I've been working on various common live scenarios and I am noticing that certain hands in certain spots need to be played agressively a percentage of the time and less agressively the rest in order to balance out ranges. Using software like pokersnowie it shows that in some cases I should be raising like 64% of the time and calling the rest. How would I implement that in a live or even in an online session? I don't typically wear a watch but I've been thinking of buying a watch that has a second hand and using that in some way for a source of semi randomness. In the case of 64% I would most likely remember that it was between 50 and 75% and before acting base my action on how far through the current minute that the second hand on my watch is.

There are loads of things you can use as pseudo random number generators, such as the colour/suit of one of your holecards (or the first card on the flop).

e.g. Suppose you wanted to 3-bet 65s pre-flop but only 25% of the time, you might decide to only do it when it's 65cc (I prefer hearts, myself), or supposing you wanted to 3-bet AQo half the time, you might choose to only do it when the ace is a red card (hearts or diamonds). If you wanted to use more specific percentages, like the 64 you mentioned, a stopwatch (e.g. on your phone) would work. Just press to start, and then randomly stop the clock and see how many hundredths of a second are shown after the decimal point. For online play, you could use an online RNG, like the one at https://www.random.org/

I don't think it's humanly possible to memorize all of the correct frequencies.
I do think it's possible to play a very strong strategy without the help of random number generators or tricks that tell us when to check/bet/call/raise.

These are the possibilities:

a) results in a very poorly constructed strategy.
b) results in a strong strategy.
c) results in a very strong strategy.

(d) and (e) need no explanation here.

(b) and (c) are closely related. While (b) is slightly exploitable, it bears very close resemblance to (c) as the only difference will be the frequencies.

The important thing here is that you're using a mixed strategy with the right hands. It would take a large sample of hands for your opponent to figure out that you're betting 10% too much here or 10% too little there. Even so, the future is never set in stone, thus there's not much your opponent can do vs strategy (b).

Trying to get exact mixing frequencies down is futile. First of all remember the action is only mixed at equilibrium, in reality when you play against a human opponent one action is almost certainly better than another.

Of course if you want some way to split hands of similar value just take the aggressive action when you have backdoors/overcards and the passive one when you have no redraws/undercards. This is what solver solutions trend towards anyways so it's the most efficient way to incorporate some pseudo mixing into your strategy while not complicating things too much.

Good point

You can put an actual RNG on your phone and use that. Or a watch and convert to a base 100 system.

this. if solver advocates to do a certain action with f.e. 3 backdoor flushdraw combos 80% of the time and to do it 20% with the 9 non-backdoor flushdraw combos, just do it 100% with the backdoor flushdraw combos and 0% with the others. almost always you will end up with frequencies sufficiently close to gto.

Would some particular situations be far more appropriate for whatever action you're assigning a frequency? For example if hero 3-bets a range 25% of the time would it be more practical to base it on table dynamic and the characteristics of the villain in this specific hand? Or would that defeat the purpose of it being random?

I have the loose theory in my head that multiway poker leads to pure strategies*, and that heads up poker leads to mixed strategies; the exception being that your opponent has an exploitable tendency, which would lead to a pure strategy when heads up. Like if the opponent calls a little bit too much on the flop, then the stronger semibluffs would be 100% bets, and the weaker semibluffs would be 100% checks.

*Preflop in 3+ way pots, for example, I just raise the range of hands that I think are profitable from my position on the button. Then if there are 3+ players postflop I wouldn't mix until it was heads up.

This is very much how i think about the game.

You don't need to have perfect frequencies of perfect hands in each spots. In a lot of spots like mixing hands otf into bet and check ranges its much more important that you have certain type of hands approximately correct amount in each range.

Like having some% of TP in xb range so that when we defend enough vs probes, our calldown range doesnt become too weak which can be exploited by valuebetting thinner. So if you look a solver solution, it would for example check like 20% of all TP with some exceptions. But in practice you can just check back like few of the weakest tp combos almost always and end up with a solid strat.

Same with like wanting few FDs in ur checkback range, just figure out which fds make most sense to XB and xb those often and bet everything else.

Where rng becomes more prevelant are some river spots where you need to choose bluffing combos from a group of bluffs where all are equal value, but you cant bluff most of them let alone all of them.


One nice tip to get 50-50 mix I've used for years which is handy for pf. If you have 65s you 3bet, if you have 56s you call.





Agree mostly. You shouldn't use "mixed strategies" just for the sake of having mixed strats. But in a lot of situations esp where ranges are tight, you just need to do some mixing in order to have a solid strategy. In many spots esp wide range spots, you can play pure strat with each hand and still come up with solid strategy by just allocating different type of hands properly into each range.

Obviously this is not something you need to do if you have a clear exploit for every decision... but how often you can actually confidently say that one option is better than other?


THIS so much. In most spots villains aren't going to get good idea what our total strategy in each spot is gonna be. But what they can quite quickly realize is the frequencies of our actions in spots. It's way quicker to realize if someone checkbacks and folds to probe often, compared to what type of hands we use in that range.