I wanted to see which professional players have the highest and lowest standard deviations among their map to map HLTV rating. I used data from the last 18 months vs top 50 teams, only considering including players with 50 or more maps. Calculating the standard deviation by player was easy but I ran into a problem. There was a clear correlation (r=0.65) between HLTV rating and standard deviation. I didn't want the "consistency" of each player to be heavily skewed by their rating so I used a linear regression model to get residuals from each player so there was no correlation between HLTV rating and "consistency".
Highest standard deviations:
donk 0.451
s1mple 0.441
cobrazera 0.427
blameF 0.415
ZywOo 0.407
Lowest standard deviations:
cadiaN 0.274
karrigan 0.276
Snax 0.282
Aleksib 0.287
Boombl4 0.290
Lowest Residuals (Most Consistent):
Moseyuh -0.054
sausol -0.053
mezii -0.044
Neityu -0.040
lauNX -0.038
degster -0.037
KRIMZ -0.037
Jame -0.036
Magisk -0.036
Boombl4 -0.035
Highest Residuals (Least Consistent):
cobrazera 0.080
susp 0.068
s1mple 0.063
FL4MUS 0.061
ultimate 0.052
piriajr 0.050
huNter 0.047
fame 0.047
hallzerk 0.040
Yekindar 0.039
Notes:
Negative consistency (residuals) means more consistent.
Data from 200 players (Mostly players on top teams and others I was curious about).
I also calculated some HLTV Rating percentiles for each player in the spreadsheet linked below.