Quote:
Originally Posted by rpriske
The 'in your head' thing aside, I have never understand the fascination with WHIP. It tells you very little. Why would you not used opp OPS (or equivalent offensive stat) instead?
If a pitcher has a WHIP of 1.00 but most of those hits are home runs he is a much worse pitcher than one with a WHIP of 1.25 if most of those hits are singles.
|
I think you should have choosen something beside OPS. it's completely mis-weighted and misrepresents data. it makes no attempt to be realistic. someone arbitrarily added two stats together without any thought on the matter. (you said other offensive stats, too. don't think i am riding you. i just have a strong distaste for the logic behnd the creation of OPS)
opponents average/obp and WHIP are the exact same thing. they are just presented differently. if one is more intuitive for you, then that will be your preference.
whip could be improved by weighting walks differently than hits, relative to % chance of scoring runs. it has flaws, and most stats do whether traditional or newer.
your example of why it's not helpful is quite biased. it presupposes that you can only use whip without other information. no rational person would use whip as in your example. the point is to use a matrix of metrics to evaluate value in a consistent and objective manner. once you have consistency, you can make use of experience more efficiently.
so you wouldn't just compare whips between a 1.00 and a 1.25 pitcher. you'd look at hr/9 and such, too, which he mentions in the initial post as stats he likes for evaluating pitchers.
the metrics that are rates, as opposed to totals, will give you immensely more useful information.
so you use whip, hr/9 bb/9 h/9 etc. and nothing prevents you from also using avg/obp/slugging against, either. where info is lacking in one stat, you use another to fill in the blank. this isn't faith. absolute belief in any particular stat as being unfallible is not required. some stats have limitations. some only tell a portion of the story. by using more than 1 or 2, you can draw a much more accurate picture of reality.
derivative stats like WAR use those stats weighted in an agreed upon manner. the standard may or may not be the most precise method possible. this explains why they are not absolute figures, like hr/9. subjectivity and human error is involved. i do believe it will be refined into a better form in the future. until then, i trust my own weights more. i prefer to use the individual parts and go from there.