So it's an incredible small sample size (8k hands at 5nl ignition zone) and these are my "river calls." Based on this, should I just be straight up folding to all bets that aren't basically the nuts? I mean this looks really bad to me.

River calls are going to be negative overall when filtered like this. That isn't particularly bad.

You'd want to look at river call efficiency, went to showdown (wtsd), won at showdown (W$SD), to have a better idea of whether or not your river calls are actually bad.

To further explain the first part, understand that if you're calling a pot sized river bet on the river then to breakeven *on that node* it only needs to win 1/3 of the time. There are an absolute ton of situations like this, so when you run the filter you did obviously it's going to be negative overall, but that doesn't mean the calls are bad.

Best way to avoid shitty river situations is to stop flooding preflop and subsequent flop and turn lines with shitty hands. This makes the river much easier in practice.

The following flow is a decent, but non comprehensive summation imo:

My opponent has bet.

I estimate his weakest and strongest value hand given his line and betsize.

Can I beat any value hands? Yes = call. No = continue below:

Can my hand beat a bluff? No = fold. Yes = continue:

Do I need to adhere to minimum defense frequency vs this opponent? No(many reasons for no). If the opponent doesnt bluff enough at this size given this line to effect indifference on this river for my bluffcatchers(fold ev = 0ev = call ev in equilibrium for bluffcatchers lacking blocking effect) I fold the portion of my range that cant beat value hands. Or no, if the opponent bluffs too much, I will pick spots to selectively exploit by calling more bluffcatchers than my estimate of river equilibrium. If the answer is yes minimum defense must be maintained, then the entire portion of my range which can beat my opponents strongest bluff (most showdown value) must be considered, and a folding range based on betsize must be determined. Then I call or fold depending on my best visualization of my own range and where the combo I'm holding falls on the spectrum relative to equilibrium fold frequency.

Note: the exception to the first q is a vs a nit turn and river bet that never bluffs turn. Everyone else goes to the river with a portion of range that my bluffcatchers can beat, and thus I believe folding a combo that beats a portion of the opponents value range is an error in strategic framework, unless vs the confirmed exception.