Lucky vs. Good
What does this mean? Well adjusting starters’ WAR against the total number or regular season wins paints an even more confusing picture: teams with great starting pitching are just as susceptible as teams with pitching that carriers less of the load for their team. While great pitching will get you into the playoffs, it isn’t guaranteed to get you anywhere.
You’ve no doubt heard the axiom “pitching wins in the post-season”. It is a phrase repeated ad nauseam at the beginning of every October as teams set their short-series rotation, knocking some poor soul from the back of the rotation into the bullpen. The 2008 Phillies, 2010 Giants, and 2011 Cardinals showed us that a team need not excel in any one particular area to win it all; rather, a team must have a distribution of talent across all areas and the appearance of good timing. The Phillies were superior in every way to the Giants in 2010, but lost to them in the NLCS because short series produce a lot of variance, and most of it fell in favor of the Giants. As Kurt Vonnegut would say, “so it goes”.
Ian contacted me and asked if I had any thoughts on what other factors could be influencing his findings and if randomness isn’t the only answer. I don’t think using WAR for pitchers, particularly fWAR, is the best statistic to use for this investigation. For one, pitcher WAR is influenced by innings pitched. Teams that bring up the stud pitcher in late June and ride him into the post-season (e.g. Madison Bumgarner) will have less pitcher WAR than the team who used the slightly above-average veteran since the start of the season (e.g. Joe Blanton). Additionally, eating innings makes a big difference over the course of 33 regular season starts, but when one ace is going up against another ace, the difference between the two will be minimal at best. Pitcher WAR will exaggerate the difference between two aces in one post-season game by accounting for regular season innings.
On a related note, Ian found that the American League seems to have much better results than the National League — roughly 4 WAR. Innings pitched comes into account here again because AL teams can leave their pitchers in longer as they are not forced to pinch-hit to attempt to score a crucial run late in the game. Last year, starting pitchers accounted for 67.8 percent of all innings in the AL and 66.9 percent in the NL. While this doesn’t seem like a lot, over the span of 20,000 innings (14 teams * 162 games * 9 innings per game), the 0.9 percent difference accounts for nearly 200 innings. Essentially, the AL is getting an extra pitcher every year when compared to the NL like this. When you account for survivor bias (better pitchers stay in the rotation longer, appear in more games, and stay in games longer), you have your 4 WAR difference.
Secondly, FanGraphs WAR (fWAR) uses FIP rather than straight runs allowed, so it is retrodictive rather than descriptive. A team with a large fWAR-ERA disparity will not be accounted for properly. For instance, the 2011 Chicago White Sox had an impressive 3.66 FIP but a 4.10 ERA. They finished second in fWAR to the Phillies, 26.4 to 27.5, even though the Phillies’ ERA was more than a full run per nine innings lower.
Since we’re interested in finding out how important pitching is in a particular post-season, we should use runs allowed to answer that question. Another method is a post-season only retrodictor, compared to the regular season ERA or its retrodictor. For instance, in the 2011 World Series, the Cardinals had a 3.86 ERA but a 5.45 FIP while the Rangers had a 4.65 ERA and a 4.64 FIP. During the regular season, the Cardinals had a 3.79 ERA and 3.75 FIP; the Rangers were at 3.79 in ERA and 3.98 in FIP. Were the Cardinals luckier than the Rangers in last year’s World Series? This particular method would say so.
Still, to Ian’s conclusion that it is better to be lucky than good, I would phrase it differently. It is better to be talented than lucky because talent will always give you more chances to capitalize on that luck. Sure, the 2011 Cardinals were an afterthought going into the post-season, but Albert Pujols did hit three home runs in the World Series while Chris Carpenter and Jaime Garcia — their two best pitchers — were phenomenal in the five combined starts they made in the World Series. For as unlucky as it seems like the Phillies were, they were still one run away from advancing to the NLCS because Jimmy Rollins and Chase Utley produced offensively while Roy Halladay and Cole Hamels nailed things down in their three starts.
The post-season is absolutely a crapshoot. When you place eight teams into a scenario where a champion is crowned in a minimum of 11 games and a maximum of 19 games, variance takes over as commanding officer. That’s less than three weeks of games. The spread in talent between teams is already small enough — Rangers position players posted 4 WAR more than those of the Cardinals during the regular season — where luck has more of an impact. I would compare post-season baseball to Texas Hold’em, where baseball talent is the poker player’s stack of chips. Every player has the same chance of getting lucky, but if you have more chips, you can make luck pay off more for you than for others. To put it more bluntly, if you were entering the post-season, which rotation would you rather have while trying to strike lightning in a bottle: Halladay-Lee-Hamels or Carpenter-Garcia-Jackson?
As it pertains to the 2012 Phillies, their best shot at post-season success is to simply have their most talented players ready to go if and when they reach the NLDS in October. From there, they simply let the chips fall where they may.