By Noodle and Son
Last time we discussed the concept of overrated; this time we move on to another often debated topic, ‘clutch’ and ‘unclutch’. Again, there is no real definition, so we will define our own, and invite readers to weigh in.
We will start by defining two terms, familiar to those in the sabermetric community: WAR and WPA.
WAR: Wins Above Replacement. The basic concept is how many wins this player adds to your team above a player you can easily find in the minor leagues. A full definition is found on http://www.insidethebook.com/ee/index.php/site/article/how_to_calculate_war/.
Obviously, the best players will have the highest WAR, and vice-versa for the worst players. This does not take clutch performance into account, but rather is based on total statistics.
WPA: Win Probability Added. Every at bat during a game begins with the team at the plate having a certain probability of achieving victory, and ends with a different one. For example, a game begins with a visiting team having about a 46% chance of winning. (Home teams win 54% of the time.) If the leadoff hitter hits a home run, the visiting team leads 1-0, and now has a 56% chance of winning. The leadoff batter gets 0.1 points of WPA – since he added 10%, or 1/10th of a win, to his team’s chance of winning. Later in the game, say the 9th inning, his team is down by 1 with 1 out and the based are loaded when he hits into a double play. At the start of the at bat his team had a 52% chance of winning (he can get a hit, and if he doesn’t the next hitter might) and at the end of this at bat his team has a 0% chance, since the game is over. In that situation he has -0.52 points for this at bat. At the end of the season, we add the WPA of each at bat of the entire season to get his seasonal WPA. Of course one expects the best players to have the highest WPA, whether they are clutch or not. (All statistics courtesy of http://winexp.walkoffbalk.com/expectancy/search. You can enter any situation at any time during a game and calculate the probability of winning.)
For example, the top WPA in baseball was Albert Pujols at 8.24. He was also second in the major leagues in WAR, behind Ben Zobrist. (A good hitter at second base is worth more than a good hitter at first base.) At 50 points for a single win, Albert’s WPA added 16 wins to his team. The other top WAR finishers were unsurprisingly Jeter, Utley and Mauer.
We will define unclutch as someone with a high WAR – meaning a very good player based on his statistics – but a low WPA, meaning he did not come through to help his team actually win games. Hitting many Home Runs in games your team is ahead or behind by numerous runs will increase your WAR but not your WPA.
Thanks to the good folks at the Hardball Times (http://www.hardballtimes.com) we were able to download and sort all the data. We took the top 100 players in baseball as sorted by WAR. We then compared their WAR to their WPA. It should be no surprise that there is some correlation between WAR and WPA: for this group of 100 players the correlation is 0.54.
On the entire list of top 100 players (by WAR) there were only 12 with negative WPAs. The player with the worst WPA on this list is Robinson Cano at -1.51. His inability to hit with runners in scoring position (see a recent article here: http://www.theyankeeu.com/2010/02/why-does-cano-fail-in-the-clutch-14717 ) cost his team 3 wins!
To put this another way, we looked at all 671 hitters listed in the Hardball Times data base. While Cano’s WAR (4.27) would rank him as the 36th best player in baseball, his WPA puts him at #641! The players below him cost their team more wins, but that list is made up of bad players – players with low WAR. His WPA is between Jose Molina and Yuniesky Betancourt!
Does this prove he is the most unclutch player in baseball? It does to us!
For those interested in clutch hitting – yes, Pujols had the highest WPA, but this does not mean he is clutch, it means he is great. After all, his WAR was huge too. A clutch player is one who rises above his ‘normal’ talent level. For 2009 we give the award to Joey Votto – whose WAR (4.43) was 33rd in baseball, yet had the 3rd highest WPA in baseball.
February 17, 2010
The most clutch and unclutch player in baseball, 2009
11 Comments »
RSS feed for comments on this post. TrackBack URL
i really get what your saying but what about all the other players.for example jorge posada cost the yankees 5 games if not more and cano actually made them win a couple, but everything else in the article could not have been more true
Comment by mathew — February 17, 2010 @ 7:43 pm
I completely agree that cano was the worst clutch hitter in baseball last year. It seemed like every time he led off an inning he got a hit and every time he came up with the bases loaded he got out.The yankees should have have traded him. I hope he was just unlucky and he will do better this year but I wouldn’t bet on it.
Comment by overrated cano — February 17, 2010 @ 8:42 pm
sorry that i am not that up to date on many of the modern players. i would love to find out about Joe Carter, Tony Perez etc. , players with very high RBI totals but are bashed by all the math lunatics who follow baseball.
Comment by mosho — February 18, 2010 @ 12:41 pm
Joe Carter was a lifetime 259 hitter with a 771 OPS.
In ‘high leverage’ sitations he batted 264 with a 763 OPS.
Runners in scoring position he was 271/805, though with With two outs and Runners in scoring position he batted 258 with a 787 OPS.
Simply put, there is nothing in his stats to show any kind of ‘clutchiness’. Yes, it seems he was a little better with RISP, but 271/805 is nothing for a corner OFielder (or 1b) to wriote home about. Rather, he knocked in 100+ runs a year becasue he played on team that had a lot of baserunners.
Tony Perez is a little more interesting. 279 and 804 for his career, but 305/859 in high leverage. That is quite a difference. Overall with RISP he is 284/833, slightly better than his career average. The bottom line is that hitting behind Morgan, Rose, etc leads to way more RBI chances that most other players ever get.
Comment by Noodle — February 18, 2010 @ 6:19 pm
what on earth are you talking about. statistics prove nothing.if the yankees would’nt have him okay he is the best clutch hitter but he provides a spark to the team.unlike some other players that can’t do any like posada plus he alsi might of saved them a few games withe his defense turning double plays in a situation where other players could not. so what are you saying
Comment by confused — February 19, 2010 @ 12:42 pm
hey noodle
i think it’s time to rethink our attitude to rbi men. after all, while you claim that perez had rose and morgan hitting in front opf him, carter played for the indians and padres before the jays and drove in 100 runs a year with brook jacoby and the like. on the other hand, throughout history, there have been many players who get on base a lot (let’s use 220 times a year, arbitrary figure) and don’t score too often (under 100, also arbitrary figure) as in the curious case of vince coleman in 1988 when the speedster had 160 hits and 49 walks yet scored…77 runs, batting leadoff! richie ashburn had a number of similar seasons where he would be on base 250 times, be a speedster, play for contending teams and not score 100 runs. Now, bobby richardson, a favorite of yours doesn’t count. whether it was scoring 45 runs in 1960 or only 80 runs in ‘61 batting ahead of the m&m boys, that’s totally beyond comprehension. that’s like patrick ewing having 1 rebound in a game. if you’re 7′1″, the ball should hit you on the heas 10 times! if you’re batting in front of 115 hr’s, you should be on base to enjoy the blasts! but, i digress. the point is, it seems like you need people to drive you hoime even if you are pete rose, richie ashburn, or rod carew who interestingly scored over 100 runs in a season once in his career. who knows? maybe the pitchers used to know how to bear down on hitters when there were men on base, maybe outfielders played more shallow and a single didn’t automatically score you from second. just i think we need to credit the rbi men more.
Comment by mosho — February 21, 2010 @ 10:26 am
When a guy hits .250 or .270 with and without men on base, he is not a very good hitter. Aaron Gleeman once wrote a brilliant analysis on why Tony Batista was a terrible player despite his 100 RBI seasons. To put it in simple terms, players who rarely walk will get more RBIs than better players who do. While the better players are walking and helping along big innings, the outmakers are hurting their team and every now and then driving in runs. Nevertheless, they are much less valuable to their teams. Joe Carter is the epitome of such a player. His lifetime .306 OBP puts him on par with Mark Belanger, for crying out loud! Joe Carter was an out maker. In the course of making outs and hurting his team, he knocked in runs. Woopie. his team woudl be better served by having him take a walk and letting the guys behind him knock the, (and him) in.
In summation: No, I will not give credit to “RBI men” unless they know how to hit and not make outs. Though I do agree that we have to figure out how Rod Carew reached base 290+ times in 1974 and scored 86 runs. It is a combination of now power to ever drive himself in (3 HRs and 38 total extra base hits) and likely because the guys behind him stunk worse than Joe Carter. Stealing 38 bases in 54 attempts is OK, but nothing to write home about.
Comment by Noodle — February 22, 2010 @ 12:40 pm
The is a formula on page 283 in the book Behind-the-Scenes Baseball which can be used to predict how many RBIs a hitter should have in a season. The formula is: multiply a hitter’s at bats times hit slugging percentage then also by his team’s on-base percentage. The problem is the formula is only designed to provide a rough estimate. For some seasons it works great. Example, 1991 Joe Carter had 638 abs, a slugging percentage of .503 and his Blue Jays had an on-base percentage of .336, multiplied out that is 108 which was Carter’s RBI total for the season. More research is needed on this formula before it can be used for studies.
Comment by Doug Decatur — February 22, 2010 @ 3:06 pm
the main problem is that the formula doesnt work for leadoff hitters or players who bat at the bottom of the order because team obp is not representative of the hitters hitting in front of them. For hitters 3-7 its a great formula.
btw, i ran it for cano and he should have had 120 rbis in 09. Instead he had only 85. Is this due to his lack of clutch hitting ability??
other conventional “clutch” stats show the same thing about cano:2009 stats(career rate in parentheses)
no runners on:.376 (.331) risp:.207(.256)
bases loaded:.259(.250) 2 outs:.263(.285)
2 outs,risp: .204 (.243) high leverage(over 1.5):.224(.253)
playoffs: .181(.217)
Comment by son of Noodle — February 22, 2010 @ 10:49 pm
we also have to think about the rbi men, if there was no one else who could do anything for their team, ex: when playing for the padres of the early 70’s, why should he take a walk. it’s a worthless ploy. true, it improves their math stats, but, when Frank howard in 1970 was on base 292 times, hit 44 homers so he’s “on base” 248 times, scored 90 runs, subtract his 44 hr’s we’re left with him scoring 46 runs in 248 on base situations, a % of 18.5%, what’s the point of him taking a walk? He will almost never score!! swing away!! joe carter played with no one behind him! example: 1987, Carter 182 times on base subtract 32 hr’s =150 opportunities, scored 83 runs-32 hr’s total 51/150 close to 33%. so, he’s much better at scoring than frank howard but willie mccovey in 1970, 59/241= 24.5%, maybe he should have not walked 137 times. ted williams for his career was at 30.7% so his walks were important even though he didn’t steal bases. maybe we should just go straight to nate colbert in 1972 should be the model for this discussion. in one of the greatest individual years of all time for a horrible team,he was on-base 211-38 hr’s=173 scoring 87-38 hr= 49/173=28% so even his walks meant something
Comment by mosho — February 23, 2010 @ 3:09 pm
When determining how much value walks have you must also take into account the runners that he is bumping up to another base increasing the chance of the other runner(s) scoring and even setting up as noodle mentioned bigger innings. Even if the guy behind him is not good, probability wise the chances of scoring the other runner(s) will increase more than his chances of driving home the runs .
Comment by Fantastic Mr. Fox — March 8, 2010 @ 7:11 pm