Post-Season Play: Do Regular Season Records Matter?

As the Diamondbacks head toward the post-season [usual disclaimer taken as read, please], thoughts inevitably turn to the play-offs, and the general consensus is that the Diamondbacks will get steam-rollered by... Well, just about any other team they face. However, is that the case? It's certainly true that every team starts the post-season with a record of 0-0. But does that mean we should throw the previous 162 games out of the window? Was Billy Beane right when he said, in Moneyball: "My job is to get us to the playoffs. What happens after that is luck"? [He was a little more terse...]

Let's see whether regular season record is a good predictor of future success - and also, why Arizona might be better off facing the Phillies in the first round....

Assume we do play the Phillies, and both teams continue at their current pace. Through Wednesday morning, the Phillies were at a .652 clip, which would project to 106 wins, the D-backs at .570, which projects to "only" 92 wins. If that 82-point gap in W% moves to direct head-to-head play, Arizona would have a .459 chance of winning each and every game, and Philadelphia .541. The chance of the Phillies winning a five-game series is only 57.65%, compared to the D-backs 42.65%. Now, this is obviously simplistic, not taking into account home advantage or pitching, but it's still a lot closer to even than you'd probably expect.

It's because of this, that there's an argument that it would be better to face the Phillies in the shorter Division Series. Underdog teams have a greater chance of a surprise victory in a short series, so if we believe our route to the World Series ha to go through Philadelphia, our greatest chance of knocking them of is over a best-of-five. Of course, if you dodge them in the Division Series, you gain the chance of someone else getting lucky and beating them there, giving you an easier match-up in the Championship Series - assuming you make it that far. But which is better, is perhaps not necessarily as obvious as you'd think.

This ties in with what Doug Pappas found, when he reviewed a century's worth of post-season play, back before the 2003 post-season. That covered 200 series: four matched teams with identical regular season records, and of the remaining 196, the 'better' team won only 106. That's just eight more, over a hundred years, than you'd expect if the outcome were decided by flipping a coin. This slim advantage evaporated almost entirely in five-game series, where the better team had a record of 35-32. Pappas found "the better club is more likely to win blowout series, less likely to win those decided by a single game."

So, the playoffs are not a crap-shoot - they're more like a coin-toss, with a nickel that's very, very slightly heavier on one side. More evidence for this can be found in the performance of wild-card teams. This was introduced in 1995, and should, theoretically, have introduced the weakest franchise to the post-season line-up. But in those sixteen years, four wild-card teams have won the World Series - exactly what you'd expect by random luck - and five more have won their league's pennant. That's nine of 32 World Series teams coming out of the wild-card, slightly better than chance - despite always having to face the "best" team in the first-round.

History is littered with examples of regular-season records being thrown out the window, none closer to home than the 2001 Diamondbacks. While they now seem a mighty juggernaut, their 92 wins was sixth-best among the teams that post-season. They were twenty-four behind the Mariners, who were held to a .211 average by the Yankees in the ALCS, and lost in five. All told, two of that season's seven playoff series were won by the 'better' team: Arizona beat the 95-win Yankees, and also the 93-win Cardinals. The following year, only one match-up went as regular season records would predict, and two wild-card teams, the Angels and Giants, duked it out for the World Series.

To a large extent, this may just be small sample size at work. The gap between the best and worst post-season teams in any given year is not usually huge. Last year, for example, the best regular record was the Phillies' 97 wins; the worst, the Rangers' 90. That's only a seven game spread, over 162 contests: next to nothing, when you're talking about the course of a five- or seven-game series. If you are playing October baseball, you're a good side; that has already been adequately proven over the course of the previous six months. Contrast the NFL, where they had the same seven-game spread last year - except there, that meant playoff teams with records of 14-2 and 7-9...

However, one theory states that wild-card teams might have to work harder to make the post-season, than a team which has been on cruise control for weeks, or even longer [the 2001 Mariners had a twenty-game lead on June 16!]. This allows them to keep their "edge" sharper, than those playing meaningless contests in the period leading up to the post-season. A counter-argument would be that teams not involved in a pennant race, can rest players and get their rotations exactly how they want for the playoffs. Or both of these may be true, canceling each other out.

To look into this, I scoped out the numbers for all 64 National League post-season franchises since the current format was adopted in 1995. For each team, I tallied the wins they had after several points in the season e.g. second half, last 25 games, etc, and also their final margin of victory over the team behind them, in the division or wild-card race. I measured post-season success simply by the number of playoff wins. That could range from zero, if you were swept in the division series, all the way to 11 if you won the World Series. Is there a correlation between this number and regular-season performance? [You can find the spreadsheet here]

On an eyeball of the data, not obviously. For example, only two of the 64 teams had losing records over the second-half of the season. However, these two were the 2005 Padres (38-43) and the 2006 Cardinals (39-42), who had radically different post-season experiences. San Diego were swept in the first round, St. Louis won the World Series. The latter basically backed into the playoffs, winning only three of their last dozen games, before suddenly rediscovering their form and going 11-5 once competition started again. At the other end, three teams went 56-25 in the second-half, including the 1999 D-backs. Two of them suffered a first-round exit.

You can do the same for just about any slice of the regular season. The 2008 Phillies went 17-8 over their last 25 games. The 1997 Marlins were 10-15. Both won the World Series. Go figure, since for any example, there appears to be an equal and opposite counter-example. To cut to the chase I ran all the numbers to see if there was any degree of correlation between regular-season wins and post-season wins. Here are the numbers:

Anything less than 0.10 is seen as "no connection", so the only two stats that have any predictive value - and it's slight at that - are a team's record over the full campaign and the last dozen games. The number is very similar, which means that [based on the NL since 1995, anyway] you can get virtually as good an idea of how a team will do in the playoffs, by looking at their last two weeks of play in the regular season, as at the entire year.

I also did the same for the correlation between a team's margin and their number of playoff victories; that came in at 0.137, so about the same as overall team record. Note that this was a positive correlation, so the bigger a margin a team had, the better they tended to do in the playoffs. It looks like the 'rest and sort your rotation' argument is better supported by the numbers, though the evidence is slight here too.

If you want the "tl;dr" version, it's good news for us. The Phillies may have easily the best record over the regular-season, but that does not translate into post-season success. In fact, the 1995 Braves were the only National League team since the post-season expanded, to have the best record during the regular campaign then win the World Series. Subsequently, six NL teams have won it all - none had the most in their league. Indeed, to find the last time any Senior Circuit team took the World Series with more than 92 regular-season wins, you have to go all the way back to the 1988 Dodgers, who won 94 - and they had to beat the 100-win Mets in the NLCS.

The bottom line is, for the Diamondbacks, Phillies, or whatever teams reach the post-season, their fate will be determined solely by what they do in the play-offs.