We have differing opinions here sounds to me like I am more in the balance camp and you are in the exploitative camp.
In poker, you should always be exploiting the situation as best you can. That's the whole game, right there.
I agree with BG here; you should be considering the other player's current image/play style as well as your current image. If you do not, you're leaving a lot of money on the table.
Why?
Because your approach, that of falling to a default strategy which is generally +EV, is inherently weak. A game-theoretic strategy may be unbeatable when pitted against a pro or against another game-theoretic thinker, but that's not my goal when playing poker. My goal is to win the maximum amount of money versus a given opponent. This requires a strategy that is at least attempting to tailor to that opponent.
Your goal shouldn't be to play in a way +EV in all situations. Your goal should be to play in a way that is Max EV in the current situation... and that will lead to a higher EV over time.
A 2 hour session with a player live is at best 50 hands this is a meaningless sample imo.
I see this thinking as your fundamental flaw.
You are correct in that this is a small sample size from which to draw a statistical inference regarding behavior in any given situation (such as how they play given starting holdings in a given position.) But you are
not limited to statistical inferences from those aggregated outcomes. If the hands you see indicate a given play style, you can, in fact, infer a great deal about that player at the moment.
Someone whose prior play was solid, but who has just been on a lucky streak and has own a bunch of money on good river cards is liable to loosen up. If you see a few behaviors that suggest this has happened, you should immediately adjust your play to take advantage of this.
Likewise, you absolutely must think about your own image. These random strangers at a live table have no prior knowledge of your play style. They've seen you for an hour. In that hour, if I've been card-dead, I've been uninvolved - therefore, most of the table thinks I'm very, very tight. I use this knowledge to increase bluffs and semibluffs and isolation plays - anything that uses a show of strength, I can leverage. Or if I've been getting a lot of solid cards and they've seen me play a lot. They won't necessarily see a lot of showdowns - I'll be taking down pots - but they've seen my play a lot. Then I know I shouldn't start going out on a limb... but when I do hit cards, I can bet more heavily. Why? There will be people saying, "he can't ALWAYS be hitting cards." So I tighten up, especially against the smart, statistics-minded game-theoretic players. They hope to take down my bluffs, so I'll just bet solid hands. I'll even overbet my good hands, making them think I'm bluffing, getting huge value from their marginal hands. But I won't tighten up against those players at the table who seem to be superstitious, who are afraid to "get in a hand" against me, because I'm "hot." This is nonsense, but it's nonsense they believe. And whenever one of the statistical guys tries to call down one of my fake bluffs (value overbets), the superstitious guy will be all the more certain I'm just "running hot."
Most people at a $1/$2 game are not thinking this deeply - they won't be adjusting their play against their current image. But you certainly should... and you should take full advantage of the knowledge they give you about their current play style. Not to do so is very -EV.