People like to say, âthe way to game the system is to schedule bad teams and beat them by 40,â but that doesnât work with WAB (it doesnât look at MOV). Beat bad teams, youâll gain next to nothing; slip up even once in those games (see VT), you wipe out all your gains. Itâs a fantastic resume summary.
(And yes, there is value in predictive metrics too. For one they, they are more⌠predictive. But I love the simplicity of WAB for resume comparison.)
The other thing I like about WAB is that you have a reasonable way to compare sayâŚMiami (OH)âs resume to Kentuckyâs in a way that you just canât do by tallying Q1-Q2 records.
One step closer to giving the committee a golf weekend and letting Steve (just gonna assume his name is Steve) push the button on the computer that runs WAB one more time on Selection Sunday.
Quads are just a way to dumb things down into discrete chunks (losing signal in the process) enough so that humans are able to do an analysis that the continuous metric does a better job of already.
But we both know the quads arenât going anywhere. The committee is going to have to miss a round of golf.
Interesting that EvanMiya has Virginia as a pretty consistent team. Also, I guess he has the Hoos as a fringy title contender (or maybe Iâm being overly optimistic).
Their results might feel bouncy to us who watch them relentlessly, but are actually less bouncy than what other teams go through (who we watch a lot less relentlessly).
This filtered for freshman with >100 minutes played. As one would expect, recruit rating is correlated with performance, but thereâs a lot of player-to-player variability.
The outlier 4-star in the PORPAG graph is Ebuka Okorie, the outlier 3-stars are Keaton Wagler from Illinois and Isaiah Johnson from Colorado. The other 3-star in the list of top-15 freshmen by PORPAG is Leroy Blyden Jr, a hot-shooting 6â1â PG at Toledo.
Mostly because whatâs being measured isnât aligned with the name of the stat!
But also because it doesnât vary much, so talking about ranks is misleading. (E.g. #50 sounds way faster than #250, but the difference between those two is only like 4 possessions a game.)
Beyond âvibe codedâ? Dunno, but itâs kinda nice that itâs raised the quality level of the hobbyist website (though there is the âincrease in slopâ tradeoff).
The authorâs GitHub is public too, which is also very nice for transparency purposes:
