PQS in the Playoffs (first in a series)

I've been trying to think of a way to analytically examine the playoffs using PQS.  I first tried looking at a pitcher's PQS for the season and then looking at what they did in the playoffs, and realized that the randomness of PQS would not yield anything in short series like the MLB's playoffs.  Then I hit upon what I think is a good way of examining the power of PQS in the playoffs, and I got that from the game of chess.

In chess, when players play each other for their championship, they either win, lose, or stalemate, which is a tie for those not into chess.  And points are awarded (much like hockey too, I just realized) thusly:  +1 for a win, +0.5 for a stalemate, and 0 for a loss.  I will do that for each playoff series and see how each series ends in terms of who won vs. who won on PQS, and cover all the playoff series in a season.

It is pretty simple.  There are three types of starts:  DOM, DIS or other, which never got a name but I'm going to call it MID.  And DOM beats MID and MID beats DIS, and if they are the same, then it is a draw, a stalemate.  The rationale here is that each type of start is mostly random, thus leading to a draw, a coin-flip on who wins, but it is pretty clear that the majority of the time, the relationships of DOM > MID > DIS will hold.  Summing up the points per the matrix, each team will get what I will call their PQS Score, with the idea of looking to see if teams that have the winning PQS Score typically wins, and once I get enough data, I can do a correlation between winning a series and winning the PQS Score.

Here is the table:



I have no idea what I will find, though obviously, I hope to find that PQS does rule in the playoffs as I've been postulating for a while.  What I hope to see is that teams with the better PQS score, per my matrix above, What I am more hoping is that it will be interesting in some way.

I don't have the time to dig through every series first and then write on it, so this is intended to be a series of posts, depending on time, inclination, and what I find.  I am covering 2011 and 2010 in this post.


2011 Playoffs

In the seven series the team with the highest PQS Wins won once and lost 3 times; there were 3 ties.  So that is not a good start for this metric I am testing out.  In terms of average PQS for the series, the team with the highest average PQS won 3 times, lost 4 times; for teams averaging at least 1 PQS higher than the other team, only 1 of the higher teams won, out of 4.  By these two metrics, high PQS is a poor indicator of success in the playoffs.

However, on a game by game basis, a clearer advantage appears.  In games where one pitcher is expected to win (per the matrix I devised above), the team expected to win (or e-win) had a record of 18-6.  When there were ties, the teams were obviously .500 overall, and there were 14 ties.

Furthermore, in games where the pitchers had a DOM start, they were 21-8 (.724 winning percentage).  And four of those losses were guaranteed because both pitchers had a DOM start.  Taking out those tie games leaves the overall record at 17-4 (.810 winning percentage) when a DOM start was countered with a non-DOM start.

Given the overall 38% DOM and 38% DIS starts, there were a whole lot of mediocre to bad starts in the 2011 playoffs.  Dominance was no guarantee, either:  the Phillies had 5 DOM starts and yet lost to the Cards.  And the Tigers had 4 DOM starts while the Rangers had none, yet the Rangers won their series.

I think what lessons to be gleaned are these:

1. DOM starts are no panacea, but,
2. DOM rules to the tune of around .750 winning percentage for this playoff, and
3. In any case, the bullpen can help a team win a lot of games you have no business winning, based on the PQS starts performance, as evidenced by the Cards and Rangers making the World Series despite bad starting pitching.

2010 Playoffs

In the seven series, the team with the highest PQS won 5 times and lost once.  Ironically, that one loss was by the Rangers in the World Series, to the Giants.  There was one tie.  That was much better overall than in 2011.  The team with the highest average PQS won in 5 series out of 6, with 1 tie.  When PQS was at least 1.0 higher, all three teams won their series.

On a game by game basis, the dominance continued.  Where the starting pitcher had a high PQS score, his team went 17-3 (.773 winning percentage).  Pitchers with a DOM start went 22-12 (.647) but there were 7 ties, so when a DOM start was matched with a non-DOM start, the team went 15-5 (.750).

The playoffs were much more pitching oriented in 2010 but there was still a lot of poor pitching.  The overall DOM/DIS was 53% DOM/27% DIS.  That 53% DOM is very good, but the 27% DIS is pretty bad.  Reflecting the higher quality of starts, DIS starts did not do as well, 6-11, but that was still good considering it was a disaster start.  MID starts (PQS of 2 or 3) were 4-9, again, like 2011, a very poor record overall.  

The Giants had 10 DOM starts and 4 DIS starts in 15 starts:  67% DOM/27% DIS.  Tim Lincecum had 4 DOM starts in 5, with 1 DIS.  Matt Cain had 2 DOM starts in 3, no DIS.  Jonathan Sanchez had 2 DOM and 2 DIS starts, the most on the team.  Madison Bumgarner had 2 DOM starts and 1 DIS start.  The Giants were 7-3 in DOM starts, but 3-1 in DIS starts.  That accounted for the 2 wins of 3 in expected loss games (e-loss).  In addition, the Giants like the Rangers and Cards in 2011, won the ties, with a 6-2 record.  

I think this good record overall is a result of three things:

1. a manager willing to pull the starter out before the game got away, and yet keep the starters happy enough to still pitch well for the manager,
2. a bullpen good enough to both shut down the opposition, as well as do that for 6-9 innings, plus a manager who can mix and match relievers to enable that long stretch of a shutdown, and
3. an offense capable enough to catch up and take the lead.

My next post will cover 2009 and 2008 playoffs, which looks like more of the same as 2010.