Last week we took a look at the American League Cy Young race using statistical metrics we created here at Stats and Information.
Now it’s the National League’s turn.
Unlike the AL, the NL sees two clear cut candidates rise to the top in almost every meaningful traditional category. Roy Halladay and Adam Wainwright are both in the top five in the NL in innings pitched, wins, strikeouts, ERA, and WHIP. Halladay leads Wainwright in innings, strikeouts, and ERA while Wainwright has more wins and a better WHIP.
So, let’s take a look at our pitcher win probability method to see if we can find some separation.
As a brief recap, my analytical colleagues and I created an algorithm that issues starting pitchers a probability they gave their team to win in each start based on their innings pitched and earned runs allowed. Using these probabilities we are able to evaluate pitchers in a unique way. We can answer questions like which starters give their team the best chance to win on average, or which starters have the most starts where they gave their team at least a 75 percent chance to win (a superior start as we like to call it). For more information read here.
In our AL Cy Young article we introduced another way you can use these probabilities. By multiplying each pitchers average start by their total number of starts we get a metric we call xWIN. xWIN can be described as the number of wins a pitcher SHOULD have based on their probabilities they gave their team to win over the course of the season.
Here are a number of different NL leaderboards using these metrics.
National League Leaderboard
Superior Starts: Any start where a pitcher gives his team at least a 75 percent chance to win.
Halladay is the ever-so-slight leader in each of these categories, with Atlanta Braves Tim Hudson possibly entering the mix a dark horse. One bad start by Halladay coupled with a superior start from Wainwright or Hudson and the entire leaderboard could change.
Undoubtedly this race is far from decided with one month left. Conventional wisdom and advanced statistical analysis each need this last month to decipher who will pick up the hardware.