The One Problem With Kris Bryant’s MVP Case
Kris Bryant is everything you could want in an MVP candidate. He hits, he runs, he plays defense, he moves around, he’s in there every day — Bryant is an outstanding player on an outstanding team. You don’t have to worry about the Cubs wasting him, and not going to the playoffs. If Bryant’s teammates are doing him any harm, it could be because there are too many good ones — even without Bryant, the Cubs would be fine. It speaks to the roster’s strength, but Bryant is the best regular. He’s maybe, or probably, the best all-around player in the National League.
There’s more than a month left to go, so various leaderboards are going to change. Performances will change, and perceptions will go along with them. That being said, Kris Bryant has to be thought of as the NL MVP front-runner. By which I mean, I assume he has the most support. And, what a player to throw your support behind! In so many ways, Bryant would be deserving. There’s just one little problem. That one little problem is basically the entire counter-argument.
Some of you are probably already ahead of me. In part this is because the same sort of argument could’ve been made a year ago, against Bryce Harper. The voters didn’t care, but that doesn’t mean we shouldn’t talk about it. Last week I wrote about how Michael Saunders has been incredibly unclutch. In fact, he’s been one of the least-clutch players of the past 40 years. Saunders has nothing to do with this year’s NL MVP, but the idea in there is significant. See, that post included a table of unclutch player-seasons. I cut it off at 10, but had I stretched it just a little bit more, you would’ve seen 2016 Kris Bryant. The overall numbers love him. Adjusted for context, they’re decidedly less affectionate.
A year ago, it was Harper who was fairly unclutch. Here are this year’s current bottom 10:
| Name | Clutch | WPA/LI | WPA |
| Michael Saunders | -3.2 | 1.8 | -1.5 |
| Kris Bryant | -2.6 | 4.6 | 1.9 |
| Nelson Cruz | -2.6 | 4.0 | 1.5 |
| Randal Grichuk | -2.4 | 1.8 | -0.6 |
| Joe Mauer | -1.9 | 1.8 | -0.1 |
| Brian Dozier | -1.9 | 3.0 | 1.2 |
| Miguel Cabrera | -1.9 | 4.1 | 2.1 |
| Jackie Bradley Jr. | -1.8 | 2.3 | 0.4 |
| Freddie Freeman | -1.8 | 3.9 | 2.2 |
| Evan Longoria | -1.8 | 4.1 | 2.2 |
You should still be floored by Saunders — he’s out there on his own. But then there’s Bryant, who’s cost the Cubs real wins with his lousy timing. Now, as the last column shows, Bryant has still been an overall offensive positive. He’s been way too good to not be helpful. But WPA/LI shows a kind of leverage-adjusted offensive value. WPA shows actual offensive value. By WPA/LI, Bryant has been as valuable as Daniel Murphy. By WPA, he’s been as valuable as Cameron Rupp.
WPA doesn’t factor in defensive performance, or versatility. It doesn’t account for position, and it doesn’t account for the majority of a player’s baserunning. Because Bryant, again, is an all-around positive, just looking at WPA sells him short. But what this gets at is that his hitting this year has been less helpful than the overall numbers would suggest. That’s not spin; I don’t have anything against Kris Bryant. It’s just how things have happened. And just as with Saunders, this is an easy thing to break down. In low-leverage situations, Bryant’s wRC+ ranks in the 99th percentile. In medium-leverage situations, it ranks in the 85th percentile. In high-leverage situations, it ranks in the eighth percentile. Bryant has done the most damage when the results have mattered the least. That’s all this says.
I want to make it absolutely clear I don’t think Kris Bryant is a naturally unclutch player. These splits can be almost random over smaller samples. Bryant this year has been horrible in high-leverage situations, but he’s still been better in those spots than Miguel Cabrera. The second-best hitter in baseball in those important spots has been Yonder Alonso. And, importantly, while Bryant has a Clutch rating of -2.6, just last year as a rookie he finished at +1.9. In consecutive years, Bryant has been one of baseball’s most-clutch hitters, and one of baseball’s least-clutch hitters. Which leads you to the natural conclusion that Bryant is neither clutch nor unclutch, like pretty much everyone else.
This isn’t a true-talent conversation. This is a stuff-that’s-happened conversation. When I wrote about Saunders, I included a table showing his 50 highest-leverage plate appearances. That table was pretty big, but I liked the idea, so here’s a version for 2016 Bryant, only going to 25 plate appearances instead.
| Leverage Index | Result | WPA |
| 9.04 | Two outs | -0.53 |
| 5.93 | Out | -0.30 |
| 5.76 | Out | -0.18 |
| 5.67 | Out | -0.15 |
| 5.22 | Single, run | 0.18 |
| 4.69 | Out | -0.17 |
| 4.61 | Out | 0.05 |
| 4.45 | Out | -0.11 |
| 4.39 | Out | -0.12 |
| 3.98 | Out | -0.11 |
| 3.47 | Out | -0.08 |
| 3.44 | Out | -0.09 |
| 3.42 | Out, run | -0.05 |
| 3.38 | Out | -0.09 |
| 3.33 | Out | -0.08 |
| 3.22 | Out, run | 0.00 |
| 3.12 | Single, two runs | 0.23 |
| 3.05 | Walk | 0.08 |
| 3.04 | Out | -0.08 |
| 3.03 | Out | -0.08 |
| 2.89 | Out | -0.07 |
| 2.89 | Double | -0.03 |
| 2.85 | Walk, run | 0.10 |
| 2.83 | Out | -0.07 |
| 2.83 | Double | 0.20 |
Of those 25 events, 18 reduced the Cubs’ chances of winning, six increased the Cubs’ chances of winning, and one accomplished nothing. The sum WPA here is -1.53. Look at those top 16 events — you have 16 outs, and three runs scored. It’s not quite as absurd as Saunders’ table, but that’s not exactly the benchmark. Kris Bryant has been fantastic. I’m sure the Cubs wouldn’t change a thing. Not about his skillset. But probably about his timing.
I can’t tell you how much to care about this stuff. Maybe you don’t like the idea of partially evaluating players based on their contexts. They’re not really that responsible for the situations in which they find themselves. But this seems like the difference between observing past value, and projecting future value. I would project Kris Bryant to have enormous future value, starting today. But he’s been so unclutch, and Addison Russell has been so clutch, that Russell has a significantly higher 2016 WPA. And Russell has been no less valuable a runner or a defender. You could make a real argument that Russell has been more valuable to the Cubs. It would seem controversial, but then, how controversial is it, really? Just how much do we really want to ignore the timing of events?
A funny consequence of looking at things this way — Corey Seager has also been unclutch. Granted, he’s been less unclutch than Bryant, but still the two remain linked. And because of their unclutch performances, in theory that opens up the MVP race to other candidates. In reality, I doubt this’ll matter. Bryce Harper just demonstrated that. But there’s nothing wrong with thinking. Kris Bryant provides an awful lot to think about.
Jeff made Lookout Landing a thing, but he does not still write there about the Mariners. He does write here, sometimes about the Mariners, but usually not.
While it is no longer as valued as it used to be, many voters still care about RBI. One consequence of unclutch performance is a lack of RBI (I would think). I wonder how much RBI and clutch correlate? If someone ends up with way more RBI than Bryant, won’t that sort of have the same effect on the voting as if they actually looked at WPA?
Weirdly his RBI’s haven’t really been affected by this, Rizzo has a far greater clutch score than Bryant but only 1 RBI more so if there not looking at WPA itself they’ll probably not pick it up.
Harper had only 99 rbis last year and almost noone even noticed. People got angry if you even pointed it out.
Yes many intelligent, sabermetrically inclined people did. Many dumb local sports writers (who were able to vote) would probably still reference RBI. That’s my point.
Haven’t seen JFTG all week, but now that there’s a Cubs article, he’s back… like clockwork. What is his obsession with the Cubs?
I’m obsessed with WPA and Clutch. I also posted about Jeff’s Michael Saunders article about Clutch and that had nothing to do with the Cubs. I just made a post about clutch hitting in the article about Joey Votto and that had nothing to do with the Cubs. Recently I posted about clutching hitting in an article about Bryce Harper which had nothing to do with the Cubs. If you only see my posts when they’re related to the Cubs, it’s because you’re only reading articles about the Cubs, so it looks like you’re the one who’s obsessed with the Cubs.
But RBIs is often correlated to WPA and Clutch, so if you think that this is at least a factor in MVP considerations (and Jeff’s point in this article is that WPA and Clutch can be a legitimate factor), then RBIs are one way to talk to less aware people about things like this.
And Rizzo probably has driven in Bryant…a lot. Bryant has camped out in the #2 hole and seems to do well there. If we slip him and Rizzo Rizzo probably loses RBI but his R total would go up. Vice-versa for Bryant.
For the record, I don’t think RBI should matter, but many people who actually vote probably do think it should matter.
Similarly, many voters think that Jered Weaver’s losses somehow impact Mike Trout’s value and that the Reds giving up all the home runs in baseball makes Joey Votto less valuable. Doesn’t make it right, but that’s how it plays out.
Here is the thing that I don’t like about WPA analysis: it doesn’t look at performance retrospectively. By that, I mean that it doesn’t go back after the game is over and look at which runs ended up being valuable.
For example, say the Cubs are up 4-0 in the 3rd and Bryant hits a 3-run home run. That’s going to add very little to his WPA relative to most 3 run homers. But what if the game ends up being a 7-5 victory. That 3-run home run was the difference in the game. How do you account for that contribution?
Conversely, let’s say the Cubs are down 7-0 in the 6th and Bryant hits a bases-clearing double to make the score 7-3. That’s a relatively low-value 3 runs, so he doesn’t get much WPA credit for it. Now let’s say the Cubs score a few runs over the next few innings to make it 7-6 in the 9th. Even if the Cubs end up losing the game, those 3 runs were valuable because they gave the Cubs a much better chance at winning than they otherwise would have had.
Now, I don’t know exactly how you should calculate value. But I believe that WPA is not the be all and end all of deciding how much a player actually contributed to wins by their hitting.
But they didn’t give the cubs a much bigger chance of winning when they were scored. That’s why there isn’t much wpa value to them.
That’s exactly what he said – and then added that they should, because in the bigger picture they did give them a better chance.
They shouldn’t because it’s not the bigger picture, it’s the much smaller picture of [in this specific circumstance, everything else being unchanged, Bryant’s 3-run home run would be the difference.]
fredfotch’s criticism is much like criticizing a projection that forecasts, say, the Cubs are 90% likely to win today’s game, if instead they lose. It’s ignoring the big picture, ignoring the value of probability, and conflating “overwhelming favorite” with “sure thing.”
If the Cubs are overwhelming favorites, sometimes they will lose–say one in ten times. And if WP only credits Kris Bryant a small amount of WPA for a three run shot in a game the Cubs are leading 4-0 in the third, it’s because the likelihood of that being the difference is very low.
It’s pretty much like all advanced stats, they should never be looked at individually to come to a conclusion but should be looked at in unison with others to help form a conclusion.
This is a valid point, really, another way of stating why context-independence is where value determinations should begin. On a larger scale, you can make the same argument about performance during the season. In 2012, one of the arguments used for Cabrera over Trout, beyond the triple crown, was that his production (context-independent) was better than Trout’s in September, when the playoffs were on the line. But of course, every win counts the same, and Trout’s production earlier in the season helped put the Angels in the position of being contenders in September.
All that said, no one knows in the second inning that driving in runners at a PA is going to tip the final score, whereas that may be very evident in the ninth inning. It’s a Bayesian thing; the importance of the PA can only be judged based on the information available at the time. So I accept the notion of high leverage situations. I think the best reason for not paying much attention to a player’s performance in them–a point Jeff made sure to mention–is that that performance tends to average out over time.
I get what you’re trying to accomplish, but there doesn’t seem to be a good mathematical way to do that. I guess you could adjust closer games to be worth more than 1.0 WPA, so all events in that game are more “valuable.” But still in this context, the first run is just as important in the win as the 7th, really.
But point still taken. WPA only takes into account past events, and you want to take into acount events before and after. I’m sure some smart person can figure out a way!
Sounds like Bill James’s Win Shares. How much each player contributed to wins. Not all that hard to calculate compared to WPA.
Bring back GWRBI?
You can look backwards and ‘delete’ the runs and see if it flips the game, but it’s hard to push that more granular than runs, to get past GWRBI. You’d like a leadoff triple that ultimately scores the deciding run to be a big deal, right? So you could try looking back and replacing each play with an out, and try to see how the rest of the game would have unfolded, but it affects later decisions for a butterfly effect.
Maybe you could do something reasonable enough to be interesting, still.
WPA shouldn’t look at performance retrospectively. Consider one of the examples you’ve given:
“For example, say the Cubs are up 4-0 in the 3rd and Bryant hits a 3-run home run. That’s going to add very little to his WPA relative to most 3 run homers. But what if the game ends up being a 7-5 victory. That 3-run home run was the difference in the game.”
If the Cubs are up 4-0 in the third, WPA knows from its vast sample that that game is not likely to end 4-0. Roughly two thirds of the game is left to be played. But a 4-0 advantage so early is generally more than enough that it will account for the difference. WPA knows this too. That’s roughly equivalent to spotting a team six+ runs from the outset, which (very inexactly) over a full season would lift the 2003 Detroit Tigers to a 120-win team.
Some small number of games will start lopsided but end close, but those shouldn’t be cherry-picked out as counterexamples criticizing the value of WP. In fact, WP very accurately estimates just how likely that 3-run home run factored into the outcome of the game and awards Bryant appropriate credit. That’s exactly what the stat does!
Even in this extremely narrow hypothetical, Bryant’s 3-run home run would not be the difference any more than whatever outs helped preserve the lead–say a rally killing ground ball double play induced by Hector Rondon in the bottom of the 8th–and unlike Bryant’s 3-run home run, which was hit when the game was not competitive, those outs would be recorded with the pressure of it being a high-leverage situation.
If you’re looking for contributions toward winning in this universe, though WP does not adjust for quality of competition, it’s about the best stat we’ll ever get. It shows us the likelihood that Bryant’s hypothetical 3-run home run will be the difference, credits him for that, and also shows us overall how well a player performed in high leverage / high pressure situations.
(and this is ignoring the fact that removing one event from a highly sensitive sequence totally changes every following event, which fact further compounds the error of retrospectively awarding Bryant credit for providing “the difference in the game”)
So if you think that performance should be evaluated retrospectively on the basis of how it ultimately affected the outcome of a game, it means for example that (1) a triple hit in an inning in which no runs are scored has the same value as a strikeout, because neither led to a run being scored, and (2) a grand slam home run in the top of the 9th to tie a game which is then lost in the bottom of the 9th has the same value as striking out, because they both result in a loss. Do you really believe this?
More generally, it means that you believe that no hitting events by a team which ultimately lost a game can possibly have a positive effect, because they were insufficient to win the game and did not have a different outcome on the final event from a strikeout–in either case the team lost. Does this seem reasonable?
I see thumbs down but I don’t see any counterarguments.
I’ve given this and other WPA oddities some thought. How about the guy who is 0 for 3 on the day, all with RISP, and as a result, his 3-run HR with 2 out in the top of the 9th gives the team the lead instead of expanding an already comfortable cushion? That’s a massive positive WPA event, and he would have had less WPA had he broken the game open in the first. I had a conversation with Clay Davenport about this at a SABR convention; the key point appears to be that WPA takes the LI as a given, while negative events can in fact increase the LI and hence multiply the impact of subsequent events.
Having raised and not addressed that problem, there is an easy way to re-do WPA retrospectively, as an alternative lens: take the net change in WPA for each half-inning and re-apportion it by changes in Run Expectancy. It acknowledges that the hitters in an inning are knowingly engaged in a collaborative effort to score as many runs as possible before three outs are made. Therefore, when they’ve finished, you measure the impact that effort had on changing their odds of winning, and you divide up the total impact by their individual contributions to the runs they scored.
I think this is unquestionably one correct way to do it. The only question is how to weight this vs. standard WPA. According to the latter, Dave Henderson’s HR off of Donnie Moore in Game 5 of the 1986 ALCS was 14 or 15 times as valuable as Don Baylor’s HR earlier in the inning, which is plainly wrong. This way, they are nearly equal in value, which is also plainly wrong. The truth is somewhere in between. I have not yet given any thought as to how they might be objectively weighted.
In the first case you mention, the 3 outs with men in scoring position would all produce significant minuses (assuming that the game was close at the time) which would largely offset the homer in the 9th. If he has broken the game open in the 1st and then gone 0 for 3 with RISP, he would get a smaller +WPA in the first inning but if the other team never got close after that, the 0 for 3 wouldn’t have much effect, and his WPA rating for the whole game might well be higher. I’ve calculated WPA by hand myself quite a lot and things like this become apparent if you do.
Regarding the second form of calculation, I’d have to look at a lot of specific cases before forming an opinion. But one general thing to consider is that the present system of calculating WPA gives a big penalty for failing to hit in high-leverage situations. If Baylor had gotten out, the -WPA would have been relatively small, but if Henderson had gotten out, the -WPA would have been much greater, so before concluding that Henderson was rewarded too much, you should figured out how much he would have been penalized if he’d gotten out.
It’s trivial to construct examples of the sort I mentioned. A guy comes up with men on 2nd and 3rd and 2 outs in the first and is retired. In the top of the 9th, he comes up with his team down a run, a man on 1st, and 2 outs, and hits a go-ahead HR. WPA for these 2 PA (1/2, HR) is -.057 + .710 = .653 (based on the tables in The Book). If, however, he singles in two runs in the first and is out trying to get to second after the two runs score, and then hits his 2-run homer with his team leading by a run rather than trailing, in these 2 PA (2/2, HR) he has a WPA of .161 + .153 = .314. Singling in the first has a net -.339!
In the 1986 game, Baylor was facing a -.026 WPA had he made an out (-.038 if he hit into a game-ending DP) and earned +.059 by homering. Henderson was facing -.082 had he ended the game and earned +.710 (The Book) or +.73 (b-Ref) by homering. Had Baylor made an out, the game would have ended before Hendu hit, but the next hitter (Dwight Evans) would have been facing -.012 versus a paltry +.027 . None of that is surprising, and all of it supports the claim that Hendu’s homer was obviously neither 50% (RBI) of the rally nor 92% of it (WPA), but somewhere in between. Whereas Baylor’s 2nd out, had he made it, does seem about twice as damaging as a subsequent 3rd out.
The thing is, if I understand WPA correctly, it simply says when the player’s AB was done they had X better/worse chance of winning the game. What happens before or after is irrelevant. A team with a 4-0 lead is already likely to win or trailing 7-0 is unlikely to win. So adding a couple runs to your teams ledger in that case is unlikely to matter. What happens later in the game isn’t really relevant to the situation. Personally, I like WPA though I am not sure it is a “skill” in the same sense that making contact is a “skill.”
An argument could be made that this should be the definition of “valuable” (in MVP): helps his team win the most. Personally I like thinking about MVP as a “Player of the Year” award. Who was the best player in baseball. There are lots of ways to arrive at this. I don’t like WPA because it values context out of proportion. It isn’t his fault that his hits come in “low/medium leverage” situations. As the article points out last year he was positive WPA by a similar amount.
Nice piece Jeff! Having watched a lot of early morning Cub games i feel Bryant has probably done more to help us win than clutch shows (much in the vein of what fredfotch mentions above) but the big problem seems to be him being too patient/not going for the big hit enough. His power has disappeared in high leverage situations and it’d be nice to see it more. Still i see him as the MVP
Also is there a negative correlation between clutch and wpa/LI the only reason i ask is the worst WPA/LI on the table for unclutch is 1.8 but for the corresponding top 10 clutch hitters only 1 matches that mark (jose ramirez number 10 in clutch with 1.9), in fact many are negative.
Do we have access to the WPA and LI splits somewhere? I can’t find it.
What do you mean by splits? Overall numbers are on the Win Probability tab. For plays for an individual player, go to his player page and check out the Play Log.
“In low-leverage situations, Bryant’s wRC+ ranks in the 99th percentile. In medium-leverage situations, it ranks in the 85th percentile. In high-leverage situations, it ranks in the eighth percentile.”
Where do I find those ‘splits’ for player and league? I’m sure I’m not using the proper terminology, sorry. 🙂
Fourth from the bottom: Double, -0.03
How can a double be a negative WPA event?
Perhaps he got thrown out trying to stretch it into a triple? Can’t see any other reason why it’d be a negative.
Or there was someone on base and they got thrown out at the plate.
If that were the case the -wpa should be charged to the runner. There’s some strange story behind this.
Runner thrown out, probably. I believe this gets charged against Bryant’s at bat whether or not that runner was Bryant – a particular flaw in WPA. Someone please correct me if I’m wrong. (I’d look it up if the chart was sufficiently detailed to tell me where to look.)
Found it-Seems odd to count this as a Bryant failure
5/8: Kris Bryant doubled to center (Fliner (Fly)). Jason Heyward out at home. Kris Bryant advanced to 2B.
This was pretty easy to look up and does illustrate how WPA works in some situations. Go to Bryant’s page, play log, sort by LI, find 2.89 and a double.
Kris Bryant doubled to center (Fliner (Fly)). Jason Heyward out at home. Kris Bryant advanced to 2B.
Here’s the play log for the game: http://www.fangraphs.com/plays.aspx?date=2016-05-08&team=Cubs&dh=0
Looking closer, that was a weird one — Bryant doubled but Jason Heyward was very easily thrown out at home. Had Heyward just stopped at third — the ball didn’t even go to the wall — then Bryant would end up with a WPA on the play of about +0.203. Bryant should get credit for that (not that it makes an enormous difference, of course).
http://www.fangraphs.com/wins.aspx?date=2016-05-08&team=Cubs&dh=0&season=2016
Bottom 11: Kris Bryant doubled to center (Fliner (Fly)). Jason Heyward out at home. Kris Bryant advanced to 2B.
Definitely not on Bryant…
When I used to have the energy to calculate WPA by hand, I had a separate category for baserunning WPA which would have put this on Heyward, not Bryant.
What was the WPA for the 3rd base coach?
Seriously, I hope the fielder’s WPA went up and not the pitcher’s in this instance. I thought most of the point of WPA was to assign value to the player involved. Penalizing Bryant for a runner being thrown out is a farce.
Absolutely. The problem is that people haven’t thought of having a separate baserunning wpa.
In response to the michael saunders article, a guy wrote saying he looks at a lot of cubs games and he’s noticed that bryant goes out of the strike zone a lot in high-leverage situations, that could be why his clutch score is so bad. Unfortunately the fangraphs data don’t make it possible to check plate discipline in different leverage situations (do they?). This would be a good thing to check somehow.
It’s on his player page in the split section. his K’s do go up from 19.8 in low to 23.5 in medium to 34.9 in high (that’s for this year over his career they remain steady). His walks stay the same in high leverage so maybe he’s being too passive? His power falls off a cliff though. ISO goes .318 to .256 to .054 that’s what baffles me
If Bryant is expanding his zone in high-leverage situations as johnforthegiants suggests, it would make sense that he is whiffing more and also making weaker contact, which explains both the K rate rise and the ISO drop. Looking at his batted ball numbers in high-leverage situations in 2016, he has an IFFB% of 27.5%(!!!) and a HR/FB of 0.0%(!!!), to go along with a line drive rate of just 8.7% and a hard-hit rate of 26.1%. Without actually looking at his high-leverage at bats, it seems like maybe he’s swinging hard at pitches upstairs to try to find something to elevate for a home run, and as a result is popping up a ton. And the fact that someone with 33 home runs and counting has yet to hit a SINGLE ONE in a high-leverage at bat is pretty remarkable.
Also of note is that last year Bryant not only had a positive clutch score as Jeff mentioned, he was downright excellent in high-leverage plate appearances, with a 127 wRC+, just a 25.8% K rate, and a 37.5% hard-hit rate. Which suggests that even if his approach this year in high-leverage situations has been worse, there’s no reason to expect that to continue. If for no other reason than you’d expect that someone on pace for 40+ bombs will eventually hit one in a high-leverage at bat just by chance.
But it could be that this year, pitchers noticed that he expands the strike zone in high-leverage situations and started to take advantage of it.
I played with a guy who was all Ted Williams in normal ABs, but when it got to clutch situations went totally Vladimir Guerrero.
Those…those are both good players.
I played with a guy who was all Babe Ruth with the bases empty but if we had a runner on 2nd he went totally Lou Gehrig. It was sad. LOL
I know about that, i was talking about whether for example his o-swing% is higher in high-leverage situations. I don’t see any way to check that.
I would have to assume that, over a large enough sample, Bryant is facing better quality pitchers in high leverage situations than he is in low leverage situations. That would explain some, but not all, of the difference in K rate.
But that would be true of everyone, no? I mean to a greater or lesser extent.
I think I wrote that. I feel he gets anxious and expands in RBI put-it-in-play situations (i.e. runners on 2nd and 3rd and one out). For the most part he is normally very good at laying off breaking balls low and away, especially early in counts, but in RBI situations I feel its the opposite. His approach feels like it changes and he gets himself behind.
I can’t help but think back to last year when he had 99 RBIs with two weeks remaining, but started pressing (in his own words) and ended up with that as his year-end total.
Jeff, I love this idea of going through the clutchness plays. However, I feel that a bit more context would be needed. How many of those plays were hard hit balls to a fielder? Were they weak ground balls? If he hit the ball hard in 20 of them, that changes it for me.
Awful lot of strikeouts in here, actually. 🙁
Then he’s a bum. Trade him to the Phils for the ghost of Jim Eisenreich.
Ryan Howard! You could platoon him!
And a TON of popouts: IFFB% of 27.5%!!!!
Edit: Which is not inflated by a lower number of fly balls, he is still hitting 48% flies in those situations, slightly higher than his overall 45% rate
Edit 2: If my numbers are right, in high-leverage PAs Bryant is putting the ball in play 53.5% of the time (100%-K%-BB%), and of those times he is popping out 7% of the time (BIP% x FB% x IFFB%). That means that Kris Bryant is striking out or popping up 42% of the time (!!!) in high leverage appearances. That speaks to a bad approach as much as bad luck, although his comparable numbers last season were excellent so clearly this is not some permanent thing.
Also just 8.7% line drives and 26.1% hard. Pretty bleak.
Why would that change your opinion? The whole point of using WPA data is to account for what the player DID – it’s very results-based – as opposed to what a player is capable of doing (or, if you prefer, “should” have done in a completely just world). I’m a pretty big believer that MVP votes should skew more toward results than talent (MVP awards are retrospective, whereas things like hard-hit-ball data are prospective). I still think Bryant may be the NL MVP this year, but I also think Jeff’s article adds a meaningful wrinkle to the discussion.
It would address a different question, which is whether the person can generally hit well in important situations, for example because he tends to chase bad pitches in these situations–and this question is taken seriously in a number of replies here which suggest that Bryant for one reason or another does chase bad pitches in these situations (Brent Henry, Imachainsaw, BritishCub). His bad Clutch score might be a result of bad luck or something systematic. If it’s the result of bad luck–that is, he hits the ball just as well in high-leverage situations but he happens to hit it at fielders– an argument can be made that nevertheless he should lose MVP points even though this does NOT suggest that he has less talent. On the other hand, if he really does hit the ball much worse in high-leverage situations (and the high K%, high IFFB%, low LD%, and low Hard% suggest that this is the case), then we need to consider the possibility that there is something systematic about this, so that his Clutch score would indicate that he is not just less valuable but also–at least for the relevant time period–also less talented.
I would guess that the higher the leverage of the situation, the more likely the opposing manager is to bring in a right-handed reliever (and probably, a good one) to face Bryant.
I would also guess that this has been more likely to happen in Bryant’s 2nd year than in his 1st.
If so, then perhaps some of what appears to be “unclutchitude” might actually be “facing the other team’s best relief pitchers.”
But those are just guesses…
I would think that’s much more likely to happen with a lefty hitter, since most teams have a LOOGY for those situations. When it comes to right handed pitchers, the team’s best two pitchers are usually in set roles – the best pitcher in the 9th, and the second best in the 8th. Maybe they’d bring in their setup man with 2 outs in the 7th to face Bryant, or they’d bring in the closer with 2 outs in the 8th for a 4-out save, but aside from that I doubt it would have much of an impact.
But having a dominant pitcher in the 9th inning of a close game should have a disproportionate impact on a hitter’s performance, regardless of any platoon split. For example, it’s no great shame grounding out against Familia, that’s what many hitters great and small do. But the differential impact for great hitters skews the clutch stat. Here’s how the game-ending GIDP against Familia would affect different batters.
Kris Bryant: MVP-level offense vs. GIDP worth -0.53 WPA
Addison Russell – league-average offense vs. GIDP worth -0.53 WPA
Willie Bloomquist: replacement-level offense vs. walking back to the bench for a pinch-hitter (worth .0.00 WPA)
You know, I’m still pretty upset about that bases loaded, 1 out, DP ball against Familia, too. Thanks for bringing it up, WPA.
As others have pointed out, there seem to be flaws in using this, but I guess I could see it as a tiebreaker perhaps.
There are other soft factors that can negate some of this too that can still help the team win. Let’s say the Cubs are up 3-0 and Bryant hits a 3 run homer in the 5th. Probably fairly low leverage index, but as a result, Arietta comes out of the game an inning early and the back of the bullpen finishes out the game instead of having to use Chapman. Just one example, but I’m sure there are other ways that performing in low leverage spots still helps a team.
my take on this is that in high leverage situations, pitchers are showing less willingness to challenge him, and in an attempt to make something happen he chases bad pitches, at least that’s the impression i’ve gathered from watching him this season. the fact that in 68 ‘late & close’ plate appearances he has a slash of .237/.338/.322 seems to be a major culprit, along with a nearly 28% K-rate in those situations. though to be fair, his OPS+ is that situation is 94, so not horrible in relation to the rest of the league. and by looking at his play log for this year, it looks like he took a major hit on July 9th when he grounded into a double play with the bases loaded and 1 out in the bottom of the ninth, but it was also against Jeurys Familia. but also considering that he’s been batting 2nd lately and is leading the majors in runs scored, his MVP value shouldn’t be solely incumbent on his rbi totals or clutchness.
Yes, this is the sort of thing that i was suggesting above
Personally, I prefer REW over WPA. REW does a better job of measuring how you contributed to your team’s run total if you are concerned about how a player performs in context. And obviously runs scored in the 2nd inning count just as much as runs scored in the 9th.
There’s a major difference between Kris Bryant this year and Bryce Harper last year. Harper lapped the field, and was very likely the most valuable player regardless of clutch performance. Bryant might be the best player, but it’s much closer this year, and his terrible clutch performance probably means he has not been the most valuable.
WPA doesn’t show “actual offensive value”!
It treats a suicide squeeze and a homer with the bases loaded in the bottom of the ninth and the game tied as being of equal value.
RE24 does appropriately weight run-outcomes relative to the 24 base-out states. Use it.
If you do look at RE24, you’ll find that Daniel Murphy’s been 5 runs more valuable offensively than Bryant on the season. From there, the question is whether you believe Bryant’s been 5 runs better in the field and on the basepath than Murphy.
I think he has.
Obviously, every game counts. But Bryant has a .415/.508/1.000 slash like with 22 hits, 9 home runs and 21 RBI in 13 games against the Reds this year. Those things happened; I’m not discounting them. But it probably speaks to the overall clutchiness of the guy that he’s padded his stats against a horrible pitching staff in games he could have been replaced with one of the Bad News Bears (non-Kelly Leak division) and the Cubs still would have won. Maybe that’s how most MVPs do it – I don’t know. But if you just look at his ‘team VS’ splits, he’s pummeled terrible teams all year.
All that said – I’d still give the guy the award. You can’t help who you play, and this is all probably noise. But it’s noise worth noting, if nothing else. Maybe, though, the Reds organization could get some sort of thank you note or a fruit basket when all this is said and done.
Just for fun, vs teams with more than 10 games played, Daniel Murphy has an OPS over 1.000 against the Mets (1.331), Braves (1.250) and Phillies (1.077). For Bryant, it’s Reds (1.508), Brewers (1.167) and Pirates (1.124).
I posted these Kris Bryant stats in a couple of replies so I apologize for cluttering up the comment section, but they are so staggering that I think they deserve their own comment (also I’ve spent the last 40 minutes researching this, sue me).
For context, here are Bryant’s overall numbers for 2016:
K%: 22.3%, IFFB%: 13.6%, K%+PU%: 26.4%, HR/FB: 20.4%, LD%: 23.5%, GB%: 31.1% Hard-Soft%: 26.1%, wRC+ 155
And now high-leverage plate appearances:
K%: 34.9%, IFFB%: 27.3%, K%+PU%: 42.0%, HR/FB: 0.0%, LD%: 8.7%, GB%: 43.5% Hard-Soft%: -4.3%, wRC+ 43
That is a truly remarkable change not just in the surface results but in the underlying numbers. In high-leverage situations, Kris Bryant is whiffing, popping up, and rolling over grounders. Small sample size aside, this is clearly more than just bad luck. According to johnforthegiant there was a comment on Sullivan’s Saunders article suggesting that Bryant expands the zone in high leverage situations. The plate discipline and batted balls stats certainly suggest that hypothesis has some validity. My thus is swinging hard at elevated pitches, which is leading to pop ups and weak choppers. If I have time (which clearly I do) I’ll look at some of his specific plate appearances to see if I’m right. (Dibs on the community post!)
And for what it’s worth, Bryant’s performance in high-leverage situations last year was excellent: wRC+ 126 supported by the batted ball and plate discipline numbers, which lends hope that this is a correctable problem and not likely to be an ongoing issue, as Bryce Harper’s bounce-back in clutch this year attests (his overall performance notwithstanding).
It could also be that opposing pitchers learned this year that Bryant will go fishing in high-leverage situation and they are pitching him accordingly, whereas last year they didn’t realize that.
It could also be that we are just building a (false) narrative based on a random set of numbers… because “clutch” is NOT a thing. Its random.
…and yet Michael Brantley and Ryan Howard exist.
“You could make a real argument that Russell has been more valuable to the Cubs.”
Eat your liver out, Billy Beane.
I understand why Jeff used Addison Russell as the foil to Bryant – his MVP candidacy is more counterintuitive, and more instructive to the WPA vs. ‘raw statline’ debate. But I’m surprised he didn’t mention that another of Bryant’s teammates – Anthony Rizzo – leads the NL in WPA. Considering Rizzo’s slash #’s are very similar to Bryant’s, and he also fields well, AND has an unusual number of huge hits, he seems to have a strong case for MVP, maybe stronger than Bryant’s. Attractive problems to have up there in Chicago.
I certainly buy that Bryant has been worse in high-leverage situations that he has been on the whole. By the numbers that obviously isn’t arguable. After glancing at his game log I think one thing that helps his case is that his performance has still been the difference in a lot of games won by the Cubs it’s just that his contribution wasn’t usually early in games. In the end, if you hit a solo home run and your team wins the game 1-0, is it a more important home run if it is hit in the 9th instead of the 1st? I’d say not really, but WPA can’t tell the future so the homer in the 1st doesn’t add nearly as winning probability.
Even playing the “what if” game it looks like the Cubs would have lost more games without Bryant’s impact on them. My theory is that for whatever reason (probably coincidence) Bryant just hasn’t hit as well late in games this year, which to me is way less damning for an MVP candidate than only producing when a game is out of hand one way or another.
Inning/OPS
1st/.810
2nd/1.433
3rd/1.026
4th/1.070
5th/1.359
6th/.943
7th/.799
8th/.726
9th/1.020
10th+/.748
Are you claiming that a walkoff solo homerun in the bottom of the ninth isn’t more important than a solo homerun with the score tied in the bottom of the first? In the first case, your team wins the game, end of story. In the second case, you’ve helped but who knows if it will have any relevance to who wins the game?
I think you are missing an important qualifier. I know that it has relevance to the game because in both cases my team won 1-0. That’s why I specified.
So if the value of a particular plate appearance is determined after the game by seeing how much it contributed to a team winning or losing a game, as you’re suggesting here, regardless of context, that means that a triple hit in an inning in an inning in which one’s team scored no runs has the same value as striking out, because they both resulted in no runs being scored. Do you actually believe that?
Are you describing why RE24 is superior to both wRAA and WPA with regards to actual value provided, to the point where WPA should never, ever be used?
Because you’re doing a great job!
No, that isn’t what I’m doing at all. Why would you think that?
wRAA, RE24, and WPA are all important stats for MVP consideration, they each have their usages and they each have their shortcomings. I wouldn’t put one ahead of another. But if you’re going to focus on one, RE24 is the one which doesn’t make sense. If you think that context shouldn’t matter, then wRAA is the obvious choice, if you think that context is primary, then WPA is the obvious choice, but there’s no way of thinking according to which RE24 is the obvious choice. That doesn’t mean that I think that RE24 is less important, but there’s no way it’s the most important.
The way it’s the most important is if you think the best way to evaluate a player’s value is based on how he’s actually impacted the run expectancy of his team.
(You know, the premise of WAR).
If you think that a run in the first is as valuable as a run in the ninth (because of that pesky “reflexive property”), but that context does matter, then RE24 is the only option.
Which is why RE24 actually measures value (to the point that i replace wRAA with RE24 to calculate a player’s WAR for the purposes of the MVP), and WPA really, really doesn’t.
Sequencing within an inning matters, and can be influenced by a player’s set of skills. Sequencing of run scoring across innings is much more heavily influenced by other items, particularly managerial decisions such as pitching changes.
RE24 captures all the situational skills related to sequencing within an inning (including stuff like productive outs). WPA adds a layer of sequencing across innings, which may or may not be important (to Dane Robbens’s point)
Dave Studeman wrote a great article about your point in THT last year:
http://www.hardballtimes.com/whats-a-most-valuable-player/
The argument for RE24 over WPA really boils down to whether or not you believe in the reflexive property.
Bringing last year’s Harper into the discussion is not really fair to Harper.
Last year Harper’s WPA was 6.16, in 2nd place
behind Rizzo’s 7.15 mark.
their wrc+’s were 198 and 145 respectively.
It might be true that the gap in their value as hitters was smaller than
the difference in wrc+ makes it out to be,
but only a handful of people would sincerely argue that Rizzo was a more valuable
batter, let alone a more valuable player.
Currently, Kris Bryant’s WPA is 1.91 in 28th place (among NL batters)
behind Rizzo’s 4.37 mark and somewhat more importantly,
behind Murphy’s 4.11 mark.
their wrc+’s are 155, 151, and 159 respectively.
So in this case, people argue that the gap in their value as hitters
is larger than the difference in wrc+.
Just for the record, if the season were to end now, I would support Kris Bryant
as the MVP. However, that Harper comparison doesn’t work at all.
Why is it more important that he is behind Murphy than he is behind Rizzo? Personally I don’t like focusing on the stat because it feeds into the whole clutch thing and if anything gives it credibility. Obviously someone who just hits in blow outs isn’t very valuable, but a solo home run in the 1st in a tie game and one in the 9th of the same game count the same but wpa treats differently.
He hit a solo shot in the 8th tonight in a 1.35 LI situation, and a 2 run HR in the 10th in a 3.48 LI situation (his 11th highest leverage situation this season per your table). He’s at +0.44 WPA on the night.
Bryant clearly read this article, took some offense to the conclusion, and decided to WPA the hell out of a ball in the 10th inning.
http://m.mlb.com/video/v1101797483/chclad-bryant-hits-no-35-puts-cubs-on-top-in-10th
Apparently you pissed him off tonight and he took it out on the Dodgers, 0.44 WPA in this game alone.
For an article about timing, yours couldn’t have been much better.
You are overthinking this.
we need a new table so Kris can be 3rd worst on it and you can put fellow MVP candidate Seager on there at 10th worst and note somewhere that murphy has negative clutch too. I’d vote Rizzo only guy in top 20 of wpa/LI with a clutch better than 0.01, 18 others negative and yelich has 0.01!
That’s true and it’s a good point about Rizzo. However I just did an xBABIP calculation for Rizzo in high-leverage situations and it turns out that he’s been extremely lucky there–his BABIP has been .457 but his xBABIP has been only .285.
So after last night bryant’s +wrc in high leverage situations is up to 72, same as delino deshields in all situations, and his clutch rating has gone up from the 2nd worst in the majors to the 4th worst. Not sure if this is mvp caliber if the idea is to help your team win games.
What last night tells us is that the season isn’t over. We’ll see where his clutch rating stands at the end of the season and how close in production his competition is for the N.L. MVP.
Yes, and as britishcub points out, most of his serious rivals also have clutch issues (other than rizzo) so this should be considered too.
Just to note as well jsut focusing on bryant for his failing is removing context from an mvp debate. Seager’s high leverage WRC is 61 which is worse than bryant and murphy’s is 87 none of the three main candidates are good in high leverage.
Yes, that’s an excellent point. it happened with Bryce Harper too last year. I wonder if it’s a coincidence, that so many batters who rate very high by context-free statistics are pretty weak in terms of Clutch.
Interesting article. I do find it funny though that he hit two clutch homers last night after the article was written. Although the analysis of the article still holds, I would say the timing of this article wasn’t very clutch…
The second hr got a much higher wpa than the first, because the first still left them behind by one run.
Cool factoid of the day….note the crazy similarity between Bryant’s second-half numbers and Mookie Betts’s second-half numbers. Those two could be twins 😉
http://www.fangraphs.com/leaders.aspx?pos=all&stats=bat&lg=all&qual=y&type=8&season=2016&month=31&season1=2016&ind=0&team=&rost=&age=0&filter=&players=
There are two separate questions here:
(1) Should the MVP be about ability or performance? If it is about performance, then it seems undeniable that WPA and Clutch should be relevant (though not necessarily central). If it is about ability, the situation is less clear, because of the second question.
(2) Although there it an assumption by some people that Clutch (and similar statistics such as those distinguishing between high-leverage, mid-leverage, and low-leverage situations) is a matter of luck rather than ability, this is not necessarily the case for all batters at all times. A number of people who have written here–BritishCub, Brent Henry, and imachainsaw–have all reported that in their observation, Bryant systematically changes his approach in clutch situations:
‘Having watched a lot of early morning Cub games i feel Bryant has probably done more to help us win than clutch shows (much in the vein of what fredfotch mentions above) but the big problem seems to be him being too patient/not going for the big hit enough. His power has disappeared in high leverage situations and it’d be nice to see it more.’
‘I feel he gets anxious and expands in RBI put-it-in-play situations (i.e. runners on 2nd and 3rd and one out). For the most part he is normally very good at laying off breaking balls low and away, especially early in counts, but in RBI situations I feel its the opposite. His approach feels like it changes and he gets himself behind.’
‘my take on this is that in high leverage situations, pitchers are showing less willingness to challenge him, and in an attempt to make something happen he chases bad pitches, at least that’s the impression i’ve gathered from watching him this season.’
Brent Henry also writes that:
‘I can’t help but think back to last year when he had 99 RBIs with two weeks remaining, but started pressing (in his own words) and ended up with that as his year-end total.’
These comments from people who watch Bryant hit a lot suggest that Bryant’s weak statistics in high-leverage situations are not an accident. Although this is heresy to those who believe that statistics like Clutch are always and invariably a matter of luck, we have no proof that this is the case and scientific procedure suggests that this possibility should be empirically investigated.
So that my position is clear, I should emphasize that in such cases, the burden of proof should always be on those who claim that Clutch data in a particular case (for a particular batter over a particular period of time) are not a matter of luck. And furthermore, each claim of a meaningful statistical difference in different leverage situations must be judged on its own merits–the fact that one case of an apparent difference turning out to be actually insignificant does not imply anything about another case.
What would constitute proof on a meaningful change in hitting behavior in different leverage situations? I can think of two types of proof:
(1) If there are enough raw data over an extended period of time, and the differences are great enough, that chance can be ruled out. Offhand, I would suggest Ryan Howard and Joey Votto as examples of this in terms of positive clutch hitting (Eric Hosmer has been around much less time but his leverage statistics during that time are so striking that it is unlikely to be due to chance). Clearly, however, this is not the case with Bryant, because there are not enough data.
(2) It may also be that other statistics suggest that there is a difference in approach. In this case, the suggestion made by some people here that Bryant changes his behavior in high-leverage situations could be tested by comparing for example his O-Swing% in different leverage situations–if it is indeed much higher in high-leverage situations (which is suggested in some of the comments), then this would explain why his Clutch data are as they are, this would be a matter of ability, not luck. Of course this may only be true for a limited period of time, because (for example) he may only press in this way for a limited period of time, but this does not mean that this should be included in a realistic evaluation of this ability for this period of time. If the period of time where this situation holds is more limited, this will naturally mean that it will be more difficult to demonstrate statistical significance, but if statistical significance can be demonstrated over a certain period of time, then it follows that this will demonstrate something about his ability for this period of time.
But again, the burden of proof in each case should be on those who claim that leverage statistics are not random.
You’re like someone that recently figured out you can prove 2+2=/=4, and all you want to do is show off your new trick. Math is great because it can be very enlightening when used correctly.
Sullivan did his best to make an argument against KB for MVP. There is nothing wrong with that. In fact, that is often a very constructive exercise when reviewing the merits of a particular premise, such as KB is the current frontrunner for NL MVP.
But don’t let that fool you into believing he isn’t the MVP right now. If you want to count RBI’s as determinative and prove your point with O-Swing% in high leverage, cool. The rest of us know that 2+2=4 and aren’t fooled by your fancy proofs.
Um, you do realize that you haven’t even tried to make a counterargument, don’t you?
Yes, that was intentional. Meritless arguments do not require counterarguments. They are without merit, and sufficiently defeat themselves.
In other words, you have no counterarguments.
Another factor (which I just thought of) which might be relevant in the case of Bryant is that he’s clearly the best fastball hitter in the majors by statistics and it may be that in high-leverage situations, pitchers are simply refusing to throw him fastballs even if they can’t throw their breaking balls for strikes, and Bryant is pressing and swinging at breaking balls which are out of the strike zone.
You don’t think any of this can be chalked up as small sample sizes? He was the 5th most clutch player in the mlb last year? Now in his 2nd year he forgot how to be clutch?
Yes, it is quite possibly due to small sample size, but it is significant that not only are his basic statistics like +wrc very weak in high leverage situations but also ld%, k%, iffb%, and hard%. But still, it certainly could be small sample size. This possibility would be further decreased, though, if it turned out that he sees a lot fewer fastballs in high-l situations and has a higher o-swing%, as that would mean that the hitting environments are fundamentally different.
It isn’t that he forgot to be clutch in his second year. It’s that pitchers decided to avoid throwing him fastballs in high-leverage situations in his second year.
Haha…this article comes out yesterday and I got to sit in the Cubs-heavy crowd at Dodger Stadium that very night screaming “MVP MVP MVP” as Bryant hits a 10th inning two run game-winner. Poetic.
Well i gotta thank him for helping beat the dodgers yesterday, that’s for sure
I really hope they can trade him for cameron rupp after the season…
How do those numbers work out now?
I wrote above. After last night he has a +wrc of 72 and the 4th worst clutch in the majors
It turns out that as bad as Bryant’s 72 +WRC in high-leverage situations has been, he’s actually been extremely lucky in this situations. His xBABIP in high-leverage situations (calculated according to the formula in http://www.fangraphs.com/fantasy/hitter-xbabip-v20-a-long-needed-update/) is .207, but his actual BABIP in these situations is .348. With normal BABIP luck, instead of hitting .237 in high-leverage situations, he would be hitting .159.
For a self proclaimed Giants fan, you sure spend a lot of time digging into Cubs statistics.
It sort of yells “closet Cubs fan”. It’s ok, you don’t have to be embarrassed. The Cubs are the best team in baseball by a lot and they boast a couple frontrunners for MVP. You obviously already know that.
Go Cubs!
As I mentioned before, I like to dig into WPA and Clutch statistics whether they’re about Cubs or anyone else, and I’ve made posts about this in recent articles about Michael Saunders, Bryce Harper, and Joey Votto. I just think this is a very interesting topic whichever player or team it’s about. I have to say though that those posts (and those articles) don’t attract as much attention as articles about the Cubs so there isn’t so much discussion there. I spend at least 10 times as much time on Giants’ statistics as any other team’s but I post comments about this in Giants’ forums.
I’m also very interested in xBABIP and make a lot of posts about this topic. Some of these relate to the Cubs because of the extremely low BABIP of their pitching staff this year but others relate to for example Xander Bogaerts, Bryce Harper, and Francisco Lindor.
Other than the Giants, I’m interested in topics, not teams.
I need your help, all. I am in a 10 team points only keeper league, (5 SP, 2 RP, C, 1st, 2nd, 3rd, SS, LF, CF, RF, 2 UTIL)
You get to keep 6 players. Who would you keep?
I have Betts, Bryant, and Correa who are all virtual locks, but then I have Gary Sanchez, Aledmys Diaz, Gerrit Cole, Johnny Cueto, Aaron Sanchez, Carlos Martinez, Julio Urias, and Dylan Bundy.
I love Cole, but he has had a tough year and I am concerned- adding on that the Pirates aren’t necessarily a team in which he will garner a lot of wins. Cueto has been great, but will we see this same pitcher next year? It’s hard to argue with Aaron Sanchez’ production, but he’s still somewhat unproven and untested in terms of a full season’s worth of innings. There is plenty to like about Carlos Martinez, but he still hasn’t eclipsed that hump that divides between really really good pitcher and being elite. Then, you have Urias who could be the next phenom as he’s already shown flashes of dominance. I am big on Bundy, as well, but he is likely to be handicapped by innings and still has quite a lot to prove despite his recent showings of success.
My instincts would be to keep three pitchers, but given how difficult it is to find a stud catcher, I am intrigued by keeping Gary Sanchez. As hard as it is to find good pitching, it’s that much harder to find good, consistent hitters, much less catchers. Obviously, he won’t continue at this pace, but seems like a legitimate top tier catcher who will also benefit from some time at DH. I am equally enticed by Aledmys Diaz. Although I already have Correa, Diaz has shown a lot of promise and seemed to have been improving as the season went on. I actually had Corey Seager this time last year and thought I could scoop him up mid-draft, but someone else got him so I suppose I’m a little wary of the same situation repeating itself.
The biggest argument against Bryant is simply the other great players on the team. As you stated, the Cubs would probably still be doing fine without him. The MVP award should not be looked at based on only an individuals stats. Other players on the team matter to determine who the true MVP is.
I don’t know if this matters towards WPA or even this article, but because of the amount of wins the Cubs have where clutch situations are less likely to occur does that factor in? Or do they have a similar amount of clutch situations (high leverage) and therefore that point is mute? In other words does team quality/ways teams win affect the amount of high leverage situations or are random clutch moments fairly similar between all teams?