Why Sabermetrics? (July 10, 2003)
The following originally appeared here.
We are trying to interpret the data as best possible. The performance of a player is due to 3 things
1 - his talent level
2 - the context in which it was put up
3 - luck
To account for the context, you have to know alot of things. For a hitter, you need to account for
- the park configuration (caters to what style/quality of batter/pitcher)
- the opposing pitcher (LP/RP, GB/FB, fastball/curve, good/bad)
- the oppoing fielders (quality at each position)
- the runners if any on base (and their speed), and outs
- the inning/score
- the weather
- other playing conditions
That's alot of context to account for. Also, every player is affected by his context in some way that is different from an average player.
A player's talent level changes year-to-year (in fact, PA by PA).
And luck doesn't even out over 1 or 3 years, especially as we try to break things down to granular levels.
So, what are we trying to do? Search for the truth in numbers, while accounting for our limitations to interpret that data.
Is that valuable? It's a worthwhile hobby to most. And to those in a position of power, a way to establish with greater confidence the market value of a player, or establish the risk/reward in various baseball strategies.
But, we, as sabermetricians, won't be able to get an acceptance to our marriage proposal from Nadia.
--posted by TangoTiger at 03:29 PM EDT