A lot of videos I see talking about GTO frequencies to keep their sizings, bluffing combos, value combos, lines balanced. IMO they do not matter because the hand has to go to showdown before they even know if you were bluffing or value betting to begin with.

Imagine you play 10 hands vs someone and you bluff 5 times and value bet 5 times. All your bluffs get called and all your value bets villain folds. Now villain thinks you over bluff when you actually only bluff half the time.

For frequencies to matter you need a big sample right? But doesn't that sample need to be vs that player, with that hand combo, with that board texture? Seems impossible to build a big enough sample for frequencies to ever really matter. I think if the bet is +EV you should bet, doesn't make sense to check because you rolled a 15.

Can someone explain this to me?