/3? Or is it 1/3?

That's a good question. If you think of starting pitchers, someone who would throw a 6-inning shutout 2/3 of the time, and completely blow up (say 6 runs in 2 innings) would be an example. That would give you a 3.85 era, so slightly better than league average but would probably lead to wins in 2/3 of the pitchers appearances, so much better than league average.

It's similar to the HOF arguments of "peak" versus "career", peak=pennants is the argument, and in this case the argument (in-season) is shutouts=wins.

back end of the bullpen is just another example of the temp/gig economy -- minimum wage, high turnover.

This points out once more that the economics/analytics-based strategy/roster rules are biased in favor of having a whole bunch of FFT's (fungible flame throwers), which means all sorts of fun things about baseball are becoming rare if not extinct, including knuckleballers.

How many teams actually have enough fungible flame throwers (FFTs) to make this work? Because what I see coming and going on the minor league shuttle is a lot of flotsam and jetsam. Throwing hard does not equal getting people out.

I may be misreading Doug's comment, but isn't that failure consistently to get people out what makes the FTs fungible?

#24-26: Right, it's more the low leverage of the innings pitched than the quality of the pitcher. You have a lot of those innings, you don't want to use quality pitchers in that role, almost no pitcher (other than a knuckleballer) can withstand a regular diet of 100+ relief innings a year but those innings still gotta get pitched. So a relief "slot" split among several pitchers over the course of the season fills those innings. So you bring a guy up, have him pitch 9 innings of meaningless relief over two weeks, send him down for the next guy. You cover 117 innings of (predominently) garbage time with 5 pitchers. You could do it with one knuckleballer but it's one roster slot either way.

I've wanted to do an "anatomy of a bullpen" thing where I track some team's bullpen usage and transactions throughout a season for years but I'm too lazy to do it. Somebody could probably come up with a cool "Napoleon's march" sort of stat graphic for it but even a stodgy color-coded Gaant chart would probably work.

#21: I'm pretty sure I've seen stuff on that, at least in a back of the envelope way, like how we do "would it be better to just walk Bonds every time?" But mainly, that's what real-world baseball performance almost always looks like. You get occasional metronomic freaks like prime Pujols but most good seasons amount to something like 3 excellent months, 2 average months and one crappy month. Averae seaons are 2/2/2; bad seasons are 1/2/3.

Vlad Jr just had an awesome year. Still, in 43 of his games, he went hitless; another 29 where he went 1 for 4 or 5 without a HR (he seemed to have a lot where his only hit was a HR). (Some of those latter would have had walks to raise the OBP to a solid level or a double giving a decent SLG.) So that's almost half a season ... or at least a third of it ... when Vlad's lack of offense hurt his team. The key think is that's better than, say, Hosmer who went hitless in just over 1/3 of his games and had more than one hit in only just over 1/5. No surprise really for a sport in which reaching base say 36% of the time is really good.

Pitchers are more likely to be consistently "good". Quality starts used to be a pretty decent way to track that -- yes some quality starts aren't so great but there really aren't many of those -- but now with pitchers, especially the non-elite, not making it to 6 innings all that often, it needs to be redefined. But if we go back to 2010 (why not), 53% of all starts were QS so presumably the elite guys were around 80%.

I noticed the phenomenon you're considering of Shawn Estes' 2003 season with the Cubs. He was the #5 starter on a top staff. He pitched like one with a 76 ERA+. Still of his 28 starts, 11 were QS (40%). In those starts and a few other pretty good ones (e.g 7 IP, 4 ER or 5 IP, 2 ER) he gave up just 25 ER in 96 IP -- that's a 2.34 ERA in exactly half of his starts, 8-4 record (team 8-6). But that means the other half of his starts were 56 innings and 72 ER. He ended up 8-11, the Cubs 12-16 in his starts, probably about what you expect for a #5 starter on a good team.

I assume, but don't know, that Jekyll/Hyde performance is pretty common for #5 starters. The only real difference I see between that performance and one where Estes threw 5.5 innings and 3.5 ER every start (neat trick) is that the bullpen load is more stable in the latter ... but hard to say whether that's preferable to needing either 2 or 6 bullpen innings whenever Estes took the mound. That sort of start would obviously require the Cubs to score at least 4 and probably 5 to win (the bullpen will give up at least one run on average over 3.5 innings). The 2003 Cubs did that 41% of the time suggesting they would be expected to win 11 of consistent Estes 28 starts instead of 12.

In short, it probably doesn't matter. Maybe a Jekyll/Hyde #5 with a 76 ERA+ is as valuable as a consistent #5 with a 84 ERA+ but that's probably all there is to it.

It's probably the same on the other end. Randy Johnson's 2000 was probably the worst of his Cy Young run but it was still a Cy Young, 181 ERA+ season. He made 35 starts, only 25 of them were QS, 71% (career 67%). So right there, 1/3 of Randy Johnson's starts were "worse" than 6 IP, 3 ER. In those 10 starts, he still managed 61 IP but 45 ER and 6 UER. Arizona went 3-7, Johnson went 3-5, winning starts of 6/4, 9/4 (really a QS) and 7.2/4 with his teammates providing 21 runs of support. I'm pretty sure most teams would be willing to tolerate 10 lousy starts a year from RJ -- the rest of the season was 189 IP, 28 ER, 38 UER.

Corbin Burnes just won the CYA with 18 QS in 28 starts (64%), the times we live in but not substantially worse than Johnson 2000. He also never gave up more than 5 runs (in a 4 IP outing). Two of the non-QS were of the 5 IP, 1 ER variety so quality by today's standards. Anyway, we can say he had 8 bad starts (29%) totally 37 IP with 30 ER plus 2 UER. He went 0-3, the team 4-4 which is good.

Anyway, even a CYA season might well feature 1/3 bad starts with an ERA over 6 across those starts (and 2/3 great). Presumably a good #2 guy is more like 60/40 excellent/bad.

Rivera 2001 appeared in 71 games with a 192 ERA+ (below average!). He gave up at least one run in 13 games so about 20% -- 15.2 IP, 24 R, 21 ER. What a bum! The Yanks still managed to go 5-8 in those games, he went 3-6 with 2 saves and 7 blown saves. Of the 4 non-save situations, 3 were extras at home and 1 tied 9th at home so there was never going to be a save situation.

The key characteristic of the FFT is that he/she cannot throw very many innings, either cumulative or sequentially, without his/her effectiveness declining dramatically. Hence a great candidate for the major league/minor league shuttle. If the use of the "shuttle" was restricted, then perhaps a knuckleballer (or in fact any pitcher that can throw more than once through the order) would be more highly valued.