padres shopping kouzmanoff.. hmm

I find your philosophy of creating a larger sample size by including part-timers to be invalid in the context of this metric’s purpose: evaluating a player’s offensive output within the context of your need to plug him in your everyday lineup.

Manny’s other 9 seasons in the last decade WOULD be part of my sample, so his everyday starting talents would not be discarded because my sample size is so much larger than one season. In judging Manny’s talents as a starting corner outfielder, one would be foolish to judge him by a single partial season’s stats instead of his last decade of productivity. The larger the relevant sample size, the truer the results of the metric.

When players play every day, they get to the 502 PA benchmark more often than not. If they fail to reach that plateau, then they don’t qualify as starting position players, or are injured/suspended part of the year, thereby making their seasonal averages from such a small sample size (within the context of a single season) that those exceptions to the norm should be discarded in favor of the seasons in which they DO play every day. You can argue to lower the bar beyond that to 400 PA’s or even 300 PA’s, but would you be happy going into a season knowing your everyday player was only going to get 400 PA’s? That doesn’t sound like a lineup solution to me, it sounds like a player-personnel problem that will leave your team hamstrung for around 40 games, which takes us away from the purpose of this analytical pursuit.

The reason that Jay Bruce, along with guys like Willy, don’t meet that standard in 2009 is that for justifiable reasons, they were not everyday starters that year.

For Bruce, it was that he was injured a good portion of the year.

For JJ Hardy, it was that he was out of the lineup for 5 different stretches of time last year with injuries to his back and shoulder, and spent part of the year in AAA (due to demotion, not injury). If a player is injured and demoted during the year, causing him to not qualify for my standards, then he is, by definition, not an everyday starter that year.

I’m throwing out partial seasons because they skew the numbers unhelpfully. In persuing/evaluating starting shortstops, I do not find JJ Hardy’s 2009 numbers valid for statistical analysis in this context. When looking at a everyday starting SS for the upcoming year, I find it far more valid to look at what everyday starting shortstops produce, on average. I don’t find including JJ’s 2009 in this sample as helpful, whereas he WAS the starting SS in 2007 & 2008, which are included in my study.

Willy is not everyday starter quality, and THAT’s why he only played in 102 games, and THAT’s why he is not a suitable benchmark for comparing starting outfielders. If you’re using Willy as your benchmark for starting CF’ers, your analysis will not be fruitful, because he sucks out loud and should have never started more than 20 G for the Reds.

You’d prefer a larger sample of player’s stats that are from players who are not everyday solutions for a team. I do not find your suggestion helpful for valid analysis. You are lowering the qualifying bar beyond the purpose of the study. In judging a starting SS’s quality of offense, comparing him to platoon/part-time players may be what you would prefer, but I wholeheartedly disagree on your proposal’s validity. In the context of evaluating a player you want to plug in your lineup everyday, you want to measure his worth against the other teams’ starting lineups, not against whom they have on the bench.

That discussion aside, I like that these OPS/pos numbers show the difference in skills sets between corner outfielders vs. center fielders: Corner outfielders tend to produce a margin of around +5.9% higher OPS (5.3% in Slyde’s analysis). That’s helpful to know when your team might not have enough offense, and you are discussing moving a CF-offensive-quality player to a corner OF spot, ala C-Dick.