Here are the assumptions: We are comparing apples to apples (among positive EV players) since poker is a negative sum game when you factor in the rake and associated fees.
Other consideration. Given that so call "winning" players is already a pre-selected sample, meaning if you took a random group of players and pre-selected the ones that continues with the game vs the ones that loss their ass in the beginning and stopped playing altogether, you are left with only the so called "winners". This is the group under observation. Therefore the sample size itself is a nonrandom distribution. Everything I just mentioned here is devoid of exceptional skills.
Bayes' theorem, for streaks and runs. There are poker players who seems to consistently win, say straight for 15 years, this is similar to outperforming the SP 500 ( actual example; The Value Trust Fund) for 15 years straight, without exceptional skills. Each year the fund has a 50/50 chance of wining or losing, in case of poker, the player at least has to beat the house rake and other associated expenses. (tipping, theft. etc...) The probability of such a streak is very small (1/2)15th or 0.00031 or 0.0031 percent. However once we factor in the pool of players say 1000 players for 5/10 and 10/20 so call "elite commerce pros" the probability of having 1 or more players having a winning streak goes up to 3 percent. The probability of having 1 or more such player(s) having a winning streak over a long period say 40 yrs (career span of some poker players) out of a pool of 1000 players is goes to about 38%.
Conclusion: Lucky is probability taken personally. The hubris of randomness and lucky mistaken for exceptional skills.