1
I'm sure a small % of the hit prevention was thanks to the pitchers, but largely its going to be the D.
How about 2003?
Did the 2001 Pitching Stars CAUSE the Great DER?
Intuitively, is it possible that the 2001 Mariners had a "below average" rotation, one that offset their sensational relieving to cause a net "average" effect? When you just look at these guys like a kid looking at baseball cards... , were they Felix plus four lousy pitchers? Were they AAA pitchers filling in for injured stars (as many ML staffs are every year)? Were they four mediocre pitchers and a guy with 82 walks, 69 strikeouts, and a 1.7 flyball ratio in the #5 slot?
.......................
1 Jamie Moyer is becoming a fringe candidate for the Hall of Fame. In 2001, he was in his prime.
2 Freddy Garcia, when young, was one of the best pitchers in the league, with one of the best K/BB/HR lines. In 2001, Garcia had the very best K/BB/HR outcome of his career. 2001 was Garcia's #1 season.
Garcia was a young star and he was at his absolute peak in 2001.
3 4 5 Aaron Sele, John Halama, and Joel Pineiro were demonstrably at least average pitchers (with very solid K/BB/HR lines). For example, Sele had just finished good seasons with Texas, and Pineiro was about to become one of the league's most promising young pitchers.
In fact, nobody could call the 2001 Joel Pineiro average. He was breaking into the league, he was throwing great, and the next year (2002) he went 14-7, 3.24 with a 136/54 control ratio. Joel was probably even better than Freddy and Moyer in 2001-02.
Paul Abbott was either average, or below average, in terms of pure pitching (Abby in 2001 had a 93 fastball and a plus-plus change, but control problem). .... assuming for a moment that he was below average, THIS ...
When a team has average pitchers in the BACK of its rotation, that's not below average and it's not average. It's better than average.
1 Star
2 Star
3 Solid
4 Mediocre-below average
5 Young star
6 Solid
... is not a "below average" rotation. It is a rotation the $100M Mariners have been trying to reproduce for eight years, without a whiff of success.
A "below average" rotation has one fairly good pitcher, one average pitcher, and three problems. Or it has a bunch of guys injured and is filling in with its AAA guys. You might have noticed this lately. :- )
...........................
The 2001 Mariners' pitching staff had two aces, a solid back-of-the-rotation, and a sensational bullpen. They didn't score their 118 ERA+ by putting 14 Jarrod Washburns in front of three center fielders. They scored it by having a pitching staff we haven't seen since.
............................
How good would Moyer, Drop Dead Freddy, Jo-El, Sele, and the 10K bullpen have been with a solid defense rather than a great one? I dunno, 112 ERA+ instead of 118 maybe.
But I doubt it.
Frankly, I'm kind of surprised that such an outstanding 10-man pitching staff, with that defense, only scored a 118 ERA+ (Oakland was at 121). I'm pretty sure that part of the DER was caused by the excellent pitching. I think a whale of a lot of guys hit weak batted balls off Rhodes and Nelson.
I mean, in 2002, the Mariners had the same guys playing defense, including Cammy in CF and Ichiro in RF. They even swapped out their one weak link, Al Martin in LF, for Mark McLemore. What happened to their "eerie" DER then (which was good, but not as good)? Maybe subbing in James Baldwin and Ismael Valdes, and Jeff Nelson's injuries, hurt their DER?
Why should it be an absolute that, of the two components of run prevention (fielding and pitching), that fielding only is the cause of DER?
Something weird happened in 2001 to cause that convergence in DER, and at least part of it was outstanding pitching (including those jaw-dropping 7th, 8th, and 9th innings). I'm not sure what it was, but it underlines ... in my mind ... how little we understand defense.
It's not like the Mariners were able to reproduce that 2001 DER. And so I'm not very confident that anybody else is going to be able to predictably reproduce it, either. :- )
Cheers,
Dr D
Comments
2
Doc...
Jamie Moyer is not what I would call a star. He was a very good pitcher in 2001 (and overall of course) but he's certainly not an ace...he's the kind of guy you like to have #3 in your rotation. Chief Garcia had a good year in 2001 and DNRA backs this up...but the entire rest of the Mariner starting staff with the exception of Joel Pineiro's partial season was quite poor.
DNRA+ figures for that rotation go like this:
Garcia: 121
Moyer: 111
Sele: 96
Halama: 75
Abbott: 84
Pineiro: 175 (but in just 70-ish innings)
Doc...I would be inclined to accept the possibility that the Mariners' DER in 2001 was caused partially by great pitching *IF* the DNRA figures looked anomalous in that season...in other words...if I saw these guys moving in and out of other teams...in front of different defenses...and in 2001 they seemed to be getting shortchanged...but that's not what happened.
Moyer's 111 was right in line with a typical Moyer season...he had a few 120s and a few 100s in his peak...his average season was a 108.
Garcia's 121 was a CAREER YEAR! He was never better than that season...in fact he was never CLOSE to that season...his second best year, he hit 112.
Halama's career looks like this:
1998 NL HOU 96 82
2000 AL SEA 494 89
2001 AL SEA 320 75
2002 AL SEA 301 104
2003 AL OAK 326 86
2004 AL TBA 355 93
2005 AL BOS 137 80
2005 NL WAS 64 92
But his peripherals in 2001 explain the drop.
Abbott's career looks like this:
1990 AL MIN 106 79
1991 AL MIN 136 83
1992 AL MIN 33 72
1993 AL CLE 55 66
1998 AL SEA 75 131
2000 AL SEA 532 114
2001 AL SEA 473 84
2002 AL SEA 76 46
2003 AL KCA 142 89
2004 NL PHI 149 67
2004 AL TBA 143 79
The anomaly was 2000 here...his career DNRA+ was lower than his 2001 DNRA+
Here's Sele:
Sele 1993 AL BOS 334 110
Sele 1994 AL BOS 427 123
Sele 1995 AL BOS 99 107
Sele 1996 AL BOS 484 106
Sele 1997 AL BOS 538 86
Sele 1998 AL TEX 657 116
Sele 2000 AL SEA 624 114
Sele 2001 AL SEA 625 96
Sele 2002 AL ANA 458 74
Sele 2003 AL ANA 366 80
Sele 2004 AL ANA 399 91
Sele 2005 AL SEA 343 71
Sele pitched in front of some AWFUL defenses and some great ones...his DNRA+ curve looks like a classic pattern of aging for a mediocre pitcher. There's no indication that the Mariners' pitching staff is getting short changed in 2001 compared to their normals...separating pitching and defense the way I do WORKS...it smooths apparent fluctuations in performance quite well...it never seems to fail to make career production curves look more natural (with the expected moves in skill caused by aging, injuries and the like). You've been staring at bubblegum cards so long that I think you have yourself completely convinced that fielding can't caused an entire team to overperform...the numbers back me up here though. The Mariners in 2001 were average with their K rate (bolstered by a phenomenal pen), they gave up way too many HRs for playing at Safeco Field (that rotation is an EXTREME flyball rotation...Sele, Halama, Moyer, Abbott, and Pineiro were all flyball pitchers in 2001. Garcia was neutral to slightly flyball heavy. It makes sense that such a staff (high contact, high FB rate) would give up too many longballs, especially given the low walk rate (they got too much of the plate a lot of the time). There's just no way around this...the '01 Mariners had average pitching and spectacular defense.
3
BTW....you do raise a good point about 2002...although I would argue that despite having the same defensive cast as 2001, the defense did in fact play somewhat worse. Boone was a big step worse in 2002 than 2001...both offensively and defensively. Olerud began to slide with the glove in 2002. Dan Wilson went from average to terrible with the glove in 2002, Cameron didn't have as good a season with the glove either...all of the defensive metrics agree his 2001 and 2003 were his good years. Still - it's true that we don't know why the Mariner outfield didn't play as well that year statistically and some of the difference might have been harder contact that we weren't seeing.
That still doesn't cancel out, in my mind, the hard-to-overlook journeyman nature of the 2001 Mariner rotation.
Abbott kicked around the minors for 10 years...for a reason. Halama was a PTBNL...for a reason. Aaron Sele was getting worse (and was let go by the Ms in 2002)...for a reason. In fact, if you recall, as we entered the 2001 season, there were two big worries about the club...
1) How were they going to replace A-Rod's production?
and
2) Was their rotation going to be deep enough and get deep enough into games to hand the ball to the good bullpen?
Those 2001 Mariners were Moyer, Garcia and pray for runs. The fact that the defense gelled and the bullpen held a lot of 5-3 and 6-4 leads to produce a good team ERA should not trick you into believing that the club was special on the mound. Did Paul Abbott, John Halama, Aaron Sele or even Freddy Garcia EVER demonstrate any special ability to alter BABIP at any point in their careers? No they did not. Only Moyer showed that ability.
Hits prevented over team (career total):
Moyer: 189
Garcia: 13
Sele: -48
Halama: -35 (in way fewer innings)
Abbott: -56 (in even fewer few innings)
Pineiro: -26 (although +11 in only 70-ish innings in 2001)
The complaint against Sele was ALWAYS that he was homer-prone, a little too hittable and made team defenses worse...especially in Arlington. The complaint against Halama was that he had very flat stuff...his command had to be perfect for him to be effective. Abbott was a lot like Ryan Franklin...4-5 pitches, none of them better than average. Moyer, even in 2001, had no margin for error - he had to be spotting his change-up perfectly or he would get splattered.
Let's not make this rotation out to be some kind of special force of devastation here, Doc. They weren't even particularly good. Just 5.5 arms you could throw out there into the teeth of one of the greatest team defenses of the last 50 years and count on to put the ball in play and work quickly and hand it to the pen in a lot of close games.
4
Some great work there, Matt.
My own growing sense of reality is this.
1) McCracken was "mostly" correct. The vast majority of pitchers don't have "significant" impact on BABIP results. But, one shouldn't pretend the occasional Moyer - or Washburn doesn't exist because of the general rule. I suspect there are a few anti-Moyer's out there, (with high K rates, and consistently worse-than-expected BABIP rates). But, the high K rates keep them in the majors. Of course, a bad BABIP is gonna end most careers quickly UNLESS they've got above average K-rates, so the pools are small, and the juicy center overwhelms the outliers).
2) I believe that Ks and BBs are much more stable stats for pitchers in general, with HRs just behind. And, because they are stable (relative to other things), they become more predictive than those other things. But, "predictive" and "important" are not synonymous.
3) I believe that defense may actually be *MORE* than 1/2 of the runs allowed picture in actual importance. But, I believe that the historical views of the component parts of defense have been so generally cloudy, that almost nobody has any real clue as to what ACTUALLY determines defensive prowess. If this is so, one would expect to find wild swings in defensive performance - and there would also be a greater tendency (compared to most other stats) for the "average" - (the center of the bell curve) - to be wider than in some other areas -- or if not wider - so unstable, that the fluctuations render the defensive stats most useless for analysis over time.
4) I believe that the #1 factor in determining runs allowed is NOT Ks, BBs or HRs -- but simply hits allowed. But, the only "stable" stat the analysts have for projecting hits allowed is Ks. So, the current models tend to begin "assuming" average defense -- because they actually understand they are attrociously bad at projecting defensive results.
5) If defensive results are volatile for "hidden" reasons, (ABILITY is NOT the primary key to defensive prowess), then all attempts to attach defensive results with ability are utlimately doomed to failure. The "noise" will continue to frustrate the analysts, as every model will continue to fail, as each successive model attempts to reconcile the fluctuations as some part of "ability" that is not yet taken into account.
My personal belief at this point is that defensive prowess is controled by two primary factors -- DESIRE to play defense 100%, every day -- and positioning that is in sync with the pitching game plan being implemented.
I believe that if the above two variables are even, THEN actual athletic ability kicks in and becomes the determining factor between two defenses (or defensive players).
I also believe that some players are SELF-motivated enough to play defense that these players track across time as defensive whizzes regardless of most other factors. This gives the "perception" of defensive ability. But, even though time slowly erases the actual ability, many of these players continue to score well in individual defensive metrics despite age and injury.
In the end, I think the number cruncher who really bears down and laser beams in on hits allowed -- and starts getting some feel for projecting hits allowed that does NOT rely solely on K-rates is going to produce the next quantum leap forward in defensive analysis. But, since I believe that a large chunk of this is actually psychologically induced, I think that effort is going to have some major hurdles to clear.
5
As usual, Sandy, a thought provoking reply.
I agree completely with your point #1, obviously DNRA was engineered with that paradigm in mind.
Your point #2 I agree with as well in the absolute sense. You're correct that Ks and BBs may not be the most valuable parts of a pitchers game (in the absolute run reference frame) despite the larger predictability they possess...afterall, the vast bulk of a pitcher's DNRA comes from the hits he allows, not the walks. I would caution you, however, that the variation in hit rates among pitchers is significantly smaller than the variation in K and BB rates...McCracken's studies revealed how using only K, BB, HR and the average fielding assumption, he could essentially completely reproduce the typical range of ERAs for a league. In other words, BABIP is a very minor contributor to the spread in pitching performance.
Numerically speaking, your point #3 begins with a position that the data doesn't currently support. Even if we assume that HR, BB and K are 100% pitching and everything else is 100% fielding, we get that pitchers are contributing about 53% of the linear weight runs created and about 57% of the variability in team LW Runs Allowed. That's the baseline you have to start with...after which point any influence pithcers do have on things like baserunning, BABIP and GB/FB pushes pitching control slightly higher. Leaving me pretty happy with my Pitching/Fielding split of 59-61 for the pitchers and 39-41 for the fielders. Now that was not the case as recently as the 1950s when K and HR rates were far lower, but in the modern game, I don't see how pitching could be responsible for less than 57% of the defensive game.
On the rest of your point #3, the standard deviation of PCA defensive wins created per 162 games is about as large as the mean win scoring rate for most positions, for example, center fielders average 2.52 wins per 162 with a standard deviation of 1.6 wins/162. But before you conclude that fielding stats are more useless than other stats, think again. The standard deviation for offensive wins created by PCA is ALSO about as large as the mean win scoring rate (4.15 +/- 3.9!). Baseball is played by humans...and humans have chaotic performance tendencies in everything...not just sports. If you stare at the defensive win cards produced by PCA, as I have done, you come to appreciate how well even a very basic uberstat model like PCA does at separating good fielders from bad ones and picking out the bad seasons sensibly (almost always, when I see a really low win figure produced by a good fielder, it comes with some story about how that player was injured that year or had legal troubles or also struggled offensively or some other obviously red flag that explains it). I do not believe that a top-down defensive analysis that starts with team data and divides credit among the players will be prone to the kinds of wild fluctuations that you seem to think are endemic to fielding stats...at least...no more so than offensive stats.
On your point #4, I think we're atrociously bad at forecasting defensive performance solely because we're not using team analysis. All of today's modern sabermetricians have taken the bait on one of the worst mistakes in the history of the analysis of this game...that being the decision to believe that zone based fielding analysis works without a team context. It doesn't. The high variability we observe in UZR comes partially from the system's failure to place players in their proper context.
As to your stated belief that fielding performance is led mostly by motivation and coordination between fielders and pitchers, that's possible, but it's not merely "a few defensive whizzes" who look good no matter where they play. It turns out that a top-down defensive analysis finds the very good players just as well as it finds the great ones. If you start at the top of any position's all-time defensive win list and work your way down, the top 15 will be very logical (the better a player is, the more he stands out)...but even once you get into the merely good gloves, you're still going to be nodding your head more than scratching it. I could bring up countless hundreds of fielding profiles for merely-good players, and you're not going to see any more inherent wobble in their win scoring rates than you do for the top fielders.
6
Matt,
I understand that there are lots and lots and lots of ways of looking at data. And, I believe among the number crunchers a LOT of your work in regards to defense is significantly ahead of the pack.
But, I'll tell you why I think defense is more than half the runs allowed story. It's grade school simplistic, but I do not believe it invalidates the position.
Outs: (in integers)
Pitchers: 7
Fielders: 20
Baseball, by its rules, allows theorhetical INFINITE scoring until 27 outs are made, (we'll ignore lost half innings and extra inning games for this argument).
True, the above doesn't take into account walks or HRs. But the walks and HRs have to then account for an additional 13 outs. Fielders are DIRECTLY responsible for almost 3 out of every 4 outs made. What prevents runs? Outs.
But, in addition to this basic reality -- is the reality that as defense improves, it ALSO lowers the number of HRs and Walks allowed by the pitchers.
In 2007, Tampa allows 1649 hits. In 2008, they allow 1349. They prevented 300 hits. (The team struck out 50 FEWER batters). So, it's really 350 hits the defense erased. If you erase 350 CHANCES to hit a HR, how many HRs did the *DEFENSE* prevent? If HRs are about 1 out of every 10 hits, then a defense the prevents 350 hits, *ALSO* prevents 35 HRs. (Tampa's HRs-allowed team total dropped by 33 from 2007 to 2008 (199 to 166).
7
OK...your best arguig point is the "saving hits saves chances to get a DIPS-event like a HR or a BB" bit. The first part is completely overruled by all modern linear weights analysis. It's true that pitchers only directly get 7 outs, but it's also true that the 20 outs fielders get tend to predominantly be extremely routine and the variation in K rate is about the same as the variation in hit rate.
But to your other point, let's say we go with your point. You can't get a HR without swinging, so we need to look at the number of additional at bats, not the number of additional plate appearances that could be gained by bad defense. Saving 350 plate appearances at a normal walk rate is about 35 walks saved...so mark that down as helping the pitchers to the tune of 12 runs. The typical HR/AB in the modern game is about 1:35, so you're talking about 9 home runs saved (we already lost 35 of our 350 PAs to the walks)...so mark down another 13 runs saved. The typical range of defensive runs allowed in a season is 300 or so. 25 runs shifted to the pitchers amounts to about 8% of the typical variation, so at the outside...in one of the most extreme defensive turnarounds in major league history...we might be talking about crediting the defense with 50% instead of 57 or 58%. BUT...that's the extreme team. The average team will have no such sliding value moving from pitching to fielding. Which means the average team defense still only accounts for 40% of the defensive game. Your point will only serve to increase the variability in the percentage fielders claim from the team defensive marginal runs...in other words...great fielding teams will be worth more and bad fielding teams will be worth less than they are under my current scheme...but the average team will be worth the same.
BTW, this is something that old-PCA doesn't account for, but that new-PCA will. Old PCA was a fine piece of accounting for what actually happened but made no attempt to correct it so that each defensive unit (P, C, IF, OF) was graded as though they were contributing to an average team defense around them. New PCA will. In other words...take the accomplishments of each unit and prorate them to the chances to produce that they would have gotten had they been in an average context.
8
Points we can agree on.
1) "General" range of runs allowed per team per season is about 300 (600 - 900)
2) "General" range of Ks generated per team per season is about 300 (900 - 1200)
3) "General" range of hits allowed per team per season is about 300 (1350 - 1650)
So, at least in this glance at the stats, the concept of a 50/50 split between defense and pitching has some "untweaked" data to support it.
BUT -- a strikeout, while being an out, does *NOT* prevent a hit. It only prevents a ball-in-play. And the ball-in-play WILL be turned into an out (by the defense) roughly 2/3 of the time. So, only 1/3 of strikeouts actually prevent a hit.
Meanwhile, a fielding out DOES prevent a hit, because fielding is the last line of defense. If they don't make the play, it's a hit.
If one were to, (for the sake of argument), divide the K column by three to take into account the fact that 2/3 of non-Ks are going to turn into outs anyway, then you would end up with a K range of 300-400, and when compared to the hits range, you'd be right back in the vicinity of the original 20/7 ratio. Hmmmm.
Now, before the steam starts pouring out your ears, (or the ears of any other SABR dudes lurking), if blithely dismissing 2/3 of the strikeouts generated as trivial seems arbitrary and capricious, I would respond that doing so for fielding plays is EQUALLY arbitrary and capricious. If one is going to ever build a complete working defensive model, then one has to account for EVERY out -- not just the easy or hard ones. And if one decides that looking only at the "above baseline" plays is required to complete a working model - then it only makes sense to utilize the same logic for BOTH fielding and pitching.
I remember that some of your foundation work was heavy into Bases-per-out. I think that is an excellent starting point for defensive analysis. But, if you take the total bases allowed by a team during a season, and then split the credit between pitching/defense with outs -- well, you run right back to that 20/7 split I'm starting with -- and then it becomes a "chore" to explain why credit for outs recorded by fielders should somehow be ignored - or the value of those outs diminished or transfered to the pitchers in some way.
One of the limitations of analysis for defense is that practically all I've seen BEGINS with the assumption of an average defense -- and given that foundation, the DIPs numbers behave very nicely. But, you get a non-standard defense, and they simply fall apart.
Well, the best run-suppression teams around are "typically" those with superior DER totals -- and these are the very teams that don't work well with the models. I see a flaw with an approach that says, essentially -- "To become really great, you have to do "something" that simply doesn't work within the confines of the model we have to work with today."
9
Not quite accurate Sandy.
A fielding out prevented not a full hit but 1/3 of a hit. Each fielding out converts a random batted ball with a 0.7 probability of being an out into a definite out.
10
Let's put it this way, for the sake of clarity.
Every time a batted ball occurs, it has a 70% chance of being converted into an out.
A K prevents a batted ball from occurring at all, thus converting a potential batted ball into a definite out.
But every time a fielder makes a play, he's converting a potential out into a definite out just the same. They have exactly the same real value (actually a K is worth slightly more because it prevents baserunner advances often enough to more than offset the loss of DP chances) and exactly the same theoretical value too.
11
Let's go to the "holodeck" for a moment.
We have a pefectly average pitcher - and a perfectly average defense in this simulation.
Baseline: (2700 outs -- 100 games - just for round numbers)
Baseline Simulation:
Pitchers: 700 Ks
Fielders: 2000 outs
Hits allowed: 1000 (.270 BA -- BABIP .333 -- we'll ignore HRs for the moment just to keep the numbers simple)
Simulation A - Plus Pitcher:
Pitchers: 800 Ks -- (100 extra outs). 14% increase in Ks.
Fielders: 1900 outs
Hits allowed: 967 hits (BABIP = .333) -- 33 hits removed
Simulation B - Plus Fielders:
Pitchers: 700 Ks
Fielders: 2000 outs
Hits allowed: 967 hits (BABIP = .326) -- 33 hits removed
The fielders increased their out-producing rate by just over 1%, to achieve the IDENTICAL result of the pitchers increasing their K-rate by 14%.
Tampa from 2007 to 2008 had a 54 point improvement in DER. That prevented 300 hits.
To achieve the SAME reduction in hits through pitching alone, (if BABIP had remained at .334), would require the pitching staff to increase their Ks by NINE HUNDRED (900).
=====
The problem is outs are a constant. They can *ONLY* be measured by hits allowed.
In truth, for the 2007 Tampa staff, a non-K only turned into an out .67% of the time.
But, in 2008, a non-K turned into an out 72% of the time.
While I don't know the exact number -- each point of DER improvement (or BABIP) from a defense is probably worth something like 6 hits over a full season. (Maybe you can get the exact number, because I'm doing some seriously dangerous ball-parking here). Maybe it's lower than I think - only 5.
But, for the sake of argument, say I'm right. A 20 point improvement in BABIP prevents 120 hits during a season.
It would require a 360 improvement in Ks to achieve the same end, (assuming identical defense).
It shouldn't be too difficult to go and find some teams with very similar K and HR numbers (walks won't matter for the DER examination), but different DER/BABIP numbers -- and get a general read on how many hits a point of DER is worth.
It also shouldn't be too difficult to go and find some teams with similar DER and HR totals, and see what the impact of K differences are on the hit totals.
12
The problem, Sandy, is that you're not looking at these numbers on the proper scales.
You're erroneously concluding that fielders make a 1% change where pitchers make a 14% change to achieve the same results. But the4 fielders AREN'T making merely a 1% adjustment...it takes a SIGNIFICANT change in fielding ability to get a 1% change in BABIP. Why? Because most batted balls are routine hits or routine outs. There are in fact very few plays in a baseball season that could go either way. What's missing from your analysis is the concept of the production margin (zero-value margin).
Here's how I would re-express your example, first in absolute units, then in marginal units.
Scenario A, better pitchers, average fielders:
800 Ks instead of 700
1900 fielded outs (of which about 500 were non-routine to at least some level)
33 hits saved
Scenario B, average pitchers, better fielders:
700 Ks
2000 fielded outs (of which 600 were non-routine - thereby producing the...)
33 hits saved
The fielders have to make about the same skill adjustment for their hits saved figures to match. Numbers are for illustration purposes. I actually think that the non-routine players are even more rare than this and that rather than having to make 100 non-routine plays to save 33 hits, they have to make, say, 45 plays to save 33 hits because non-routine plays do not carry the standard BABIP...the average team misses most of them
Seriously Sandy...think carefully about this...when you watch a baseball game, how many times do you see the batter make contact and IMMEDIATELY say "out." How many times does the announcer raise his voice after witnessing an out recorded?
The problem we have here is scale...it's not correct to assume that 2000 outs have significantly more value than 700 Ks...you should know this intuitively...the margin on those outs is going to be exceedingly high compared to the margin on the Ks (because you stick some schmuck like A-Rod in at third base and he's still going to make like 95% of the out plays of the best third baseman in the league in the absolute counts. Look at the leaders in putouts in the outfield each year...the range for guys with at least 150 games played is something like 250 to 350 for the corner outfielders and 300-450 for the center fielders. That's the range from the absolute worst outfielders to the absolute best. But we KNOW that the worst outfielders are not worth 2/3 of the value of the best ones...they're worth NOTHING (or even negative value) while the best ones are worth multiple wins. The reason is as described above...most plays are routine and contribute no real value to a team. Only the plays that get missed by the bad fielding teams but made by the good fielding teams have value.
Compare that to the typical ranges of Ks...750 to 1100 or more for some teams. Not only is the spread somewhat bigger, but the base value is a MUCH smaller percentage of the top value for Ks than it is for fielding outs per batted ball.
You're getting yourself all twisted up trying to account for things on a linear surface, but you need a PYTHAGOREAN surface with a marginal baseline to make any sense of team data.
13
But where is you adjustment for the "routine" strikeout?
Back in the 30s, the league average for Ks was 4 -- today it is 7.
Today, the bottom of the pile stinkiest pitcher allowed to hold a job get 4/9.
Let's say for the sake of argument that 15 of the 20 fielding outs are "givens". That leaves 5 outs per game that "matter".
Well, why don't we also say that those base-line "worst" pitcher Ks are "given", also? If so, then we get 3 outs that matter - and we end up with a ration of 5 field to 3 pitcher -- STILL giving defense more impact on run prevention.
=======
One of the underlying problems I see with many of the defense/pitching paradigms is that I don't see any adjustment being even attempted (outside of your Matrix), to adjust for the fact that either side can skew the results for the other.
Almost all the baseline stats created today are "per nine" stats. HR/9; BB/9; K/9. The divisor is a constant and gives "truthiness" to the results. The problem is that the starting point should be "per plate appearance". Because the strength of the defense or pitching CHANGES the chances for the other side.
When the defense produces 300 extra outs, that's 300 fewer plate appearances for the pitchers. So, an improved defense, with an IDENTICAL pitching performance is going to drive the K/9, HR/9 and BB/9 numbers *ALL* down for the pitchers. Conversely, a horrid defense can create a similarly skewed picture in the opposite direction -- allowing 300 more hits to drop is 300 more chances for a pitching staff to rack of Ks. The K/9, HR/9 and BB/9 numbers should ALL go up as defense wanes. And the problem here is that while those are the TTO numbers -- the "good" direction for Ks is opposite that for HR and BB.
==========
Well, your 750 to 1100 range is patently wrong. Look at Ks BY DEFENSE. The last AL team to fan fewer than 900 batters in a season was KC in 2004, (887). The range from best to worst pitching Ks (AL) is a lot closer to 250 than the 350 you imply. (And AL is a cleaner analysis, because you don't have pitcher Ks complicating matters).
My glance at hits allowed versus Ks rung up does NOT support any significant edge for Ks. The last two seasons, the hit-range has been wider than the K range. Prior to that -- it's been random which range is wider.
14
Sandy...you're still missing the point.
You're still trying to do things on a linear scale...yes you subtracted out the baseline for Ks and outs and found that non-routine outs are still more common than "non-routine" Ks, but that's NOT how baseball works. Things don't accumulate linearly toward value...they accumulate on Pythagorean surfaces toward wins and losses...it is non-linear and you can't treat the data as though it should be linear as a result. Even if we take the last four years as gospel as you insist and say that the range in Ks and hits is about the same (250)...the percent of top performer is WAY different.
Worst K team usually around 900, best K team close to 1200...75% ratio
Worst hit/BIP team usually around .320, best usually around .280, 88% ratio
That's where we're starting here...that means that, in reality, it's not 5 out of every 20 outs that is non-routine...it's about 3 out of every 20 outs that is non-routine...and here's the tricky part...you're perfectly willing to look at the negative events for fielders (hit rate) but you ignored HR on the pitching side willingly in your example. They may be more range...but guess what...the range in HRs (take them out of the hit column) is about 100 from good to bad...and eahc HR is worth 1.4 runs! An order of magnitude more than each out (outs usually cost a batting team about 0.11 runs). The big things that pitchers can do to generate value (preventing HRs is a really big thing, preventing balls in play is a fairly big one too) outweigh the things fielders can do (convert a 70% out to a full out).
Here's where I come down on this, Sandy...your way of trying to simplify the problem so you can get at the core logic is a GREAT analysis technique 99% of the time, but it's going to cause you to be flat wrong when you try to apply that kind of thinking to a problem of this complexity (how important is pitching vs. fielding)...you NEED to see the WHOLE picture...you need to count the range of possible runs for both sides.
The range of walks is about 300 (that's about 100 runs different)
The range of HRs is about 100 (that's about 140 runs different)
The range of Ks is about 300 (that's about 90 runs different - .3 * 300 balls in play)
The range of HBP + WP is about 80 (that's about 25 runs different)
The range of defensive play-making rate is about 0.04 hit sper BIP (given a typical BIP count of 4000, that's about 160 runs different)
The range of 2B+3B per in play hit is about 0.1 (this ends up being worth as much as 50 runs...I did this math for Dr. D some months ago - ironically enough to justify why he was wrong that pitchers accounted for 70% or more of the defensive picture...LOL...now I'm using the same logic to refute a way lower estimate)
The range of DP/chance is relatively small, but let's give this credit for possibly being worth 25 runs in extreme cases
The range of SB and other baserunning runs is on the order of +/- 40 runs
On the pitching side, I get a much bigger potential influence than you seem to believe possible...I get something like 350 for the pitching side vs. 275 for the fielders...which is more or less right in line with my assertion that pitching is 60% of the defensive game.
15
whoops...I have a bad figure for the HR count spread...the range is actually closer to 75...which is the figure I used last year when explaining this to Doc...so adjust the pitching-side run estimate down to about 330...which is what I had the last time I did this...that'll teach me to round things off in too much of a rush. Sorry about the potential confusion.