skrum,
From my perspective, home runs per root game has the same skew problems as kills per root game. Let's look at some numbers, yay.

Babe Ruth hit 60 home runs (And Maris 59) in the first 154 games of the season. (I'm too lazy to find end-season numbers.) Bonds hit 70 homers in 159 games. Mark McGwire hit 70 in 162 games. Matt Williams hit 43 in ~110 games (1994, yay labor disputes).

Babe Ruth: 0.389 home runs/game.
Roger Maris: 0.383
Barry Bonds: 0.440
Mark McGwire: 0.432
Matt Williams: ~0.390

Babe Ruth: 4.835 home runs/root game.
Roger Maris: 4.754
Barry Bonds: 5.551
Mark McGwire: 5.500
Matt Williams: ~4.100

So in home runs per game, Bonds is ahead of Maris by 12.95%. And in home runs per root game, Bonds is ahead by 14.35%. Bonds had more games than Maris, and because of this, extended his 'lead'.
Even better, compare Matt Williams to Babe Ruth. In home runs per game, Matt Williams leads by 0.256% but in home runs per root game, Babe Ruth leads by a whopping 17.9%. Babe Ruth led in games by 44.

So basically, this favours Babe Ruth because he far more games than Matt Williams. Remind me of how using root games doesn't favour playing a shit ton of games, skrum?

You can also look at the stats from February.
https://spreadsheets.google.com/ccc?key=0AlHnqBf0i6tkdFJBVHpTSkZmcHdPSGowUHAyTjNRaGc&hl=en
Sorting by kills/games (and ignoring fiero since he only had 7 games at the time), Troy is in 13th. His score is 55.9% of detenmile's score. Det and Troy played 77 and 1005 games respectively. However, in kprg Troy's score is 202% of detenmile's.
I dunno about you, but from what I can see, kprg clearly favours playing a large number of games.

why not require 35 games as you do for tapl/tazd?

skrum, your KPRG may be statistically correct, but what is the use of a statistic that simply doesn't show what you were looking for in the first place?
The point of this is to determine who are the most deadly players. But now we have some average deadly players who play a lot at the top, which makes this list futile.

xchloe:

Good way to bias your results. You take the full amount of the smaller group and a partial amount of the group that has the larger number of games. So your two samples have about the same number of games. The five players have the same order in both sets of standings. I wouldn't expect anything different. If all had **exactly** the same number of games, it wouldn't matter whether you used total runs, runs per game, runs per root game, runs per cube root game, as long as the exponent is nonnegative.

xchloe:

35 games are required for tapl/tazd because there are seven cells for expected and observed values to go into and the minimum for Yate's Correction is 5 per cell.

In kills or home runs there is only one cell. If Yate's Correction still applies, a minimum of 5 games would be needed. It might be a good thing for me in future to require a five game minimum for KRPG.

Shevar:

I use the top 100 because the data are easy to collect. There are tazzledazzlers out there who are better but who are not in the top 100 but it takes too much work to find them. So I don't pay attention to the not top 100's except when I do the multimonth or yearly competitions. Then the situation is reversed. It it too costly to find multimonth top 100's; better to invite anyone and everyone to post and I can easily access their data once I know who they are.

But I could add the best total number of kills player(s) to the my KRPG list.

Wait skrum, are you denying that kills per root game biases people who play more games?

Nope.

and that it therefore takes very little skill?

The numbers are barely off from the full season, but just for you, I'll redo the stats since I'm not in class now and have time to look up full season statistics for the players. I'll even throw in Mickey Mantle, since you referenced him. Numbers came from baseball-reference.com

sorted by home runs/game.
Games played, home runs, home runs/game, home runs/root game, player name.
155 70 0.452 5.623 Mark McGwire
159 66 0.415 5.234 Sammy Sosa
151 60 0.397 4.883 Babe Ruth
112 43 0.384 4.063 Matt Williams
161 61 0.379 4.807 Roger Maris
150 52 0.347 4.246 Mickey Mantle

sorted by home runs/root game.
155 70 0.452 5.623 Mark McGwire
159 66 0.415 5.234 Sammy Sosa
151 60 0.397 4.883 Babe Ruth
161 61 0.379 4.807 Roger Maris
150 52 0.347 4.246 Mickey Mantle
112 43 0.384 4.063 Matt Williams

I'm sure you notice that Matt Williams moved to the bottom of the list because he has far fewer games than the other players, whose number of games are similar.

"The five players have the same order in both sets of standings."

Williams leads over Maris and Ruth in home runs per game but falls behind in home runs per root game, so they don't have the same order. I stated in my mini-analysis that in hr/g Maris was ahead (just barely) of Babe Ruth but in hr/rg Ruth was far ahead. So I guess you didn't read very thoroughly?

even if you don't like my baseball stats look at the february dice stats. the order for kpg is very different from kprg, as I noted with Troy as an example.

So basically, play tons of games and get a few kills if you want to win kprg? thus making kprg a pretty crappy stat since it's so easy to achieve?