Taking a sample size of only 5 or so hands doesn't prove anything. Statisticians won't take a sample size under 30 seriously. You need at least 30 sample's (in this case hands) to get a normal distribution. At this point the true odds "may" be visible. In the case of poker 30 hands is way to small due to the varying odds on each hand. If you have 30 hands of say AK vs QQ you will see the appropriate odds, but if all you have is 30 hands of coin flips and 60/40's the odds may still not be accurate to the true value.

Poker is a game of variance, you get cold. For example, I just lost A10vA4 and A10vA2, two hands that I should be winning the majority of the time. But that's poker, shit happens. Just have to move on