November 21, 2011
Is Assist-to-Turnover Ratio a fair way to judge a point guard?

Marquis Teague

Above: Marquis Teague, probably throwing the ball out of bounds. I KEED

I made the godawful mistake of wading into the comments section of Kentucky Sports Radio this afternoon when I followed a link from A Sea of Blue mentioning Teague’s progress measured against Kentucky’s last two starting freshman point guards, John Wall and Brandon Knight (note: he’s doing okay). Aside from the general conclusory dumbassery that is prevalent in KSR’s comments section, one guy paused to say that the best way to measure a point guard’s effectiveness is the assist-to-turnover ratio.

Is that true, though? Think about what the PG’s job is: run the offense and defend the perimeter. Involved in that is taking care of the ball and doing your best to ensure that your team scores when it has the ball and making sure the other team doesn’t score when it has the ball. Turnovers mean that the guard didn’t help the offense score; assists mean that he passed the ball to someone who made a basket. I think that’s too narrow a metric, though. Shouldn’t made baskets by the point guard count? What about steals? Plus-minus?

I’m a lawyer, not an accountant or some other kind of statistician, so I don’t know the answer, but I think that steals and made baskets should have some influence on how we rate a PG’s effectiveness, not just the assist. Sure, passing the ball to the guy who’s open is an important skill, but what if the PG is open and scores, and he called or initiated the play that got him open? Shouldn’t he get credit for that? I think so. So from here on out, I’m going to chart Marquis Teague’s over-all performance in games in a statistic that looks like this: (Assists + FGM + STL)/(TO).

Yes, these are solely derived from the box score. Other factors could and probably should play into it, such as plays on defense where he bottles up a scorer or gets scored on, but that takes too much time and paperwork. AFGMS/TO it is for now, and bully for the guy who comes up with a way better name for that.

Teauge’s AFGMS/TO in the four games so far:

Marist: 4 assists, 7 FGM, 3 STL, 3 TO; 4.66 AFGMS/TO

Kansas: 3 assists, 4 FGM, 0 STL, 6 TO; 1.16 AFGMS/TO

Penn State: 2 assists, 3 FGM, 1 STL, 3 TO; 2 AFGMS/TO

Old Dominion: 2 assists, 3 FGM, 1 STL, 6 TO; 1 AFGMS/TO

Okay, so he’s regressing a little bit. I have to think, though, that these numbers will improve to some kind of median much closer to the Marist game as the season goes along. At least let’s hope so. If you have any suggestions or tweaks to the formula, let me know. For example, should missed field goals count against him? I’m bad at this.

  1. anythingbutgatorade posted this