APBRmetrics

[ryan.]

I've yet to see anywhere PER being applied to single game performances. While I understand it's commonly a measure of season totality for a single players effort, I see no reason as to why it couldn't be applied to a single game performance. I don't, however, know how it could be done.

Perhaps multiplying the player's totals in that game by the average amount of games played for league starting fives (or take the easy route and simply multiply by 82) and attribute those numbers to the PER formula?

[davis21wylie2121]

For single games, John used to have something called "game score," like the Bill James pitching stat of the same name:

GS = Pts + (0.4*FG) - (0.7*FGA) - (0.4*FTMiss) + (0.7*ORb) + (0.3*DRb) + Stl + (0.7*Ast) + (0.7*Blk) - (0.4*PF) - TO

The scale went like:
50+ is a historic performance.
40-49.9 One of the best games of the season by anyone.
30-39.9 Probably best performance in the NBA that night.
20-29.9 Player of the game.
15-19.9 A very strong effort.
12-14.9 Resonably solid performance.
8-11.9 Average.
5-7.9 Subpar game or did not play many minutes.
0-4.9 Poor game or riding the pine.

[ryan.]

Interesting. Being relatively new to basketball statistics indepth and not having read any of Hollinger's books yet, I haven't heard of GS. Is it a relatively outdated way of measuring a players single game performance, or still the standard?

I realise this is probably a point not often touched on because single game performances are relatively easy to analyse merely by first-look perception, but I'm working on a little something that would benefit from a particular system. I was looking at how I could incorporate PER to single game performances, if and how that would be relevant and how I could do it.

But if GS is still a standard for this aspect, it will suffice.

Cheers.

The only problem I have with this is that there's no relation to minute production, rather just total production.

[deepak_e]

[Analyze This]

[John Beattie]

Of course there's always the generic efficiency stat such as offered on nba.com where everything gets one unit, in case you haven't seen that:

http://www.nba.com/statistics/efficiency.html#LeagueEfficiencyLeaders .

[ryan.]

Thanks for the replies, I'll check it them out. I was thinking about generic eff stat beforehand.

"If you're interested on how much a player helped his team win in a given game, why would you adjust for minutes?"

Because I think the rate at which he done so is just as important, and should attribute for something. For example, Player A) Notches 54 points, 7 rebounds, 6 assists on 53% FG in 46 minutes. Player B)Notches 54 points, 4 rebounds, 4 assists on 54%FG in only 30 minutes.

There's a difference there that the Hollinger formula doesn't take into account.

[gabefarkas]

[Mike G]

[tHe_pEsTiLeNcE]

[deepak_e]

[supersub15]

Can somebody explain the difference between FG and FGA in that formula? Is FG = FGM?

[ryan.]

^Yeah, it is. Usually if you see FT, FG, 3P it means FTM, FGM, 3PM when concerning individual's statistics. All else will have tm (team) or lg (league) infront of them, or A (attempted) after them.

You probably know all of that already, but just in case...

[John Beattie]

Is it correct that the game score formula

GS = Pts + (0.4*FG) - (0.7*FGA) - (0.4*FTmiss) + (0.7*ORb) + (0.3*DRb) + Stl + (0.7*Ast) + (0.7*Blk) - (0.4*PF) - TO

is equivalent to

2PM........+1.7 (= +2.0 + 0.4 - 0.7)
3PM........+2.7 (= +3.0 + 0.4 - 0.7)
FTM........+1.0
2Pmiss.....-0.7
3Pmiss.....-0.7
FTmiss.....-0.4
oreb.......+0.7
dreb.......+0.3
ast........+0.7
stl........+1.0
blk........+0.7
to.........-1.0
pf.........-0.4

...as a way of assigning a single exact value to each type of event so as to make the equation easier to visualize, rather than mixing values between points and made shots and attempted shots and misses.

By comparison the nba.com efficiency formula

EFF = ((Points + Rebounds + Assists + Steals + Blocks) - ((Field Goals Att. - Field Goals Made) + (Free Throws Att. - Free Throws Made) + Turnovers))

from that same location

http://www.nba.com/statistics/efficiency.html#LeagueEfficiencyLeaders

comes out equivalent to

2PM........+2.0
3PM........+3.0
FTM........+1.0
2Pmiss.....-1.0
3Pmiss.....-1.0
FTmiss.....-1.0
oreb.......+1.0
dreb.......+1.0
ast........+1.0
stl........+1.0
blk........+1.0
to.........-1.0
pf..........[0]

...is that correct.



Edit: correct FTmiss in the game score formula to be -0.4 instead of -0.7.

[asimpkins]

I think it's supposed to be -0.4 for a missed free throw, not -0.7.

(And yes, your reconfiguration looks correct, except for the FTMiss value.)