Dr. Doom here (via Doug) again with an awards-voting post. We’re switching leagues (and awards), as we vote for the Cy Young in the AL. I don’t think I’ve ever written a Cy Young post before, so here we go.
There are two clear frontrunners for the award. Let’s look first at Chris Sale. Sale paced the AL in strikeouts with 308. (When was the last time someone had 308 strikeouts in a season? It was in 2002 – a little fella named Randy Johnson.) Sale also led the league in strikeout rate – 12.9 per 9 innings, the third-highest rate in history, and placed second in strikeout-to-walk ratio (7.16). Sale’s 2.90 ERA and 157 ERA+ were also second ranked, while his 2.45 FIP and 214.1 IP paced the junior circuit. And, oh yeah, those traditional numbers? Sale also won 17 games, only one off the league lead.
The other strong candidate in the AL is Corey Kluber. Kluber posted an unbelievably low .869 WHIP – second lowest (behind Pedro‘s .737 in 2000) in the AL in the DH era, and a healthy .046 better than Felix Hernandez (2014) in third place. Kluber pretty much finished first or second in the league in everything: K:BB (7.36, 1st); FIP (2.50, 2nd); WHIP (.970, 2nd); walks (1.6/9, 1st); strikeouts (265, 2nd); ERA (2.25, 1st); and wins (18, first).
Of course, while there are two really strong candidates, we all have five-man ballots to fill out, so who else has an interesting case?
- Justin Verlander split his time between two teams, but was third in the league in innings and pitched lights out down the stretch for Houston. He also led the league in starts at 34.
- Luis Severino was third in the league in a lot of things: ERA (2.98), ERA+ (3rd), FIP (3.08), and WHIP (1.040).
- Ervin Santana started incredibly hot and managed the league’s third-best win total (16) and 4th best ERA (3.09).
- Carlos Carrasco tied for the league lead in wins with 18, pitched 200 innings, and ranked third in K:BB (4.91).
There are other candidates – plenty of ’em. So enjoy, and good luck sorting out who’s going where on your ballot!
DIRECTIONS: Please list 5 players on your Cy Young ballot in a NEW comment below (ballots with fewer than 5 candidates will be thrown out; I ask for a new comment because it’s easy to lose one if it’s in a reply, especially since we got rid of numbered comments). Ballots will be scored as per BBWAA scoring (7-4-3-2-1). Strategic voting is discouraged, though that’s unenforceable, so please just don’t do it, as the goal here is to (somewhat) mimic the BBWAA process. The post will be live for about a week; I will comment shortly after the post goes live to tell you when ballots are due. Please discuss and vote whenever you’d like, but there will be NO vote changes, so don’t vote until you’re sure you’re ready!
Let’s keep this open through Monday night, 11:59. Can’t wait to read your thoughts and votes! Also, happy Thanksgiving, everyone!
Do we vote for 3 or…?
Just saw it- 5-man ballots.
1. Kluber
2. Sale
3. Carrasco
4. Severino
5. Verlander
Almost went with Sale because he had two more starts and more IP, but Kluber was more effective overall with 5 CGs and 3 ShOs. Kluber had more innings per start, a higher average Game Score (70 vs. 67 for Sale) and was dominant down the stretch.
I always find seasons without 20-game winners disappointing. They are far more common than they used to be, of course, due to pitchers being pulled earlier in games. One thing I think needs to be discussed is the way Ks affect Wins in the pitch-count era. We all know high Ks mean high pitch counts. Guys like Unit and Rocket would throw up to 150 pitches in some of their starts, and they continued a high-K/high-W tradition that marked the ’60s and ’70s. But if you put a strikeout pitcher on a pitch count diet, you’re going to get a game log like Sale’s, where he’s clearly being removed from game after game when he’s pitching well, because he’s reached a 110-115 pitch limit. The result is going to be a certain number of potential Ws forgone.
Take, for example, Sale’s performance on July 15 this year against the Yankees — an important game, two contending teams head-to-head. Sale pitched a shutout through seven with only two hits, and started the 8th strong: a groundout on an 0-2 count, a single into right on a 1-2 count — a short, opposite field hit — followed by a strikeout on a 1-2 count, Sale’s 13th in 7.2 innings. But that brought Sale to 118 pitches, his highest season count. Up comes Aaron Judge. Wouldn’t you want your top starter, who’s steaming along throwing strikes in a shutout, to be the one to face Judge, despite having no righty/righty advantage — Judge’s 2017 average against Sale was .000, and his K rate was .667 (I didn’t look it up, but I think we can infer that on July 15, Judge’s average against Sale was not too far from .000). But no: Farrell goes to Kimbrel right there. Kimbrel gets Judge, but gives up a game-tying homer to Holliday first thing in the 9th; the Sox go on to lose in extras and Sale gets no W.
If the tyranny of pitch count is going to be a given, then it’s the high strikeout pitchers who will pay the highest price in terms of wins (and, naturally, CGs and ShOs). This is not the way overpowering pitchers in the past (Walter and Randy Johnson, Clemens, Koufax, Ryan, etc.) were handled. The value of the K-pitcher is really twofold: 1) there are (almost) no errors on a K, the batter’s guaranteed to be out and runners rarely advance; 2) high Ks are a sign of effective pitch delivery that produces an edge in every PA, whether a K or not (unless, perhaps, like Ryan, you offer batters almost equal opportunities for walks). But if high Ks also comes to mean above-norm reductions in CG and W opportunities, then they may reduce the value of a pitcher more than a lower K total would. The result may be that Cy Young candidates whose WAR and averages are boosted by high Ks, and those are the stats that we look more at now, but who compete on the basis of increasingly low-value seasons, compared to their predecessors, if you value their control over generating actual team wins — which, advanced stats or not, is one important version of the bottom line.
epm:
It used to be—in my mind, at least—a sign of a good starter on a good team that the team’s W/L record in his no decisions wasn’t too far off from his personal W/L record. If he was pulled in the late innings of a tied or a close game, his teammates had the firepower to take that situation as an advantage for a late inning win. In 2017, at least, this was not the case with Corey Kluber and the Tribe. Kluber’s no decisions ended up 2-5, not quite the reverse of his 18-4 record, but I’m wondering if this isn’t connected somewhat to your investigation about pitch count. Every one of Kluber’s No Decision losses came in a well pitched game, where he left with his team in a good position and an exceptional bullpen to hold that position.
Sale’s Bosox fared better in his No Decisions, going the reverse of Kluber’s at 5-2, but in the two losses the pattern is there, and in three of the five wins, most of the scoring happened after Sale’s departure.
Slavish adherence to pitch counts does seem a little silly when a pitcher is still seemingly in command. But, changing that approach can only happen gradually as pitchers today are conditioned to expect to be removed, so would probably not perform well if managers suddenly started leaving them in games in the late innings, much less doing so if they got into a jamb in those innings.
The other factor is that managers are second-guessed to death if they leave a pitcher in too long, yet somehow escape that treatment for going to a reliever and then regretting doing so. It was sort of the opposite in the old days, with heavy scrutiny on pulling your starter too soon, even if it did work out; instead, your best pitchers stayed in games because they were your best pitchers, and if it didn’t work out, well that was just baseball – things don’t always work out the way you planned.
On WAR and WAA, I’ll go with Kluber as the easy winner.
1. Kluber – 8.0, 6.2
2. Verlander – 6.4, 4.5
3. Sale – 6.0, 4.1
4. Stroman – 5.8, 4.0
5. Severino – 5.3, 3.6
Carrasco (5.4, 3.6) was a virtual tie with Severino, but I’ll give the nod to the Yankee ace, with less help in the rotation and, thus, more pressure to perform.
Some comments on contenders:
Corey Kluber looks the obvious best pitcher, but some chinks in his armor exist. He was 1.02 better in ERA at home. Only 8 of his 29 starts were vs winning teams. His stats with RISP aren’t as good of those of several contenders for the CY.
Chris Sale: he was obviously held back some as far as pitching in Fenway is concerned, only 13 of 32 starts came at home. Only 9 of 32 starts vs winning teams, record 2-4. Aug/September 4-4 record with ERA over 4.00.
Justin Verlander was 7-6 with a 4.16 ERA away. He pitched indifferently for a bad team and excellently for his one month with a good team.
Luis Severino: Splits in ERA 3.71home/ 2.24 away; 3.54 1st half/2.28 2nd half. 11 of 31 starts vs winning teams.
Ervin Santana: Like Severino he was much batter away than at home. Unlike Severino he cooled greatly, not just in the second half but in the last 5 months. Without his hot April, would he be under consideration for the CY?
Carlos Carrasco inverted his teammate Kluber’s ERA prowess, being far better away than at home. He got outstanding runs support. Only started 6 of 32 games vs winning teams.
Marcus Stroman pitched much better at home than away, got amazing run support from a team that finished last in AL scoring.
A small part of those low start totals against .500 teams is selection bias, as in it’s tough to be a winning team facing the likes of Kluber and Sale. If you’re curious which starters faced .500 teams most often, it’s these guys with a minimum 10 starts against .500 clubs, comprising 50% of total starts.
I’m going to pick at a few of these points, simply because I think the issues raised are interesting ones. NSB has dug deeper to try and refine and enrich our view of these pitchers, which is precisely what a project like Doom’s invites us to do (if we can find the time and energy). My interest is in looking for cases where the added data may look different if you dig yet one level deeper still. (It’s part of a theme for me: looking at the different perspectives provided by aggregate stats and the data from game logs, play-by-play, and contemporary narrative; that is, recovering what aggregation necessarily loses in order to retain as much qualitative richness as possible to supplement the power that quantitative reduction provides.)
On the home/away split issue, where there may be a tendency to value consistency between the two figures, or to slightly devalue strength on home grounds, Kluber’s home/road ERA split is an interesting angle to think about. Progressive Field was a mild pitchers’ park in 2017, and that may contribute to the split a bit, but it’s close enough to neutral that it may have been Kluber’s record that tilted the overall team stats in that direction. Assuming the park was neutral, why would you not consider it a strength rather than a weakness that a pitcher does best in his home park, where he pitches about 50% of his games? While we may like the narrative of a pitcher so dominant that he wins anywhere, realistically speaking, a pitcher who can’t figure out how to get an edge in a neutral home park, where he has far more opportunity to master the angles of the park’s geography, is either one who has been suboptimally recruited for that home team, or one who doesn’t make use of added opportunities to adapt his skills. So it seems to me that Kluber’s record in this regard could be viewed as superior to those of Severino and Santana (although it would be unfair to cast shade for these reasons on a player of Severino’s youth and relative inexperience in his home park).
Still on the home/away split issue, but on a more specific point NSB makes: I’m not sure that Sale was, in fact, held back from pitching games at Fenway. Sale pitched on a very regular rhythm all season: 29 of his 32 starts were with 4 or 5 days’ rest. On two occasions he pitched on six days’ rest, and the resulting starts were away once and at Fenway once. There was also one instance of pitching with 8 days’ rest, perhaps injury-related: the next start was at Fenway. It is true that he slightly underperformed at Fenway in terms of ERA — the park was almost pitcher-neutral last year, and Sale’s ERA was 0.45 higher. But he got the job done, going 7-3 in 13 starts (the Sox’ record with Sale starting was 9-4 at home and 13-6 away — essentially neutral, with a slight home edge: nothing that would have caused Farrell to hold Sale back at Fenway later in the season, which is when the imbalance developed).
On the .500>/<.500 opponent issue: Since in the AL there were only five .500+ teams and ten below .500, we should expect pitchers to have lopsided splits in that regard, especially if they pitched for a .500+ team. In interleague games, the NL opponents were slightly more likely to be below .500, and the teams encountered were, in the end, the schedule-makers’ choice, rather than a product of chance or managerial strategy (as it turned out, all the Sox IL games were vs. the largely unimpressive NL Central teams, plus the hapless Phillies). Sale is an interesting illustration of how aggregate figures can be misleading in this regard: He pitched against only two NL opponents, both sub-.500. His record was 0-1. Sounds disappointing. But his ERA in those two games was 0.60 (with no unearned runs): one run, seven hits, one walk, and 17 Ks in 15 IP. The disappointment was that the Sox scored zero runs behind him; Sale was as dominant over those weak teams as one could possibly ask.
In writing about Santana in my ballot post, I addressed the issue of weighting the value of early-season vs. late-season strength. This is a conversation we’ve had here on and off for years in various forms. I do think there is validity in judging how a player performs under the increased pressures of late-season (or, in a game context, late-inning) competitive tension: the “clutch player” issue. But in terms of season value provided, I think clutch performances are far less relevant than total contribution. There is, by the way, an interesting string over on the Tangotiger Blog concerning this issue — the general issue of “high-leverage” performance vs. total contribution — and I’d be happy if we had an opportunity to go into more detail on that here. There are, on that string, some particularly interesting (and, I think, somewhat off-target) comments concerning prospective vs. retrospective assessments of leverage that I think we might like to take up.
epm:
Did you by chance read the Bill James article on Altuve and Judge that was referenced in the thread? If there’s a starting point to any discussion here on interpreting stats it should be this article:
https://www.billjamesonline.com/judge_and_altuve/
I have never been able to get links to work at HHS, unfortunately.
By golly, this one worked.
I’m reading James’ article, and there is a statistical point he makes about WAR to which he infers but does not explain. He bops Judge’s WAR down from 8.1 to 6.8 because of the Yanx’ crappy pythagorean.
He says that because they won 91 instead of the expected 102, Judge’s WAR has to be docked at least 11 percent.
My question: is a team’s pythagorean a factor in how WAR is calculated, or is James taking liberties because he is James?
As I read James, Voomo, and as I understand WAR, the pythagorean factor does not figure into WAR, but WAR calculations will correlate to the pythagorean projection, rather than to actual wins, something very different from James’s Win Shares formula.
Maybe I am missing something here but according to what Voomo said in the comment immediately above the team’s pythagorean does figure into the WAR calculation.
Yes and no; at no point in the WAR calculation is the “Pythagorean theorem” actually calculated. Rather, the formula attempts to account for things at the level of RUNS, rather than WINS; since the Pythagorean formula and WAR are both gathering their data from runs, they correspond much more closely to one another than they do to wins at a team level. So, in a way, those who are saying the PT is used are correct, as are those who are saying that it isn’t. Your mileage may vary on how you’d prefer to think about it.
Voomo, I want to try to be a little clearer — although my lack of understanding may make that an illusion.
The unique feature of James’s Win Shares system, as I recall it (and I just re-read the intro to his 15 year-old book), is that after every hitter’s batting win shares, every pitcher’s pitching win shares, and every fielder’s fielding win shares have been calculated, the figures are adjusted so that the total for all individual players on a team match the actual totals of team wins (times three, but that’s just a constant of convenience). So, most basically, the system is about allocating credit for real team wins, rather than assessing the value of individual contributions relative to a league figure (average or replacement level), adjusted for home park, etc. As I understand it, team “wins above replacement” are not derived from or limited by actual team wins, although there will usually be a general correlation because strong/weak individual performances tend to produce many/few team wins — except when they don’t, which is the case when pythagorean expectations fall short of or exceed real results.
So in the case of Aaron Judge, you can’t treat his performance as if he were producing actual wins above replacement if, in fact, the “wins” never materialized in wins. When you adjust hypothetical “wins” above replacement by limiting the available wins to real Yankee wins, Judge’s figure has to be reduced by 11% — the “wins” (above replacement) he was credited with didn’t exist in the real world.
In his post, James tries to show the Justice in Judge’s case by showing how Judge was, in fact, responsible for a large share of the Yankee underperformance — actually more than 16% — as shown through analysis of the leverage value of his various batting acts. Pythagorean misfires are generally matters of productivity/leverage mismatch, and so Judge’s poor showing on a leverage analysis ties him to that result. . . . I think I’m now on ground you’re actually more familiar with than I am so I’ll stop.
Hi all – Haven’t had time to participate here for a while but wanted to add my thoughts to this discussion:
1) Seems to me that WAR and Win Shares are trying to answer two different questions. WAR is trying to answer the question: “How much better was Player X than a replacement level player?”, whereas Win Shares is trying to answer the question “What portion of a team’s wins was Player X responsible for?”. Part of the problem may be nomenclature. WAR might be better recast as PAR – Performance Above Replacement.
2) Win Shares has the potential to lead to absurd conclusions. Take two teams, Team A and Team B. Team A goes 162-0, winning every game by one run. Team B also goes 162-0 but wins every game by 10 runs. Win Shares would say those two teams are exactly equal. But does anyone actually believe that to be true? In a head to head series of sufficient length of eliminate random variability, Team B would absolutely dominate Team A.
3) In the comments on Tango’s article, Guy gives concrete examples in which seemingly meaningless hits turned out to actually play a role in the eventual outcome of a game.
4) One issue with both WAR and Win Shares is that they look at each game as an isolated circumstance, when in fact we know that what happens in one game can effect what happens in subsequent games (particularly with modern bullpen usage). Let’s take the following hypothetical:
Bottom of the 6th, the home team is down 7-2, no outs, runner on 1st. The batter hits a two run home run, cutting the lead to 7-4. The manager decides to pull his starter since the game is now closer and the bullpen pitches the final 4 innings, with no more runs scoring for either team. So that 2 run home run was completely meaningless right, since it didn’t affect the outcome of the game? But what if it affected the outcome of the following game? Because the home run resulted in the visiting team’s bullpen pitching 4 innings, the bullpen is now depleted, helping the home team win the subsequent game. In an alternate universe, the home run isn’t hit, the starter stays in the game through the 8th inning, the back end of the bullpen pitches the 9th inning, and the visiting team’s bullpen is still fresh for the next game.
WAR likely overvalues that home run since it values it the same as any other home run. Win Shares though likely undervalues it.
5) Neither Win Shares nor WAR looks at how batters affect their teammates’ performance. An example. Aaron Judge took 4.47 pitches per plate appearance, second most in the majors among 146 qualified batters. Jose Altuve, on the other hand, only took 3.49 pitches per plate appearance, 141st in the majors. Those extras pitches that Judge took have to have value don’t they? Tiring pitchers out so that his teammates get better pitches to hit.
Anyway, just a few thoughts on the issue at hand. 🙂
David P:
Wondered what happened to you, since several threads you missed here have been just your meat. What you’re saying is basically what I was planning on saying in response to epm two comments below, that not only does context supply meaning, but that in a baseball game—and in life outside baseball—every action (possibly every thought or emotion as well, but that’s a side issue here) has potential repercussions moving forward. You can’t isolate the leadoff homer from the game winning homer in the ninth, nor from what happens in between. Nor, as you argue, from what happens the following day. The Cleveland 20-game run this season was hardly a series of unconnected contests, nor was the team’s collapse against the Yankees in the playoffs. Something not said here so far, but a thing that the stats worshippers implicitly deny, is that men aren’t mechanical devices but responsive, living beings. Below I remark that context supplies meaning, but it also has the power to drive the situation, given the responses of the men involved.
nsb, I don’t actually see any important divergence in our ways of thinking, so I have a hard time finding a focus of disagreement to analyze. Both of us seem to me to be pointing to the importance of sustaining awareness of the individual qualitative factors that bear on the human exercise of skill in concrete situations, which lie behind the elements that are homogenized to produce the abstract quantitative reductions we rely on for comparative assessments over the course of a 162-game season or 20-year career.
I guess maybe the narrower point I’m trying to make is that while context is always significant, “it” isn’t always significant in the same way (or, perhaps, better to say that “context” does not always denote an identical thing). The specific example I keep returning to is that I think there is different significance to contexts about early-inning/late-inning acts and contexts about close-game/blowout acts — both are contextual issues, but I think the context-significance is different between them.
However, to shift issues slightly, I do want to resist what seems to me an implication of the comments both you and David are making (one I don’t think either of you actually intends), which is that the flow of human events over the course of a game, series, season, or career is too complex and unknowable in specifics to reduce to quantitative measures. I agree that reality cannot be reduced without cost, but the cost may nevertheless be worth it if it allows us to add analytic assessments to narrative descriptions (and since everyone here, statisticians and non-, like stats, I think this is not a controversial statement). So long as we don’t give up the narrative perspective and keep exploring specific game logs, play-by-plays , and so forth, we can have it both ways.
This is turning into a real discussion! These are great observations, David. Let me comment on them one by one.
1) This seems to me exactly right: WAR does not actually seem to be about wins, but about deviations from norms of performance. The formula derives wins from runs, and it may be more appropriate to link the unit to runs directly (RAR).
2) As you note in (1), Win Shares is trying to do something different from WAR: it is not measuring player quality directly, it is measuring player contributions to actual winning outcomes. So the situation you envision may only seem absurd if you treat Win Shares like WAR: the question for James is not which team is “better,” but which team won more games and how each player contributed to that outcome.
As for your hypothetical, I think it needs some refinement. As you present the example, the teams must be from different leagues, and the strength of schedule may account for the apparent differences in team quality. The outcome of a series between them could be as counterintuitive as the 1906 World Series.
Even so, let’s assume that the teams have played equally strong opponents and reached the result you stipulate. As nsb points out below, in real life there are factors of human psychology that bear on outcomes in ways statistics can’t capture. When a team of power hitters that has played nothing but winning blowouts faces a team that has made scraping its way to narrow victories an inevitable habit, which is better prepared for the match-up? What happens when the scrabbling team encounters strength beyond what it has faced before; what happens when the power team discovers a team that doesn’t allow itself to let down if it falls behind?
3) I assume that the hits Tangotiger Blog’s “Guy” describes that you note are Judge’s early-inning home runs, hit when the Yankees were far ahead or far behind in games that later became close. I think the key issue is the degree to which Judge (or any such batter) is responsible for the ultimate effect of his hit. When WPA is low, the value of the hit in the abstract is still credited to the player via WAR, but the value of the hit in its concrete context is assessed to be very low. Later, it may turn out to have played a crucial role (although it can’t be stressed enough that the patterns of causation involved are not simple), but Judge may have had no role to play in those developments, and if he has, then he will pick up WPA credit accordingly. The case is different when an early-inning home run significantly alters win probabilities at that point, as the batter has indeed played a direct role in that shift. The issue, it seems to me, concerns how much weight to assign to various roles that involve similar acts in different contexts. I’m uncomfortable with the way WPA distinguished between early-/late-inning contexts, but I’m comfortable with the way it distinguishes between degrees to which PA outcomes alter the state of the game at the points where they occur.
I also want to note that I do see “Guy” suggesting that a player who hits a lead-off triple and is stranded after three strikeouts has, in the perspective of hindsight, “wasted” his triple. Guy’s a smart guy, but I don’t think that’s a smart analysis: it assigns the responsibility for the waste to the wrong player, and the notion that the player has somehow misused part of his “quota” of annual triples seems equally wrongheaded. (I’m not suggesting you agree with Guy’s approach here.)
4) This is a really good point, but I think it relies on assumptions that are very hard to verify. To illustrate: in your example, the bullpen is weakened from the starter’s early exit and loses the game the next day. But when the starter returns four days after that, his arm may be stronger for having thrown 30 fewer pitches in his previous start, allowing him to go further into the game, saving the bullpen for the day after that. The flow of events from game to game over the course of a season is of a complexity that goes far beyond knowability, and if we stretch too far in assigning new value to one set of elusive factors that have come to our attention and for which we have devised some new quantitative measure, we risk devaluing by contrast other sets of factors that are either beyond our attention or beyond our ability to reduce to quantitative measures.
5) Bearing my last comment in #4 in mind, I’d say: Yeah — this looks like a very interesting stat to track and build into a WAR-like master value. But perhaps we should think through what other factors we might overlook in that reduction. For example, what proportion of Judge’s extra pitches were second strikes, shifting the advantage to the pitcher in exchange for a one-pitch toll on his arm? (That may or may not be relevant for Judge, but might be relevant if we want to assess the implied assumption that more pitches simply means better team-batting outcomes.)
Anticipating that you’ll be pointing out flaws in my reasoning, it’s good to have you back contributing!
Regarding the value of hitting well in blowouts vs close games, sure, those late & close numbers certainly favor Altuve, and make Judge look like a rattled rookie. But to illustrate the point made by David P above, here’s an example of Judge hitting his 2nd homer of a game, a three-run shot with the Yanx ahead 8-4:
https://www.baseball-reference.com/boxes/NYA/NYA201705020.shtml
After that, NYY used the back end of it’s pen to close it out (Layne/Holder).
And the following day, after coming back from 6-0, Chapman was fresh to polish off an 8-6 win.
There are too many psychological and strategic variables in the game for any one number to describe a player’s value. It sure makes for a fun discussion to try, though.
(Are you reading this, BIll James? This is fun!)
Good point Voomo.
I agree: it is a good point, reinforcing David’s. But once we move into the way a blowout changes the calculus by shifting value from today’s game to tomorrow’s, we can’t restrict that to one team.
In the case Voomo cites, while Girardi’s first thought may have been that the Yankee bullpen A-team can rest up for tomorrow, Gibbons’ first thought may have been the same for the Blue Jay bullpen, so Loup comes in instead of, say, Leone, and the net “tomorrow” value becomes the difference in predicted performance quality by a better rested Yankee bullpen A-team vs. the Blue Jays’. In the case of two teams, this may often work out in favor of the team that loses today, suggesting that the HR, which has modest positive value for today (since it merely changes a commanding lead into a blowout), may have larger negative value for tomorrow. But that’s not the mission that the batter who hits the HR is pursuing, and it would be perverse to assign his HR negative value.
(Parenthetically [hence the parentheses] a somewhat similar perverseness actually does apply under current stats when a batter gets negative WPA for a hit because it leads to a runner on base being thrown out trying to take an extra base, something outside the purview of the batter’s mission and control.)
Voom,
To his credit, when asked what he would do if confronted with a statistic that measured everything about a player’s value, James supposedly responded, “Come up with another statistic”.
epm:
Re your commentary on my commentary:
The remarks on Kluber were merely by way of pointing out some areas of his performance that were less that spotless, since he was so damn good generally. I do think that home/road splits are meaningful, insofar as the closer a good player comes to performing at a high level in all venues, the more admirable he is. Players whose home park is outrageously skewed one way or the other, of course, should get a partial pass in this regard.
As far as the .500± comments are concerned, yeah, there were far more losers in the AL this year than winners, but that doesn’t mean that the stats of pitchers on winning teams mightn’t have been boosted a little by the fact of it.
The point I didn’t make about Sale was that he did so well in Fenway for a lefty that, if he was held back, it was through stupidity on the part of the manager.
Santana. Your championing of him is OK by me, but my point was simply that if he had gone 2-2 with a 3.66 ERA in April, on the basis of the resultant stats he wouldn’t be in the mix for the award, however much his manager may have held him back (Parenthetically, Carrasco seems to have been pulled early quite a bit, too). And I can’t agree that all human events exist in and of themselves, isolated and equally important. Context supplies meaning. A bomb exploded at a testing site does not have the same impact as a bomb exploded in a mosque. Or, if it does, all life and effort are futile and meaningless, whether I kill you or let you live are opposite side of the same coin and nothing else, and, as ‘Guy’ on the Tango site attempts to argue at one point, the guy who triples to open an inning is as much at fault for not scoring as the three teammates who struck out after him. Can’t buy it. Sorry.
nsb, Thanks for keeping me honest and sending me to the James link; I had not actually gone there — I read the Tangotiger Blog after visiting family went to bed over the holiday, and I couldn’t risk staying up later by checking out James. But the James post is a really important one that raises a terrific issue I’ve been wondering about for years. I read “Win Shares” when James published it and thought it was great, and I’ve never been clear on the relation between WAR and Win Shares, or why Win Shares seems no longer to be a major part of advanced stats. It seems James is now ready to address that issue, saying that the only reason Win Shares dropped away was because he was being a Prince of a Guy (ok: apologies to James; I really admire him, but that was too much temptation to resist). My own feeling is that the best outcome would merge both forms of calculation; each captures a valid perspective, and the key would be figuring out a formula that can represent relative validity mathematically, but it’s just a feeling: I’ve lost control of Win Shares and never had much control of WAR, so I don’t know what I’m talking about anymore (I believe I have a better claim to your screen name than you).
There are too many issues I’d like to discuss to fit into this comment, but your point about context supplying meaning is at the heart of all of this. I agree with you entirely, I really do, but I think we are construing “context” in different ways. For example, I don’t think the inning in which a home run is hit matters very much — I think there are context issues around the margins, but that their weight is not great. But I do think the score when a home run is hit is a context issue that matters significantly. I don’t think BA with RISP gives us a lot of information — some, but not as much as it may seem — but I think play-by-play charts that show us under what circumstances batters performed well or poorly with RISP tell us a lot. (To borrow your analogy, we can stipulate that a “testing site bomb” doesn’t have the impact of a “mosque bomb,” but if the testing site bomb was intentionally set off by surprise when the head of state was visiting and the mosque bomb was set off when the structure was empty, the “categorical” context has led us towards a wrong conclusion: we need the play-by-play.)
I do think there are clutch hits, but I think the circumstances of a clutch hit can be varied. To give the simplest type of illustration, a clutch hit can be construed as a product of unusually focused skill under pressure by the batter, or as a product of unusually inept skill performance under pressure by the pitcher. It may be a combination with degrees of both (my default theory would be that it usually is), or it might be simple “luck,” which term I’d only use if the outcome was the product of some phenomenon outside the skill control of either player, such as a gust of wind leading the ball towards the batter’s area of strength and away from his area of weakness, which the pitcher would have successfully targeted if the gust had not occurred. (I think James has an idea of “luck” that makes the term nearly useless, and I don’t like that section of his post.)
When a batter comes up with the bases empty in the first inning, facing a fresh starting pitcher, and hits a home run, he has maximized his opportunity to help his team. If his teammate comes to the plate in the bottom of the ninth with the game tied 1-1 and hits a home run off a fresh closer, he too has maximized his opportunity. Without Player A’s performance, there is no way to know whether Player B’s performance would occur (something that “Guy” on the Tango blog does not understand), but as things actually stand, it is Player A’s home run that has created the team basis for winning the game on a solo home run in the ninth. Why is Player B’s home run more valuable? I’m willing to grant that it may be, because the narrative tension under which Player B performs creates a threshold for him to cross that Player A didn’t face in the first inning, but I’m not willing to give that too much weight, because it could just as easily be that Player B didn’t show what a clutch guy he is so much as Closer A for some reason “choked” at that point. Of course, if Player B does this ten times a season, those data alter the calculus.
I don’t think it would be useful to make this comment longer, although nothing would be more natural for me than to keep going (as my family would assure you). But in the past, HHS has had long strings on these sorts of topics in which we challenged each other to such length that the arguments seemed to me to become clearer and much better configured (unfortunately, they sometimes also were undermined by short tempers rarely seen on HHS). I’d really welcome it if we all wanted to see how well we could do arguing these things out at length, and modifying our positions as good correctives emerged. I appreciated how the posters on Tangotiger Blog were tackling the issues as a group, even though I felt that they were headed in a wrong direction.
I like long arguments. Perhaps some content here just teeing up an area of controversy (along with regular postings) might add more traffic and substance?
I thought the BBWAA got the top three correct.
Kluber
Sale
Severino
Verlander
Carrasco
My vote:
1) Kluber—head shoulders above the rest.
2) Severino—pitched well generally, but especially down the stretch in the chase after Boston.
3) Verlander—on the basis of his September in Houston.
4) Carrasco—went from 10-5, 4.06, to 18-6, 3.29, from Aug 11 onward. On August 10 the tribe’s lead was only 3.5 games.
5) Sale—Flashy strikeout total aside, in Aug-Sept he went 4-4 with an ERA that rose from 2.37 to 2.90.
So, yeah, I’m sort of making a case for the CY as a pitching MVP by dropping Sale so low.
I can’t see any challenge to Kluber on the top spot, and my thoughts are mostly like others’, but I’m going to be a little contrarian on Ervin Santana.
I think Santana may be failing to attract support because his contributions were frontloaded in the season. I don’t think that’s a good approach. Minnesota may have fallen short in the end because Santana ran out of gas, but they wouldn’t have been in wild-card contention late in the season if he hadn’t contributed such a terrific start. The net contribution was the same as if all those shutouts and complete games had come at the end, but too late to quite get the Twins into the divisional contention.
Moreover, there seems to me an odd pattern in Santana’s season that I don’t understand, but that suggests his record could have been better but for managerial decisions. Santana had nine no-decisions to go with his 16-8 record. In those nine NDs, his ERA was a very strong 2.44, and in most of those cases, looking at the play-by-play, I don’t see compelling reasons to lift Santana. He was over 100 pitches in only three of those instances, and in those cases his ERA was 2.74, meaning he was lifted with fewer than 100 pitches in six games where he had a 2.34 ERA. In two of those sub-100 pitch games he’d yielded zero runs and was not in a jam. Obviously, there was no pinch-hitting reason for lifting him — it does seem to me that had he been allowed to go further into those games (and he was the league’s CG leader), he’d have been a much stronger Cy Young candidate. (Of course, there may be reasons that Molitor had for his decisions that don’t show up in the play-by-play.)
In any case, with that in mind, Santana is going to add some slight spice to my otherwise bland list:
1. Kluber
2. Sale
3. Verlander
4. Severino
5. Santana
I totally agree with your points about Santana. I watched two games where I thought he was pulled early this year. While hee did certainly frontload his season, it was awfully good. I’m not sure I’ll have space for him, but I think your points are absolutely spot-on.
I appreciate your saying so, Doom.
1. Kluber
2. Sale
3. Severino
4. Carrasco
5. Verlander
1. Corey Kluber – An outstanding year.
2. Chris Sale – Actually posted a better fWAR than Kluber. It’s easy to dismiss him because of his performance down the stretch, but I could see myself being persuaded into putting him in the top spot.
3. Justin Verlander – It’s hard to pretend I don’t know how the playoffs turned out. I admit that probably has some influence on Verlander this high; still, it was a nice comeback season. I’m rooting for a HOF career for Verlander, and I fear it will take another few seasons like this one. Here’s hoping!
4. Carlos Carrasco – see below
5. Luis Severino – Carrasco and Severino had the same season; Carrasco had a marginally better SO:BB ratio, and that’s how I made my decision. Severino also had a minor edge in BABIP, and since their other numbers were so similar, I couldn’t help but give Carrasco an edge. I could’ve put Ervin Santana here to for that flaming-hot start, but his dreadful FIP scares me a little.
There’s a lot of good discussion going on here, and (even after the vote I just posted) less than ten votes up on the site. I’m going to make the executive decision to extend our deadline by another 24 hours. New deadline is TOMORROW night (Tuesday 11/28) at 11:59 PM. (As always, that’s a flexible deadline; the hard-and-fast deadline is “whenever I get up on Wednesday morning, so don’t push it too much.
I would like to add some thoughts to the WAR-Win Shares discussion above, but I want to do so outside the parameters of the discussion itself, because they’re really side-issues dealing with calculations, and not at the heart of the discussion. In no particular order:
1.) WAR does have a nomenclature problem. I feel confident that if you took ’21 Ruth, ’01 Bonds, ’12 Trout, ’80 Brett, ’91 Ripken, ’24 Hornsby, ’31 Gehrig, and ’72 Bench, then add in ’01 Randy Johnson, ’72 Carlton, ’85 Gooden, ’97 Clemens, and ’68 Gibson, you end up with a team with more than 162 wins. Obviously, that doesn’t make sense. As mentioned above, WAR is not about team wins, it’s about performance above a replacement player. So why not just use RAR? When it first started, that was my impulse. I would STILL like that, actually. But the reason is simple: run environments. A Baker Bowl run in 1930 and an Astrodome run in 1968 are two very different things, but one game is still one game, and therefore one win is still one win. Even in less extreme cases (like within one individual season), you find that the runs-to-wins calculation is different. Yes, you COULD solve this issue by adjusting to a “historical run environment…” but that is literally what WAR is already doing. So if you want to, multiply it by ten and call it Normalized RAR, but you end up with the same thing. Maybe the name is bad, but the math makes sense, in its own way.
2. I’m a religious reader of Bill James’s website, but he can be a mean cur when he wants to. He’s pompous and believes he’s never wrong, and he can be as intolerant of other ways of thinking as anyone. That said, he’s been a profound influence on how I think, and I will continue to read him, because as obnoxious as he can be, he also still occasionally has absolutely brilliant insights into how to think, about baseball in particular. Anyway, he had some real criticisms of Judge’s season in relation to Altuve’s, and they should be considered in an MVP discussion. However, he was asking WAR to be something it isn’t in that post, and it’s tremendously unfair. Sean Foreman and Dave Cameron wrote beautiful defenses of WAR; Sean’s is particularly worth reading, and it also has links to about a million other pieces on the subject, all of which are worth reading (least so the Baseball Prospectus one, which I think fundamentally misunderstood some of Bill’s arguments, but that’s neither here nor there). WAR makes choices, just as Win Shares does. On the balance, I think WAR’s choices are BETTER. One of those choices is to adhere more closely to team runs than team wins. It’s a choice. You can criticize it, but it’s a reasonable choice. Bill does not believe so, and it’s a place where I think he just hasn’t thought through the issue in a balanced way.
3. Finally, someone above asked, “Whatever Happened to Win Shares?” I can’t find the comment, but I can answer with some really basic fundamental flaws with Win Shares, that subsequent analysis (including WAR) has thought better of. First of all, as I stated above, Win Shares forces team wins rather than team runs; this is a choice, not a reason it’s been passed by. Just a difference. Anyway, the more important choices come down to some arbitrary choices Bill made. First, offense received 53% of all Win Shares; defense received 47%. This was an arbitrary choice Bill made, but not in the way that I said above. This is an arbitrary choice which fundamentally changes how we think about baseball. It says that somehow runs scored are worth more than runs allowed. That’s not how baseball worked. Basically, when it was all said and done, Bill didn’t like the numbers, so he changed them. It’s right there in the introduction to the Win Shares system in the NBJHBA. Second, he lauds his defensive analysis in the book; he says that it’s lightyears ahead of any defensive analysis that came before; he’s not wrong. It WAS. The problem is, in the time since, much BETTER defensive analysis has come about. TotalZone and Michael Humphries’ DRA are both better uses of traditional defensive statistics. As always, kudos to Bill as the godfather; however, just because Newton had the first great physics insights, that doesn’t mean their more correct than Einstein’s, you know? Anyway, the final problem I’ll point out is, to me, the most egregious. Win Shares compresses all performance, such that negative numbers are not allowed. This is a problem. If a guy playing first base never catches a ball thrown to him all year, that’s not a zero; that’s a deeply negative number, because there are literally millions of Americans who could do better. I’d say 80-85% of the time, it makes no difference. But then Ralph Kiner and Dave Kingman get their numbers inflated because clearly costly defensive performance is not actually taking enough away; similarly, players like Cal Ripken have their numbers compressed, because negative defensive players on their teams are being scored with positive numbers, which makes Ripken’s own numbers slightly less-positive than they should be. These are major drawbacks that really hurt serious analysis. Not so much that it ruins things; you know, as it turns out, Ted Williams and Joe DiMaggio and Hank Aaron are great, no matter how you measure. But there are reasons that other, superior systems took up the mantle in the popular imagination – they make better choices, on balance. The wins/runs choice is fine; the others, not so much.
Great comment, Doom. I spent so long bloviating on my comment below that I didn’t see it till after posting. I appreciate your pointing us to Sean Foreman’s response to James.
By the way, I’m the one who wondered whatever happened to Win Shares — after all the effort of reading James’s original book on the subject, I felt distressed that I never got to show off that I’d done it.
I do think that there are deep structural strengths and flaws in both Win Shares and WAR (or RAR), and that they are not likely to be reconcilable. If that’s the case, the optimal solution may be a system that provides some weighted combination of two irreconcilable, significantly valid perspectives.
Wow.
Some really interesting and informative discussions happening. I may not be any smarter than I was before but my ignorance and confusion has reached a higher level than ever before. This is how I see it:
1) Kluber- as others have said before me, there’s just too much here to come to any other conclusion
2) Verlander- I have to admit that because of personal prejudice I tried to come up with a logical reason to make him my top pick.
3) Sale- again, just too much to overlook. He also deserved serious consideration for the #1 spot.
4) Stroman- I’m a little baffled why he’s not a bigger part of the discussion.
5) Carrasco
There’s only a tiny difference between Carrasco and Severino but I think it favors Carrasco plus I didn’t think it reasonable to penalize him for having Kluber as a teammate and I have to admit to a slight anti-Yankee bias so it wasn’t a hard decision.
I also gave Santana a hard look, largely because of epm’s comments. I may not have voted for him for the Cy Young but it’s entirely possible he would have been on my MVP ballot over several of those that I did.
You’ll still have your chance, Hartvig! The NL Cy Young is next, but then we get the AL MVP!
Agreed. Great last 48 hours on this site.
I even learned a fantastic new word: bloviate
I’m picking up the set of exchanges about WAR vs. Win Shares, Altuve s. Judge, etc., that appears midway on this string and began this comment hoping to make it easier to continue that discussion by adding to it with a new bottom-of-the-page comment here. Unfortunately, I may be squelching discussion by adding a comment that’s simply too long. But anyway . . .
A lot of the discussion in the earlier exchanges was prompted by the Tangotiger Blog and Bill James online blog. I wanted to note that the Tango blog also includes a link to a post on Joe Posnanski’s blog that seems helpful to me. In it, Posnanski raises the issues that David P raised in his point #1 (and says much better than I what I meant to convey in my response agreeing with David). I thought the post was worth linking to directly here. (As a side point, I want to mention that Tango blog poster DavidS [comment #11] raises some of the points I was trying to make about using retrospective measures to assess value, except he does so more clearly and in about one-tenth the space.)
But the main issue I wanted to raise again here after reading Posnanski’s comments was the matter of what James calls “luck” and Posnanski calls “chance.” We’ve discussed this many times on HHS, but I continue to think it is a barrier to clear analysis of the contributions players make in the course of a game.
I suppose somewhere the major baseball bloggers have defined what they take “chance” or “luck” to mean, but I don’t know where they’ve done that, and I think part of the problem we have in thinking about baseball performance is that we have no consensus on what constitutes luck/chance. I’m going to try out a definitional strategy here., and I’m going to define the terms “chance” and “luck” to denote different things. The entire argument of this long comment is just a shot at clarifying issues by trying to define these terms.
Most (for simplicity) baseball events result from interplays among at least three elements: 1) the attempt by a pitcher to employ his pitching skill to make a batter fail to get a base hit or progress towards a walk; 2) the attempt by a batter to employ his batting skill to get a base hit or progress towards a walk; 3) the attempt by a fielder to employ fielding skills to make a batter fail to get a base hit. Any event that is the result of success or failure in the deployment of these skills, which means most baseball events, seems to me outside the range of chance.
Let me use a concrete case (one I’ve tried out before on HHS, that eliminates fielding skill for simplicity). The batter hits a line drive that falls one yard beyond the reach of an outfielder. Is it chance that the ball was hit so that it dropped in rather than being caught? Let’s suppose that the batter is Mike Trout and the pitcher is me (imagining I can still throw a ball sixty feet). If I’m throwing my best fastball (34 mph), my skill level is so low that Mike Trout will be able to look at it arching in and say, “I’m going to punch that nonsense right to the left of the outfielder’s reach.” No problem, and no chance involved. The placement of the ball is entirely within Trout’s control because my skill level is so low and his is so high. (Think about how easy it is to hit fungos to a kid just about where you want to: the “pitch” offers no challenge, and an average adult has the skill to get within a couple of yards of where he — or she — means the ball to go.) But when the pitcher is Corey Kluber, Trout can’t predict where the ball will go; he can’t even know whether he’ll make contact. Why is that? Has chance suddenly trotted out onto the field? No, a skilled pitcher has. It is Kluber’s skill at pitch delivery that weakens Trout’s control over ball placement. Now, should Trout hit the ball a yard from the left fielder’s grasp it is true that his skill was not adequate to control the exact placement of the ball, but the exact placement of the ball can be expressed as a product of Kluber’s pitching skill in attempting to deny Trout a hit and Trout’s batting skill in attempting to get one off of Kluber. What does chance have to do with it? You might think chance expresses the fact that the ball dropped one yard east of the fielder’s glove rather than in the glove itself, but I think that’s an error: the precise placement is an exact product of the goals of the two players and their combined skills in attaining their goals. The reason we are tempted to call it chance is because we don’t understand exactly how the ball came to be where it did, and neither do the pitcher and batter: no one person intended that specific outcome, so it matches no one specific human intention. But it was nevertheless entire the product of the human intention, modified by the level of skill each participating human was able to bring to bear to attain the goal each was intending. What does chance have to do with it?
“Chance” denotes outcomes that are not products of agency; it should not denote outcomes that are action products of multiple human agencies. Sudden gusts of wind or infield pebbles that affect a batted ball’s trajectory are examples of chance. Ordinary balls in play are the resultants of two vectors of skill agency (we can add the fielder[s] back in later).
But there is another angle to all of this which I think is better served by the term “luck.” I’ll try to be briefer: I’m going to define “luck” as an outcome that affects one of the actors in a play independent of his own agency, but through the agency of another player. If Judge hits a bomb to center and Trout makes an unbelievable circus catch to rob him of a home run or extra base hit, it’s Judge’s bad luck. His skill delivered a result that overcame the pitcher’s skill and the skill of an ordinarily talented fielder, but an exceptional effort on Trout’s part, something over which Judge had no control, overcame his skill. There’s no chance involved — it’s all a matter of vectors of skill-intentions, but from Judge’s perspective he earned a hit and was denied one. (Now, if Trout’s catch was enabled by a sudden shift of wind that kept the ball in reach, that would be chance.)
As a final example, if Sale pitches with the flu and Judge finally gets a hit off him, it’s Judge’s good luck that he faced a Sale dulled by diminished skills. But the cause was not chance but pitching skills impaired by illness. We could say that Sale was unlucky to get the flu or contracted it by chance, and this may be true, but to pitch with the flu requires Sale and/or his manager to assume responsibility for the game outcome; it is no longer a matter of chance, and we can’t say Sale was unlucky to decide to pitch with the flu (though he could be unlucky to have been ordered to do so). (However, to echo nsb’s earlier comment, there are behind every player’s conscious performance a host of physical and psychological causes outside our and his ranges of perception that influence events . . . I don’t think we can attend to them; they never emerge into view.)
Given the new tools that Statcast has introduced and improved understanding of probabilities represented by strategies such as fielding shifts, I suspect that at some point during the 2020s we’ll begin hearing less about chance and luck, find that outcome-based stats lose a little of their centrality, and confront a new set of stats that measure how well each baseball act conforms to ideals of maximized probabilities — for pitchers, aggregates of pitch type/placement relative to batter strengths/weaknesses; for batters, batted ball velocity and launch angle aggregates; fielder sprint speed and path efficiency, and so forth. Perhaps then our assessments of players will give credit both for outcomes and for the quality of skill performance relative to intent.
epm: I am so impressed with your comments that I have created a Word document for them for my future perusal. I have a ways to go until I fully understand WAR. I have also included comments by some of the other commenters.
Wow — Richard — thanks! I have at least as far to go on WAR as you, and I’m beginning to think it’s time I invested the effort to work through to the details.
epm:
Two points: 1) Should we start referring to this site as HHE—High Heat Existentialism—from now on? 2) Do you think anyone over at the Tango site or the Posnanski site, Bill James, or even John Autin, are following this thread?
The whole discussion, anyway—from the cited James article to your comment above— is good in that it at least has had the effect of pointing out a basic flaw in WAR, something I began arguing years ago and gave up on in frustration, that WAR is mis-titled. If the people who titled it don’t even understand their own use of language in this regard, how can we, who are denied the priestly mysteries, accept uncritically the orthodoxy they espouse? “Trust me” is the last refuge of the obfuscater, the con man, and the politician, not just the expert in complex analysis. I personally don’t give a free pass to any of these people. I grant that WAR is a useful tool in evaluating batting especially, but —to get back to a an issue I raised in an earlier thread—the uncritical reliance on WAR that more and more people seem to have adopted in place of critical evaluation is disturbing. I might also add that Bill James’s Win Share idea is also a useful tool. Both it and WAR are flawed, but both, if taken as tools, not as truth, are helpful in reaching understanding.
nsb, I thought I was an existentialist in the 60s, and took a college course on it, which I recall mostly for two things: (a) my embrace of the existentialist zeitgeist and the freedom it said I was condemned to, skipping virtually every class; (b) the D on my transcript.
I don’t suppose Posnanski and James have time for HHS, but John, if you’re lurking in silence (and anyone could understand why you’d treat HHS as though it were the object of a 12-step program), we miss you!
I think you’re right, nsb: most of us ultimately take WAR on faith, and that’s unhealthy. I don’t think the folks at B-R or Fangraphs intend to suppress criticism, and I admire what they have accomplished, but I think it would be healthy if both sites highlighted exactly what the shortcuts and weak links are, as James and Posnanski are doing. I’m sure they are aware of them and working on revisions that will keep adding to WAR’s validity. Maybe now that Bill James has decided it’s time to push for his own system, also flawed, perhaps some momentum will build for the next generation of WAR, RAR, whatever. If so, I assume it will debut the day after I feel I’ve finally figured out how WAR works.
You and your 34 mph fastball and the ball lands 1 yard outside of the outfielders reach?
I was not aware there were any ballparks where the outfield fences were 700 feet;)
Hartvig, If you’re in total control, why settle for a timeworn cliche when you can make the ball do circus tricks?
Can I suggest this: https://www.youtube.com/watch?v=IOt2iw-7n6A
Who won?
Here are your 2017 AL Cy Young Results:
There were 8 ballots cast. 7 points for first place, then 4-3-2-1 for the remaining five. Points listed after the name, first place votes in parentheses.
1. Corey Kluber, 56 (8)
2. Chris Sale, 27
3. Justin Verlander, 21
4. Luis Severino, 16
5. Carlos Carrasco, 11
6. Marcus Stroman, 4
7. Ervin Santana, 1
A clean sweep for Kluber!
Five people had all of the top five on their ballot, but no one had them in order. Only one voter had the top four in order (that was e pluribus munu), and he picked a different player fifth! Before we started this, I thought there was a pretty good chance that someone would have a perfect “consensus” ballot.
Kluber, Sale, and Verlander were named on all eight ballots. Severino was named on seven, Carrasco on six, Stroman on two, and Santana only on one.
Obviously, there wouldn’t be any reason to talk about where the first place votes went, but the second place ones were moderately interesting. Sale took five of the eight runner-up votes, Verlander took two, and Severino took the last one.
Looking forward to the comments on the NL post!
For those interested in the WAR discussion, Tango has a WONDERFUL post about Judge right now:
http://tangotiger.com/index.php/site/comments/so-what-is-all-the-hulabaloo-about-war
I agree that it’s a good analysis, Doom. Tango is absolutely right that if we clarify our question we will be better able to discern what metics answer it and what don’t, and not get caught up trying to make the wrong metric fit our question. On the other hand, it’s also possible to claim that questions are not all of equal interest and significance. (It’s a lot like the discussion we’re having about FIP on the next thread. You can say that FIP is better for measuring pitcher stuff and control and ERA is better for measuring pitcher game success, but you can go on then to debate whether stuff & control or game success is more meaningful and significant in assessing pitcher quality.)
But what blew me away on the Tango thread was the comment from ‘Guy”, which included this: “If Judge hit poorly with men on base, some of that debit must go to the teammates who got on base at the wrong time.” This seems to me like saying to the next of kin of an airline passenger whose plane went down: “Well, her death was partly her fault, because she should have chosen a plane that wasn’t going to crash.”
Hindsight allows us to look both forward and backward to appreciate the implications of event chains, but agency and responsibility concern causation, and causation only goes in one direction (at least in our universe).