I am def no expert on this but I think you should go lower than 1% tbh. Watched a sauce vid yesterday and I think he solves for something like .2% but that might be a bit overkill.

First of all, understand what dEV means. It's the difference between the current strategy (found "by chance" at the current iteration) and the (biggest) max-exploit-strategy of either opponents. If for example IP would stick with the calculated strategy and OOP would choose his maxEV strat for every single combo in his range, the resulting difference in EV (which got to be positive!) would be dEV.

In other words, dEV is the maximal possible exploitation we "allow" to happen. And it's measured as % of the current potsize.

Now, to decide what % is acceptable, we should look at the details. I will explain it with two examples:

Scenario 1: We are in a SRP, BTN vs. BB, pot is 5bb, we solve flop. When we stop at a dEV of 2%, it means, we are "forfeiting" 0.1bb - which sums up to 10bb/100. Given we are talking about a winrate of 30bb/100 for BTN, this would be quite a desaster, as the situation happens incredibly frequently, hence will have a huge impact on our button winrate - which again will have a significant winrate on our overall winrate! Let's estimate a BTN open of 25% (everybody folding times BTN open range), SB folding of 75%, BB flatting of 40%. That rounds up to 8%, which means, overall we're potentially (!) losing 2.25bb / 100 in overall winrate! That could turn a winning player into a losing one.

Scenario 2: It's a 4-bet pot, UTG vs. MP, small cbet the flop, check-down on turn and a shove on the river. The pot is 137bb. We solve river for a dEV of 5%. That makes roughly whopping 7bb. How often do we get into this situation? Let's say, we are opening 15%, getting 3bet 3% and 4-bet 3%. We cbet 50%, get called 50%, checkdown the turn in 40% and face a shove on the river in 20%. The result is almost neglible (it's 0.0004%). When we multiply that with our 7bb potential EV loss, it gives 0,000025bb/100. Mmmh ... unless we take some significant overall learning stuff from solving that scenario or at least enjoyed the time we spent it seems like a waste of time, huh? :)

You see, 1% can be too high in one case and 5% can be massively precise in another scenario.

Summary: dEV is the potential EV loss against max exploits, "optimal" percentage depends on the scenario (especially potsize and frequency of actually that situation happening, determining the impact on our overall winrate).

Hope that helps!

BigFiszh

Thank you, and just noticed that mistake, doh!