Situationalbuntingapologist recently made a fanpost about the lack of a metric that showed a hitters ability against the better pitchers in the league. I found the concept interesting and decided to take a crack at refining it a bit.
The general consensus on the original method was that the sample sizes were small and the selection of top-tier pitchers was somewhat arbitrary. The inherent problem here is that if you select more pitchers in order to increase the number of plate appearances, you will naturally be sacrificing the skill of the pitchers selected. Career stats are also tricky because a pitcher can be great one year and horrible another, skewing the numbers. What I ultimately decided to do is determine the best pitchers from each season and take the Tigers numbers against those pitchers during each season, for their entire careers.
For my "top-tier" I selected the pitchers from each league that ranked in the top 15% in FIP or xFIP. This method usually netted me the top 7-10 starters in each league for each season. For example, these are the top-tier pitchers I used for 2012:
|Yu Darvish||15||9||176.2|| 4.02
|Cliff Lee|| 6
|Justin Verlander|| 14
2012 may not be the best example because the Tigers didn't face any of the top NL pitchers, and two of the AL pitchers are Tigers themselves. However, the method did work better for past seasons and usually resulted in at least 35 PAs for everyday players (Infante set the team record, facing top-tier pitchers 72 times in 2011).
After the top-tier pitchers were determined for each season, it was just a matter of going through each batters plate appearances against those pitchers in those respective seasons. So, without further adieu, here are the results:
|M. Cabrera||Career vs. Top-tier||446||.267||.334||.421||.331||9.2%||20.0%|
|O. Infante||Career vs. Top-tier||376||.254||.282||.370||.284||3.5%||20.2%|
|J. Peralta||Career vs. Top-tier||344||.224||.294||.353||.288||9.0%||27.6%|
|P. Fielder||Career vs. Top-tier||314||.301||.389||.515||.391||11.1%||22.0%|
|D. Young||Career vs. Top-tier||279||.276||.297||.448||.319||3.2%||20.8%|
|R. Santiago||Career vs. Top-tier||187||.172||.251||.258||.235||5.9%||18.7%|
|G. Laird||Career vs. Top-tier||177||.201||.254||.287||.244||5.1%||26.6%|
|A. Jackson||Career vs. Top-tier||129||.250||.318||.362||.304||7.8%||31.8%|
|B. Boesch||Career vs. Top-tier||90||.282||.322||.553||.369||3.3%||23.3%|
|A. Avila||Career vs. Top-tier||72||.138||.208||.231||.200||8.3%||44.4%|
Good pitchers are good. Unless your name is Prince Fielder, in which case good pitchers are pitchers.
Honestly, the sample is still too small to determine much of anything. This could be remedied by adding plate appearances against the top relievers, which would also provide some more interesting insight about late game heroics and such. I also probably could have increased the selected pitchers from the top 15% to 20% and still have had a very good group of pitchers. However, I don't care to put in the work anymore, so this is what you get. But still, for players who have only been in the league a few years (Jackson, Boesch, Avila, Dirks, etc.) the sample size will not be large enough for several years.
Leave your thoughts/analysis in the comments and if anyone would like to contact me with questions or take a look at my data, send me an email at firstname.lastname@example.org.