I will explain this, because it is grammar, not logic or facts... I gave up on those...
Not sure why the semantics changed.
We use the world correlation describing the connection between 2 or more events that
progress in time. It is usually expressed on a chart where the horizontal line is time and the vertical line is the value.
So we can say, there is a positive correlation between global warming and the CO2 gas in the air, over time. Higher the average temperture, when CO2 increases. (correlation =/=causation but I don't even dare to go there)
Or there is no correlation between gold and milk futures. One stays flat when the other rallies or drops big. Their prices are completely independent from each other, aka not correlated.
On the other hand we use words as similarity, sameness, kind of,etc. to describe 2 things that have characteristics that overlap, aka, similar or the same. These things just exist, no progress in time is necessary. So we say, oranges are a lot like apples (they are both fruits), and peanuts aren't like those 2.
So when comparing poker and blackjack, correlation shouldn't be used. Now if you really, really want to use it, you could say: "There is a positive correlation between my poker and blackjack gains. (over time again) Anytime I win in poker, I tend to win in blackjack too."
But here you are describing the result of playing those games, not the games themselves.
One last example: You can not say Pekelo and Bones are correlating. You can say they look alike, or they post similar views, but they are never correlating.