+1

And furthermore, you don't have to reply to trolls imo

Everyone is talking about resolving the entire NLHE game... But does someone ever solve a spot? In order to solve the game you need to solve any subgame in it... Can you provide any link?

Just downloaded it. I like the 6max simulation mode they do. I've fed a couple of my own hands into it but already Ive come across a strange one.

I hold 89hh. Flop is 9s3sQc3h9d. You can click on the hand range and it will tell you what villains range is and how many combinations of each hand he can have. Well villain had Q9 of which there are 3 combos - but pokersnowie for some reason is not showing this as an option. Nor is it showing QQ as an option. Am I right in thinking it only shows combos of cards that are feasible and has assumed that QQ or Q9 would have raised earlier on in the hand (didn't raise till river)

Ok, here goes.... You promised to read and respond. 8)

Disclaimers:
1) I think Poker Snowie's marketing is hyperbolic and is probably dishonest.

2) That said I have a subscription, learn useful things from it, and think it is a good product if used properly.

These last two paragraphs are DIRECTLY contradictory. First you say that it is absurd that Snowie helped someone beat 25NL THEN you claim that even a half-assed GTO approximation is enough to learn to beat the micros.

Do YOU even know what GTO is? Sure, the micros are full of bad players playing badly. But, if you don't understand GTO you will do a sub-optimal job of exploiting them. It is very hard to see and understand exploitable play, and how to best exploit it, without understanding unexploitable play. DUCY?

Sure, maybe you can learn a few rules of thumb and beat up on 'the player pool' where they are wildly exploitable. You can't do it well or optimally without understanding optimal play.

Finally, trying to understand GTO play is a perfectly reasonable way to learn. If done well it will lead you on a path to beat any stakes and to truly get good as poker - as opposed to learning standard plays and exploits that happen to work against the current player pool at the current time. I mostly agree with the words in this last part but not in the spirit of it.

Snowie's claims are misleading. BUT I still get a lot of value from looking at Snowie's plays in a LOT of situations. There are plenty of times when the sample size seems too small - and I really wish Snowie would point these out when the recommendations are shown. But there are plenty of other times when you can look through the various players' ranges and get useful insight into why Snowie's recommendations are what they are. Even when I find that I don't agree with Snowie it is a worthwhile mental exercise to figure out what is going on and why.

So, Snowie's marketing claims are misleading. BUT Snowie still provides a VERY useful framework for thinking about and studying the game. I wouldn't recommend that anyone learn Snowie's recommendations by rote - that's not the point - I get a lot of value from it in thinking about GTO and why Snowie recommends that it does.

Hi,

Yeah, I can see how the first parts could seem contradictory. I will admit I did not present my thoughts very clearly. However, I believe my inherent idea was correct so let me explain again.

First off, so we are clear, I certainly understand what GTO means. It means you play a strategy that is optimal against another perfect strategy, i.e. a strategy that cannot be exploited. This yields the best possible payoff considering all other players are aware of their own equilibrium strategies and choose to employ them. In this case, no player can gain anything by unilaterally deviating and we have a Nash Equilibrium. Several of these may exist in the form of pure or mixed strategies but then we get into technicalities and the translation into a poker example gets complicated so I'll leave it at that.


Now, in my first statement I said if someone assumes Snowie is in fact perfectly GTO, using it to play 25NL is absurd. The reasoning follows from my above explanation of GTO. At 25NL, people play very poorly, have unbalanced ranges, etc. Hence, even though you certainly can't go wrong playing GTO against them, it's MUCH better to play exploitative and deviate from an unexploitable strategy. To give a very basic example, if someone simply never folds any piece of the board or A-high, having an air bluffing range in many spots is going to be pointless and is going to cost you money, even if it is GTO. Furthermore, if someone folds to 3-bet 95% and NEVER adjusts, you should 3-bet them way more than optimal because it is profitable in a vacuum. The simplest way to understand this is to look at a basic game like below:



{B,Y} is clearly the NE. However, say the rec player is a complete donk and decides to play strategy X because he doesn't really understand the game or his own optimal strategies. If you play B which is GTO, you gain 5 points. Playing A gives you 100 points, so you cost yourself 95 points of profit per round if the rec. play doesn't adjust. The fact that the rec player can exploit you by switching to Y is hugely irrelevant if he's too bad to do so. Similarly, playing GTO poker (if you somehow magically find a GTO strategy) against micro players is not optimal and is going to cost you alot of potential profit. Yeah, it's an alright starting point IF you understand what GTO is and why you should adapt and not just blindly do what Snowie tells you. However, with the way Snowie is marketed, I doubt many customers will make this distinction.


--

Now, on to my second point where I said the fact that Snowie's advice can help someone beat cheeseburger stakes doesn't make it a good program. If someone needs help from a training program to beat micros, they probably don't have a very good grasp of the game. Snowie teaches them basics which will likely lead them to play better than they currently do at their limits. Assume we add a new strategy for the agent in the above diagram called C, with payoffs of -1000 regardless of what the rec player chooses. Playing the GTO strategy of B will always be better (since C is dominated), but A is still the best exploitative strategy against a donk playing X. Equivalently, if someone is a complete donk at poker who limps every hand and never folds a pair, Snowie will teach them to play better and maybe this will be enough to beat the micros, but it still doesn't mean its teaching them a very good strategy; it's just teaching them something better than what they were doing before. Try using Snowie anywhere past NL50 and I'm pretty sure it will be losing, so overall the program is very limited.

--

Finally, my point on why I think Snowie is a scam still stands. The marketing is simply false. "play PERFECT game theory optimal poker" is what the site says. Based on my analysis and the analysis of many others, that's extremely far from the truth. I can appreciate that you are clever enough to understand GTO, understand Snowie has flaws, take Snowie's guidance with a grain of salt, and use it to understand spots better while applying your own reasoning and logic to come up with meaningful conclusions. However, I feel many people that buy Snowie will be fooled by the advertising and just blindly follow its advice. As I have already gone over, I think this is a terrible idea.

Looking at the disclaimers you made, it already seems we are on the same page. I just view it more negatively as I think its false advertising that will fool many people that are less knowledgeable of game theory. It essentially is selling itself as something that it is not, and seems alot like a way to extract a quick buck from a few desperate players who think Snowie will be the secret that will turn them from break-even players into sick 10bb/100 mid-stakes crushers.

Any2Suited-

I think we do more or less agree.

---

I think that learning about and understanding poker from a GTO angle is a good way to learn poker and get better. I think Snowie can be a useful tool in doing that.

What if someone folds to post-sized 3-bets 45% of the time and re-raises 30% of the time? Now what? You created a hyperbolic cartoon of a micro player where it is easy-ish to see the exploitative play.

You basically are saying it is better to play exploitatively.

I'm saying that thinking about GTO poker is a way to see where the exploits are.

What is wrong with using Snowie to help understand GTO so I can see where NL25 players are exploitable?

Snowie's advertising is misleading. This is no excuse but whose advertising isn't misleading? Are you going to argue against the virtues of beer because it doesn't guarantee you hot chicks?

when you drink enough beers -- all the chicks become hot, and thus the guarantee again becomes valid

which

why is he folding that much on the river? is it normal?

You get these spots all the time even with 6-max or full ring, you can completely crush Snowie if you have its strategy next to you which basically proves that it's nowhere near GTO because a GTO strategy would be unexploitable even if you knew the whole strategy. Folding 95%+ of your range to river bets on most boards runouts is pretty exploitable.

Since your second clause is correct, it means that if your first clause is correct there has been a lot of wasted words in this thread.

^ @hand above, does snowy want to potbet that river in the first place? you should also have the opponent raise to 36 instead of 40 because snowie is only trained with potbets


I played around a lot with snowie and even under snowies abstractions its pretty clear that its not gto so it seems pointless to discuss that further. The more interesting discussion now is how much value snowie actually has, and to me the answer seems to be a lot. Although for beginners the value will be much less as with copy/pasting snowies play you wont get far esp because it plays pretty bad at the river. But on earlier streets it does a lot of interesting/smart things for sure. It requires a lot of thinking/interpretation and you never know for sure how strong the ranges are in any given spot but if you use snowie as a tool to get new ideas for the way you could divide ranges and to help understand general patterns of the game better i think its a valuable tool. Its a shame they fcked up with the advertising, If they just named it "one of the best poker bots in the world" instead of using the word GTO this thread could have been much more interesting

Does that table mean it folds the straight 97% of the time or just Ts4c? Isn't it possible it calls most of the time with other straights?

that's just bad there's no argument for folding some % and raising some % in this spot. Is he turning a str8 into a bluff then?

It has to mean range? That's bad too though. I dunno.

Snowie clearly makes mistakes. I had a similar situation i think it involved possible quads as well where I did not think snowies fold % was reasonable. The way i checked it was by re-playing the scenario and checking how snowie would come up with opponents ranges. In my case the range was wrong. I.e. snowie assumed a stronger range of opponent that it would have itself.

I think snowie is clearly not perfect and you have to review these cases sometimes manually to verify if snowies advice makes sense or not.

no abusing betsizing, just using .5 - 1x pot.
i saw snowies cbet range, which was extremely weak basically just some gutters and some few TP (but not if it also had BDFD), and it's pretty obvious a PSB will print money.
Sure enough Snowie wants to cbet/fold 99.6%

IMO Snowie has a LOT of sample size problems. The weirdest recommendations I've seen are where someone has played a hand differently than Snowie would have and then Snowie gives absurd advice on a later street.

not that Rant post is about my post, but just wanted to mention that what I posted is how snowie would play snowie (so nothing about sizings or hands snowie wouldn't have played).

They probably don't run enough 3 handed.

I assume 'Hero' in your hand is you. 8)
Maybe Snowie doesn't bet first-to-act on that board 3-handed.

I don't know what part you're not getting, but there is no manipulation here.
3 handed or 6 handed makes no difference, it's BVB with action folded to us.
Snowie does bet first-to-act on that board, and it's range is extremely weak, which was exactly my point. I'm looking specifically at range-vs-range and that should be obvious as I haven't selected any hole-cards.
If anything it's snowie being weak at these stack sizes, it probably doesn't have a sample at these sizes and makes guesses based on what works well deep - but even then I'd have to do some thinking to figure out why it would ever suggest a range like that.
I was under the impression that snowie worked to do well against all of the possible snowie-sized bets (.5x,1x,2x) but now I'm guessing it just works to be balanced at what it thinks is 'optimal sizing' and is possibly very exploitable by just using "sub-optimal sizing" (according to snowie) even if those are .5x,1x,2x bets.
Part of me wants to play around with this more, part of me doesn't want to risk helping the thing that might kill online poker in a few years

3 handed probably does make a difference. It is unlikely that they are using their 6-handed 'learning' runs to inform decisions in 3 handed games.

Non-standard stack sizes make a difference too.

In short - there are multiple things about this hand that make it very likely that Snowie has too small of a sample and is not giving reasonable advice.

can someone give me cliffs? is this snowie stuff legit or what?

Everytime I return to this thread, this line makes me smile

No money in hold'em, everyone's playing GTO because of Snowie. Need to switch to Badugi or Razz.

pokersnowie preflop advisor does not seem to make sense. Yesterday I made charts of every action it would advise me to take preflop. It advises a certain raising range UTG that does not include 56s. Then if you get advice for facing 3 bet UTG it advises to call with 56s. Well thats not really possible because if I followed its origional advice I never would have raised with 56s to begin with.

Thoughts?