One thing that has to be noted here is that sample size is hugely relevant for most of the stats. You can't really rely on a small sample of hands to be even close to correct for other stats than vpip and pfr. Someone running 90/40 over say 25 hands is likely to still be a maniac at 250 hands, but someone with a cbet stat of 85 at 25 hands could very well just have been hitting his flops.