Murray is certainly one player with such a reputation. Many have said that he was a great clutch hitter. Did he really excel when the game was on the line? If so, was he better than normal or than expected? Did his clutch performance have a significant impact on his team's record? Did it ever impact pennant races or the playoffs? These are some of the questions I looked at.
Pitcher Mike Flanagan, one of Murray's teammates with the Orioles, said Murray was the best clutch hitter he ever saw (see sources for links to those citing Murray's clutch prowess). His clutch hitting was mentioned when he was elected to the Hall of Fame. To analyze this, I looked a variety of evidence.
Let's start with some conventional clutch stats. The table below shows Murray's stats for various situations during his career. The Close and Late and Non-Close and Late stats are only for the years 1984-97.
(OBP = on-base percentage, SLG = slugging percentage, and OPS = OBP + SLG)
Murray batted .300 with Runners on Base (ROB) while hitting .275 with None on (NONE), for a gain of .025. The normal gain was .011 over the years 1991-2000, so he was .014 higher than that (see General Clutch Data). Since he had 5,409 ROB ABs, the .014 means 75.73 more hits than expected or over and above what we would expect him to do with ROB. In ROB situations, Murray had 1,670 RBIs with ROB. But 128 of them were sacrifice flies (SFs). That leaves 1,542 RBIs on hits with ROB. He had about .95 RBIs per hit with ROB. Since he had an extra 75.73 hits with ROB, that means, approximately, means 71.9 more RBIs (I say approximately since I am not taking the type of hits into account-it's probably a little higher since his SLG went up so much with ROB). Murray played the equivalent of about 18.28 full seasons, assuming 700 plate appearances per season. That is an extra 3.93 RBIs per season. Since it usually takes about 10 additional runs to add one more win, Murray added less than half a win per season with his performance with ROB (since ROB and Runners in Scoring Position (RISP) situations overlap somewhat, I won't specifically discuss his RISP performance-but you can got to the General Clutch Data to see how his relative improvement with RISP compares to what players normally do).
Murray's SLG goes up by .053 in ROB situations over NONE situations. The normal gain is about .011, so he is .042 above that. That means an additional 227 total bases (TB). In ROB situations he had about .566 RBIs per TB and his gain of .042 in SLG means and additional 227 TBs. That means an additional 128 RBIs since .566*227 = 128. Over 18.28 seasons, that is about 7 more RBIs per season, much more than based on AVG alone, but its still adds less than 1 win per season.
You might wonder about taking the SFs out. One thing that helps a player's average with ROB or with Runners in Scoring Position is that he can hit SFs which don't count as ABs so they don't lower batting average when they happen. For a good discussion on what effect this might have in evaluating clutch hitting see Tom Ruane's article "In Search Of Clutch Hitting". Elan Fuld's work takes SFs into account when analyzing clutch hitting (see sources). More on this when I discuss ""Total Clutch Hitting Statistics" below. But, if we try to see if he was especially good at hitting SFS, we run into a problem-we don't know how often he hit SF-type fly balls when there was not a man on third. Maybe he had a lot of SFs because he hit alot of flyballs.
Another thing that might have helped Murray with ROB is that he was a switch hitter. The opposing team could not bring in a righty or lefty to face him with ROB and have the platoon advantage. But maybe in Murray's case this was not true. His AVG-OBP-SLG against righties was .292-.368-.491 while vs. lefties it was .276-.341-.446. His dropoff against lefties is maybe a little less than what we would expect from a pure left-handed hitter. So perhaps the switch hitting helped Murray (see David Smith's article "The Effectiveness of Platooning"). I did find in a study of players from 1987-2001 that switch hitters increased their average about .0027 points more than other hitters when going from NONE situations to ROB situations. So part of Murray's success with ROB might be due to this.
But maybe we should just look at his hitting clutch situations, period, without looking at how he did then compared to other times. His clutch numbers are certainly good. But they are not astronomical. To see what the best clutch performances look like, go to my site "Clutch Hitting Leaders, 1987-2001". In fairness to Murray, the 1987-2001 period was better offensively than his career of 1977-1997 and some of the leaders I have listed at this site had not yet played their whole career, so their averages had not started to decline. To give you an idea of the differences, Murray faced league averages of .264-.331-.403 in AVG-OBP-SLG while Frank Thomas faced .270-.340-.422. Those differences are probably not big enough to significantly affect where Murray might rank in any stat for the different clutch situations.
Moving to Close and Late situations (CL-situations when the game is in the 7th inning or later and the batting team is leading by one run, tied, or has the potential tying run on base, at bat or on deck.), the stats show Murray did no better here than at other times. That's a little misleading because if you go to my General Clutch Data site, you will see that normally hitters do worse in those situations. So just maintaining your normal OPS is pretty good. The normal dropoff is about 4.5%. But for a small goup of switch hitters, I found that it only fell about 3%. If you got to my "Clutch Hitting Leaders, 1987-2001" site, however, you will see that some players have actually done better in CL situations than NONCL situations, although Murray's maintaining his normal OPS would be near the top in terms of the CL/NONCL ratio.
It is not clear how many more games Murray's teams won because of this. If Murray maintained his normal OPS when it was supposed to drop 4.5%, that raises his team's OPS in CL situations by about .5%. The impact OPS in CL situations has on team winning percentage can be seen at my site "Does Team Clutch Matter in Baseball?". Using the parameters found there, I estimate that Murray added about .23 wins a season from his CL performance.
There is a more comprehensive way to look at clutch hitting. That is to take into account every plate appearance (PA), not just certain situations. Then check to see how much a team's probability of winning changes as a result of that PA. For example, hitting a HR with the score tied in the bottom of the 9th will be a huge increase in your team's chances of winning while hitting one when your up 12-0 will be a very small increase. I call such stats that take this into account ""Total Clutch Hitting Statistics". They take everything into account, hits, walks, outs (which lower a team's chances of winning and are a negative for the player), SFs, etc. If you go to this site, you can see that Murray ranked just 18th in this kind of stat among the top 100 in PAs from 1972-2002. I also found that a hitter's OBP and SLG relative to the league average, even though they are not adjusted for situations, pretty much explains their "win probability added" or "player win value" or PWV, as some of these stats are called. I found that the following equation for PWV/PA explains it very well
PWV/PA = -.024 + .00015*OBP + .000095*SLG
For both OBP and SLG, it is relative to the league average. For Murray, those numbers are 109 and 118, respectively (from Lee Sinins' Complete Baseball Encyclopedia). The 108, for example, means Murray was 8% better than the league average. This formula predicts that Murray would have .00356 PWV/PA while he actually had .0037. That difference over 700 PAs (perhaps a full season) is about .098 wins. Not a huge amount (if I plugged in the exact relative OBP and LSG, 108.5 and 118.11, it would be about .14 wins)
What about pennant races? Did Murray have a significant impact there? His career September AVG-OBP-SLG were .316-.392-.520. Generally, hitting stats are down just a bit in September. Although Murray went up and it seems like a lot, the key question is how often did he elevate his game in September when his team was in the hunt for a playoff spot and what impact did this have?
The table below shows how many games from first place or a playoff spot Murray's team was as of August 31st and at the end of the season. At the end of play on August 31, 1977, the orioles were 4 games out of first. They finished 2.5 games out. A positive number means his team was that many games ahead. The AVG column is simply the average of the Aug, 31 and end of season figures. The next column, OPS ratio, shows what Murray's OPS was that year relative to his OPS for the rest of the season. In 1977, for example, Murray's OPS was 30% better in September/October than it was for the rest of the season prior to that.
The last column is an estimate of how many more or fewer wins Murray's team had due to his Sept/Oct performance relative to his earlier performance. In 1977, his OPS up through August was .803. In Sept/Oct it was 1.043, a gain of .240. Assuming he was one-ninth of his team's PAs, that raises his team's OPS by about .027. My site on team clutch mentioned earlier shows that winning percentage goes up by 1.27 for every .001 gain in OPS. A gain of .024 for team OPS would raise pct by .034. Over 30 games (about the last month or so), that is 1.02 wins. So that is a very big impact for one guy to have on the pennant race and it was his biggest impact in his career.
In some years, Murray's team had pretty much no chance to make the playoffs once September came. In one year, 1995 with the Indians, they were so far out in front there was no race. So I took all the critical years, where the AVG games from first place was 5.25 or less and found Murray's average OPS ratios. In those years, it was 1.126, meaning that his OPS in Sept/Oct was 12.6% higher than the rest of the season. In other years, it was only 2% higher. So Murray really did elevate his game in September when a playoff spot was at stake. But I did not attempt to see if this was statistically significant. I included 1979, a year the Orioles finished in first place, as a non-critical. He did 4% better in Sept/Oct that year.
The one last thing to check is how he did in post-season play. The table below shows how he did during the regular season and the post-season in years his team made the playoffs. The post-season numbers includes divisional series, league championship series and World Series play.
I did not include 1996 since he had only 17 post-season PAs that year while he had at least 39 in all of the other years and averaged about 45. There are no signs of clutch hitting here. Sure, he was facing better than average pitchers in the post-season. But there is nothing extraordinary going on since his numbers go way down from what he did during the regular season.
"Clutch and Choke Hitters in Major League Baseball: Romantic Myth or Empirical Fact. By: Elan Fuld
STATS, INC Player Profiles books
The Great American Baseball STAT Book