SnowAndFire 's avatar


341 points

Comment | SnowAndFire commented on Where to begin ??

Book: Mathematics of Poker
Essential Subscription to RunItOnce: Pro Videos (From The Ground Up)

March 26, 2019 | 12:10 a.m.

Comment | SnowAndFire commented on x

Pre-flop - Looks good.

Flop - Looks good. Villain is polarizing his range here meaning his value range should have lots of equity and the rest will likely be low equity - take note.

Turn - Villains give-up range should be very low equity which means our range at this point has a ton of equity. When our range has a ton of equity we want to be using small sizes to extract EV. If you are going to bet here the proper size is likely 25-33%. That being said, this hand, after villain polarized flop, likely favors a check-back. Bet 25%-33% or check-back are both fine.

River - As played, after you bet large OTT, shove is good.

March 24, 2019 | 5:47 p.m.

The only likely mistake you made during this hand was $0.40 OTF. Generally multi-way spots favor very small c-bet sizes. That said, 66% was okay, not bad. Otherwise, you can't do much. You played it well.

March 24, 2019 | 5:35 p.m.

SnowAndFire I have several comments here. I use GTO+, and not Pio, and I'm very new to the software, so I may make some errors.

No worries. They should all end up at the same place if we use the same abstractions.

First, your comment about the solver preferring a (larger than geometric) overbet by Villain on turn is completely correct. When I give the solver the option to use bet sizes like 130% or 110%, it prefers that size with almost all its betting hands. The actual size chosen by villain (66%) is almost never chosen by the solver.

In my solution I did not give these sizes as I was just simulating what happened in-game.

Geometric overbets should be used when one player has a nut advantage and the ranges stay static from one street to the next. In this case no draws were completed so there is no huge shift in equity there. The one thing that was added was another flush draw which increases OOP's EV slightly.

However, if we force Villain to only use 66% pot bet size on the turn, then in my sim, T8hh is almost entirely a call for Hero, with a very tiny fraction of raises mixed in (might be due to some sort of convergence issues?).

Likely not convergence issues as I am seeing a check/raise at about 20% frequency as well, as was stated above. When 2 options are the same EV they are 'mixed'. Against a GTO player it does not matter which option you take. Exploitatively you want to take the option that is more likely to cause a mistake by your opponent. Exploitative strategies require no mixing - they can be pure strategies everywhere.

I'd like to note that our ranges and trees are built slightly different which can shift strategies.

Overall, we can see that calling or raising are both fine options in a vacuum. If we have info on villain then we can lean towards one strategy or the other to exploit them.

March 18, 2019 | 12:39 a.m.

I know it seems counter-intuitive but I have always had the belief that rake should be paid for inverse to a players VPIP.

ie If you are really loose, and therefore give a lot of action, you should be paying less rake than the person who is playing a tight VPIP. This incentivizes better games and therefore should help the economy and poker sites grow.

It would be cool to see a site implement a system which each player contributes rake in this way. Not only do I think it would be better for the games but since there will always be a changing rake structure in multi-way pots, it will be much harder to "solve".

Imagine you are playing 6max NLHE and someone opens the CO and you are in the SB with a low-medium strength hand that would just be folded if the rake was high. Now imagine that your rake is paid relative to your contribution to the game (inverse to VPIP). In this scenario you'd be more than likely willing to VPIP this hand because you know that folding it causes your rake to go up.

Now imagine the player in the BB has been tight for the past 300 hands and you know that his money will pay for the rake more often than your money. This would incentivize you to flat more often to get him to enter the pot. As your VPIP in the SB goes up he now becomes more incentivized to play hands out of the BB because your range is wider and therefore his hands become more profitable.

Not sure how a system like this could be implemented but I am sure someone smart could come up with a way to structure it.

March 17, 2019 | 11:23 p.m.

First I am going to say a couple things.. Multi-way spots are not solved and we should be careful about trying to take the information from a HU solver and thinking that everything applies - this is not always the case. Also, this sim was quickly thrown together. The level of abstraction may be high and there may be convergence issues. Though I do believe it is close enough for what we want.

So here you can see that the EV of T8hh call and raise are very close when solved. It shows that the EV of raise is slightly higher but if we were able to solve to 0 exploitability I would be willing to bet they would be the same. Therefore, against a GTO opponent it does not matter which option we take. Against a nemesis opponent (one who will always counter us perfectly) we should be mixing to remain unexploitable. Against a sub-optimal opponent who does not adjust quickly the best option would be to take the one which causes them to make the most mistakes.

If the opponent cbets too often, which is likely, we should be reducing the EV of his range by shoving thinner value hands as well as hands such as this at 100% frequency because he will fold too often which will increase the EV of hands such as this. That will make this a pure shove rather than mixed.

If opponent cbets not enough then we should be calling 100% (most likely) as to not be handing his range, which is value heavy, more EV.

Let´s be realistic tho.
How often will a Micro Player bet turn with his draws and how often with value hands. I´d say he checks his draws at a damn high frequency and his bet means 85% of the time just a strong jack, overpairs and sets.

This is low stakes but not the micros. Anyways, we dont have much information on this player so we can't say he bets too much or too little without reads. The only information I see is that he opened for 4x and that he chose 66% OTF in a multi-way spot. We know that these sizes are sub-optimal and too large (from having solved the game enough). Since he is using sizes which are too large I believe it would be safe to assume that he also bets to frequently to be using these sizes. Other than that, it is tough to definitively lean one way or the other.

Where did you get this 85% number and the range? Are these something off the top of your head?

Just because calling is +EV does not mean you should always be taking that line. Other options can also be +EV but have higher EV.

So I prefer the call for sure since we also get good odds and then just lead rivers if you hit. Shove if you hit a straight with no flush possibilities and block bet flushes.
Thats what I´d do and I think its most profitable way of playing vs these players.

This is very exploitatitve and I don't want to use the wrong words but I would normally call this range construction fishy. When you deviate this far from equilibrium you open yourself to a world of counter-exploitation that many 100NL players this day in age are able to act on.

March 17, 2019 | 10:18 p.m.

When the villain bets the turn I would venture to guess that he should be overbetting (larger than geometric) due to the wet nature of the board. Though it could be that a smaller size such as this is correct being that hero's flatting range from the sb should be quite condensed. Either way I believe that villains turn betting range should be quite polarized (low freq/high hand strength) as he will end up cbet/folding too often or cbet/calling too light if he bets with all his draws and thin value hands.

In these scenarios we tend to see that our turn bluff-shoving range should be hands with very robust equity when called and not as much worried about blocker effects as you would think. My first choice for bluff-shoving hands would be hands like [A5hh, A4hh] as these hands tend to dominate villains draws that call the shove and have very robust equity versus villains value-calling-range. Combo draws are fine to bluff shove with as they have robust equity as well. I will run a PioSOLVER sim to see what it favors. Calling may be the best option for hero's specific hand, though being that people c-bet this type of turn too often I would guess that we should widen our turn raising range to punish villains EV and hands like this would be the first to go in there as a bluff (if it is not already).

March 17, 2019 | 9:52 p.m.

Preflop - Fold.

Flop - Check/Raise or call are both fine.

Turn - Looks fine.

March 17, 2019 | 9:18 p.m.

Probably some type of server request error message for spamming their servers. Do you open videos then move on immediately? If so, try watching the whole video then moving on. Never had this issue myself.

March 17, 2019 | 9:16 p.m.

Block bet flop for $16. Your range has a ton of equity and you should be using block bets to realize equity with weaker hands. It will also make opponent have a much more difficult strategy to play against you as he should be raising you frequently when you use small sized bets. After block betting flop and getting called villain is likely to have much more Tx in his range relative to the amount of Tx in your range. Therefore, Check and Call is the play. This hand blocks his most frequent Tx combos [JT] so I would not be insanely shocked to see a solver use this hand [JJ] as a merged jam on the turn. Though call is for sure not going to be a mistake so I would always take that line.

As played with large size on flop and check on turn, call is the only option. You are getting almost immediate odds to continue witha set if he has a straight. Many times he does not have a straight and therefore you should be able to immediately call profitably.

March 17, 2019 | 9:12 p.m.

Comment | SnowAndFire commented on x

I own many copies of Pio Edge and Jesolver, have used GTO+, Monker, Simple Holdem, etc for several years. For a player trying to learn quickly I highly recommend Snowie. It has a couple issues with its level of abstraction but for the most part is insanely strong and will get you much better at a much faster rate then having to build and interpret solutions in these other solvers which do not allow you to play against the solutions.

March 17, 2019 | 9:01 p.m.

With uncapped or very high cap rake you will see optimal play looking to 3-bet or fold most of your continuing range in most spots. That said, if you feel a player is weak and will make mistakes in SRP's then the EV of flatting increases.

March 17, 2019 | 6:57 p.m.

Just because a play is +EV or auto-profits from betting does not mean you should.

The strategic option of checking-back is always >0EV as any hand will be able to realize some sort of equity versus a portion of your opponent's range.

Also, if you c-bet at 100% frequency it is very easy for your opponent to counter this strategy by raising you liberally. When this happens you will be forced to fold a lot of hands that could have gained EV by checking back. Not only this but your opponent will be effectively reducing the EV of your range because you will have put money in the pot and then folded.

With every play in poker there are many moving parts and the fact that you are taking the time to ask these questions will help you 'put the pieces of the puzzle together' in the future.

Keep it up!

March 17, 2019 | 6:53 p.m.

Yo! Thanks guys. GL to everyone this year

Feb. 4, 2019 | 3:16 a.m.

More value in betting small. If you node-lock in a solver for players not check-raising often enough it will always stab. Denying equity is much more valuable than villain betting after you check-back.

Jan. 26, 2019 | 12:06 a.m.

Comment | SnowAndFire commented on 50z Folding a set

Way too tight of a fold. Get it in here every single time.

Jan. 25, 2019 | 2:29 a.m.

I have studied with Pio for quite a long time and when I did a small amount of play on Snowie it showed no errors so I would assume that the strategy it creates is pretty strong. I would bet it is a great tool for learning a good strategy quickly. As you move up or want to create exploits I would say move to a real equilibrium solver.

Jan. 24, 2019 | 1:14 a.m.

I will def have to do some pre-flop work to see how limping strategies can affect players strategies around them. Have you done this work yourself? If so, what program did you do it in? SPH?

Also, do you think there is quite a bit of value to be had by implementing limping strategies before many other players do that work? Seems that is one area that is under-studied...

Jan. 23, 2019 | 4:48 p.m.

You need to look at a few different spots to try to understand why this is happening. While posters above are probably correct there may be other reasons it happens in different spots. It depends on range construction.

Start by going back to the Bet 60 node and look at villains strategy. What combos does 8d affect? Now go to the river and look at the EV of this combo on different runouts.

Easiest way to do this is to use hotkeys that are built into piosolver for changing runouts (rank up = Ctrl+Up, etc). This can also be done for previous streets.

Another useful tool is hotness tool (ctrl+H).

Jan. 23, 2019 | 4:38 p.m.

Yes. And not only should you read it once, but multiple times. This is pretty much the bible of poker. Also, read it before you use solvers. There are things you will learn in MoP that will help you use solvers.

Jan. 23, 2019 | 3:59 a.m.

Comment | SnowAndFire commented on 4BP B-X-B Bluff

Pre-flop - 4-betting this hand is too wide if you are opening to any size, especially for 3x. This is a fold.

River - You are repping a thin value range and with 4-betting this hand pre-flop I could see you getting out of line (having way too many bluffs) quite frequently. If that is the case then he may call you wide.

Note - We know it is Global Poker. You could just post the Share Hand link ;)

Jan. 23, 2019 | 3:38 a.m.

Flop won't get raised often enough to deincentivize small cbets. 25%-33% OTF. Then OTT I would (exploitatively) use 33% again to force him into tough spot with many hands. I would use this size with any hands that bet as well. again, I don't think they will attack these bets often enough to merit choosing any other size.

As played OTF, turn I would use 33% pot. He has already shown weakness and is unlikely to have much equity here. You can bet small to get value from very small hands.

As played OTT, call river. I think you see a lot of nutted hands with this line but I also believe that players spazz often enough at this level to call here. Ad helps the decision.

Jan. 22, 2019 | 11:52 p.m.

23:45 - Table 2 - Thoughts on geometric block-betting turn for 25% pot?

My thoughts are:

-You block the biggest portion of his Kx range (KQs). I imagine players would get in AK quite a bit pre-flop versus you from these positions. So QQ feels good enough to block bet as an equity-denial value-bet with quite a bit of hands in your range.
-The turn doesn't shift equity much so it feels like you should be growing the pot geometrically with your betting range. Being that the board is dry and your range has a lot of equity some type of small geometric sizing feels right.

30:05 - Table 3 - Your range is extremely capped here allowing both players to play almost perfectly against you. Do you think that there is often enough folds that this play is higher value than flatting OOP?

It feels like JJ in a BTNvsBB spot where BTN opens JJ for 2.5x and BB 3bets (for the optimal sizing) making JJ indifferent between shoving and 4betting small. In this scenario JJ doesn't really seem to care about balance but rather that there is enough dead money to incentivize getting it in now. Does this feel the same to you?

33:14 - Person Cam - Shot of whiskey mid-session? Sometimes my kids make my take mid-session shots of whiskey as well.. def exploitative! XD

Good video.. as always!

Jan. 22, 2019 | 11:39 p.m.

Nice! I have been watching and re-watching your videos so much lately as your style feels so close to mine. I haven't even watched this yet but I can say how excited I am! Whoop whoop!

Jan. 22, 2019 | 11:08 p.m.

No.. some hands will do very well in 3bet pots.. filter for each hand in each position to find out what each hand is worth and where you need to fold more or make post-flop adjustments. But remember if you filter for 'Call vs 3bet' then you are already starting the pot with money involved. so breakeven after opening to 2.5bb is -250bb/100. Therefore if a hand is -100bb/100 and you always open 2.5bb from that position then it is likely printing money to call the 3bet.

Jan. 22, 2019 | 8:52 p.m.

Don't think too much about "what it looks like" when picking a sizing. Just pick the size that you would with value hands. This will end up looking to your opponents like you have value. You may be over-thinking it which is causing you to make mistakes. If they are over-folding to 3bets then make this squeezing range one pip wider...

Squeezing range vs utg1 and fish in CO would be something like [77+, AQo+, AJs+, KJs+, QJs, JTs]. No suited connectors. From SB you should be squeezing rather linear. Being shallow makes this much more the case.

If they fold too much then I would squeeze [77+, AJo, ATs+, KTs, QJs, JTs]. You dont want to get out of line when exploiting players as counter-strategies are more costly than initial exploits make you.

Jan. 22, 2019 | 7:54 p.m.

Main villain UTG+1 plays a tag style, but over cbets even multiway.

One of the most profitable mental notes you can make in poker.

Villain two in the CO was the player I was targeting in this hand playing over 80% of his hands.

Is UTG+1 c-betting too often because players are too passive? One thing you should consider as it may build onto the meta-game later on in a session. eg if you start attacking his c-betting too often and he wants to counter you

Effective stacks hero $435 in this hand.

Likely not deep enough to be flatting 87s out of the SB vs a 5x raise. You are OOP and versus an aggressive reg. The fish makes it better but versus fish you want to be 1.) IP 2.) Playing big cards as these play better versus ATC than suited connectors

*His bet size is a little on the small size here and I find this is a spot I could check raise as he would likely to bet $40-$60 with with his made hands. I was ranging him at this time on some draw or middle pocket pair.

Be careful about putting players on absolute ranges due to your reads. It is better to say "I think it is slightly more likely he has draws in his range due to this read." but do not be absolute about it. Just use a heavier weight.

Flop ($80) Qc4d5d

Your range has some very strong hands here. If you want to continue the best option is likely raise flop and shove club or 6 turns. Players will have a hard time dealing with a flop x/r. Most will over-fold if you make it $125.

Turn ($150) Qc4d5d2c
Hero ($375) checks, UTG+1 bets $70, Hero raises to $220 with $155 behind.

Shove here. Do not call. Double-suited board really wants you to over-bet. You have sets and straights here so there is no reason this hand should be a small raise. Besides the small raise making no sense it gives him good odds to call.

River ($590) Qc 4d 5d 2c 2s
Hero all in $155, UTg+1 ?

River he has to call a ton because his odds are so good. Should have shoved the turn.

Overall I would say that you are playing too passive and fishy. Tighten up your ranges OOP and be more aggressive.

Jan. 22, 2019 | 7:03 p.m.

When filtering for 'Call vs 3bet' you have to remember that you already have committed money to the pot in the form of a raise. So, if you are opening to 2.5bb then you are already -250bb/100 if you fold. Therefore the 'Call vs 3bet' must only be better than -250bb/100 to be profitable. It is a little counter-intuitive if you did not already know this but I am sure makes sense if you think about it.

It has been awhile since I have used a database but I believe they still work this way. You must do the calculations.

When constructing pre-flop ranges you should look at some solutions of what optimal looks like and figure out why they look the way they do then take to your database and see which types of players certain hands are profitable against. Then construct pre-flop ranges versus each player type. Hint: In poker the hands which stop becoming profitable to call are normally your first bluff raises ie polarization.

Jan. 22, 2019 | 6:36 p.m.

Yeah. Either are fine. But dont expect it to be any easier to play once he calls.

Jan. 22, 2019 | 3:59 p.m.

Load more uses cookies to give you the best experience. Learn more about our Cookie Policy