matlittle

Always enjoy hearing your analysis of these matchups!

Was interesting to see a small raise here:

When I put these parameters into GTOWizard it only has a 0.5% raise frequency. Seems like a spot to just play call/fold and not have a raising range. So my question is whether you think Davy is:
1. Attempting to play a 0.5% frequency small raising range
2. Overestimating the raise frequency/need for a raising range
3. Attempting to achieve something else here with this strategy?

On the KQ99 board near the end you were trying to work out why K9 would cbet turn more often than Q9, despite K9 blocking more of the BB calling range. Whilst it's not massively significant, I did find it interesting. I had a few theories:

K9 blocks some of the BB river bets, whereas Q9 unblocks them. IP will therefore face more bets on the river with Q9 than K9. Important to note that BB only bets the block sizing though with top pair hands.

In your sim IP was checking turn with KK reasonably often. This prevents BB from blocking river for value with KX then having a really profitably 3bet shove vs a raise. By checking back Q9 on the turn you prevent BB from block/3betting river with QX profitably. QQ bets turn ~pure, so Q9 acts as the trap here instead.

On the theme of donking - we saw a number of spots where the BB should have a decent donking range on the turn. That would also affect IP's turn cbetting strategy significantly. I'm sure at the stakes you are playing that most regs will find the turn donks, especially the more common or higher frequency donking spots. But at midstakes and lower it can be difficult to know whether a reg has turn donks or not, so it's difficult to know what turn cbet strategy to use as IP. Do you have any advice on how to deal with this?

At one point in the video you said that you couldn't remove a turn donking range with a turn subtree, and that instead you would have to re-do the whole sim. Could you explain briefly why that was the case please?

If you compare the EVs at the start of the street, the small bet size for BB gives it more EV than the bigger size, hence here the dynamic bet sizing was working fine in this instance.

At 30.00 you were debating whether to use the 40% or the 150% turn bet size. You are right that the dynamic option on GTOWizard doesn't always select the correct sizing, but the way you compared EV's in the video was not the correct way to do it. After adjusting the turn bet sizes, you should click on the SB decision point and compare the EVs there. The reason being that the SB might have a different checking range in the 2 different scenarios and in this case it does. In your sim where the BB bets 150% on the turn, the SB will develop a small donking range. This removes some of the strong hands from its range, hence the BB will have much more EV when checked to compared to the original sim (BB bets 40% on turn, SB no donking range).

At 10.20 it was interesting to see that more 3x bluff raises the turn than 4x. Presumably it's because SB range bet the flop, so the 4x is not more relevant for blocking SB's strong hands (2 pair, sets), than the 3x? At the same time BB is value raising turn with more 3x, so uses more 3x as bluffs to mirror this?

I thought timing tells would be very unlikely from a high stakes reg, but having seen it a few times in some of your videos, I am under the impression that it is still reasonably common?

That's a fair counter-point. I think you are probably right that population is under check-raising the turn. I nodelocked this in to the sim, but BB still prefers the small bet sizing. When checked to, the BB (green) has a slight advantage in the middle of the ranges, so this becomes the main driver of the betting range:

Hands in this area of the graph are mainly 2nd-4th pair hands which require a lot of protection, hence why the BB likes the small bet.

I tried also a 2 sizing turn strategy for BB with small/big/check mix, but it really doesn't seem to add much EV to the overall strategy, mainly because there aren't many hands that are desperate to go into the big sizing even when SB doesn't check raise frequently enough. So I think the answer is that against good regs we want to bet small, against recs and unobservant players we probably just want to bet big with the nut hands and otherwise bet small.

I thought this play was pretty creative and probably generates a decent amount of EV against most regs. I remember that Freenachos stated his data showed that people fast play flushes way more often on monotone boards than a solver would, making them susceptible to river overbets. That was for a different line, but I would assume this phenomenon would also occur in this delayed cbet line too, especially when your delayed cbet is so small.

Was this play vs a reg? Would you consider the same river sizing vs a fish, or do you think that too many of them would snap call a straight and and potentially also be reluctant to put you on a flush, thus making them call too often to make it +EV?

This exact spot came up when I was drilling this spot the other day on GTOWizard. Like you, I assume it would be best as an overbet/check strategy on the turn for IP when checked to (assuming we are playing 1 size here). Interestingly, the solver prefers a small sizing here - the reason being that lots of paired hands are low and need protection. That only changes once the 2nd highest card on the board is 8 or higher, given that SB and BB will both have a decent chunk of offsuit broadway 8X. Of course a block/big bet/check mix would improve EV, but obviously way harder to play. It's not super significant, but hopefully its interesting, and can also apply on other similar boards like K72 vs K82.

Yes GTOWizard has 99 here in the SB shoving range fairly often so TT is a snap call. In earlier, tighter formations I agree with idea that TT should be a 0EV call and that people are probably calling it most/every time, so then AKo doesn't become a particularly attractive shoving hand.

On the 543r board you figured that facing a flop shove from the SB, the BB has a -14BB call with TT. If I run it on GTOWizard AI it comes out with TT being a +55BB call. Even if I nodelock SB shoving a much stronger range (most of the hands you used except for AA which I don't think people are open jamming the flop with) then it still comes out at +11BB to call TT and 99 then becomes the strongest fold.

I’d guess that, because of this, the suits blocking the FD would be used less/never.

I would have guessed the exact same, but my sim has the opposite effect - AKo only delayed cbetting turn with a flush draw blocker. If IP calls the turn, then faces a check on river, they get to bluff their KXss combos and even some of their AXss combos - perhaps this is why?

Based on a few data points my read here is that this was a push from RIO to create more live poker content. Will your videos be returning to online poker content after this or is this more of a long term direction?

At 24.45 when looking at delayed betting the turn 3BP BB vs UTG on 755s2s we saw that AK bet sometimes but AT was never betting. Aside from having more outs against pocket pairs, do you think that AK sometimes gets value here from flush draws and is part of the reason it bets here as a pseudo-value bet? Perhaps it folds the occasional better hand too (low pocket pairs)?

Unfortunately this bet sizing bug you got in the video is quite common on GTOWizard AI. I find that in a bunch of spots I'm able to pick a better bet size than AI has picked, and after re-running the sim with my bet size the EV is higher. Hopefully they fix this in the future. Other than that, the site is very impressive (although as I type this the site is currently completely down for some reason).

Yeh its a pretty unnatural way to look at EV on each street, way more natural to just consider your EV from that decision point, but for some reason this software has decided to subtract what you invested on previous streets from your EV

At 48.26 you faced an MP 3x open and a BTN flat.

The GTOWizard ranges (NL500) for a MP 2x open and BTN flat show the BB pure folding A5o, so I'm curious as to how your ranges can be so different? Are they monker ranges, and are they for a raise size less than 3bb?

FYI the EV numbers seem to add up to 0 in all the sims, so the EV at each node is the overall EV for the entire hand, hence why some players had -EV numbers by them at certain nodes: they had already invested money into the pot on earlier streets and now caught an unfavourable runout for example.

Hello Frankie, thanks for the follow up on multi-way pots, these sims are always very eye opening and in no way resemble what I see in-game when playing multiway pots. You mentioned briefly that in a 3way pot BB/CO/BTN the CO player checks very often. I am guessing that the CO's cbet strategy is somewhat similar to a CO/BTN 2-way SRP, albeit with a reduced cbet frequency due to the added player?

Similarly, the BB checked at a huge frequency on the turn. I am assuming this should be the case on most boards due to the BB having the weakest preflop range and usually facing a stronger, mostly still uncapped range? Are there any board types you know of where the BB can be more aggressive? Is the block sizing the most common sizing for BB turn probe, even on boards that play polar?

I tried to run the AK hand on J64dd2Kd where you donked the river and called the shove on GTOWizard AI. Like you pointed out on the turn, the IP sizing should be bigger than block, but if I forced IP to have only a block sizing, then you don't get any river leads on the Kd anymore. If you look at the bluff shoving range for IP vs the river block, it contains a lot of pretty wild bluffs like T9ss/hh, 98ss/hh, 87ss/hh at some frequency - I don't think the vast majority of regs at most stakes (especially 100nl) will find these.

Or "they're all pretty bad, this ones egregious".
To answer why the play is so bad, the line-ups are basically recs, plus old school live regs that still use the same strategy they used 20 years ago. They can get away with this strategy because of the softness of the games they can play in based on their fame value.
There is also some incentive for the 'regs' to play a little looser preflop - someone playing 24/20/10 in these games would be probably be considered a 'nit' and unless they offer some other reason for viewers to watch would unlikely be invited back to these TV stream games. Some shows will track VPIP of players, shame the tighter players and praise the looser ones in order to generate more action. Some, I believe, even stipulate a minimum VPIP for regs to play in order to be asked back on the show. A large variable in live streams rather than other poker games is essentially "entertainment" value.

"It feels like an old school move from 2005" kinda sums up the play in these games.

Seemed like Yakobishvili was giving off constant timing tells. Sure, they are not 100% indicative of hand strength, but there were plenty of spots where both you and I were able to narrow their range down and remove/reduce combos of certain hand classes from their range based on how quickly they acted on that street or a previous street. Did you feel that during the match T Duthweiler was making exploits based on this? I saw one river overbet from TD after GY snap checked back the turn, and I thought that was potentially a good way to bluff and counter this, so perhaps that was a bluff but we didn't get to see it.

Most people will end up probing way too many bluffs here - any 8 out straight draw, a bunch of gutters, more of the combodraws and SDV flush draws etc. Nodelocking this gives a much increased raise % for the IP player facing this at the main exploit.

Hello Luke, glad to see you are back with some theory videos. At 44.00 you were a bit puzzled as to why the J8hh would not probe the turn. The answer is that you just don't have enough bluff allowance here to bluff every 8 out straight draw. In most scenarios you would, but here there a few things happening:
1. The large probe sizing means you are probing at a low frequency, and hence don't have a huge bluff allowance.
2. The board is 2-tone on the turn, so you now have twice as many flush draws as on a board where there is just one flush draw present. Flush draws obviously are better draws than straight draws so are higher in the probing bluff hierarchy (aside from the ones with showdown value or some combodraws as you mentioned in video).
3. This board is just about the most straight-draw heavy board possible.
4. Higher straight draws get probed less frequently here than lower ones, as the higher ones block more of the folding range.

Hello Callum, thanks for another great video!

In this hand you correctly noted that your opponent should be using a bigger sizing here. I see this a lot in the games I play - regs snap betting for the small sizing on these boards in 3bet pots (often a weaker reg). I assume some/most are doing this as a range bet but it's of course impossible to know for sure. What is your advice for countering such strategies?

Also, on turn you do get to have a reasonable shoving range here, AQ with Qh seems a high frequency candidate because, as you mentioned, the Qh should unblock villain's bluffs/high equity draws as villain's Qh bluffs should give up turn more often than other combos.

At ~9.50 we saw T Duthweiler min cbet A82ccc. You mentioned at the time that you thought this sizing down meant he was cbetting range, then we saw him Q9hh hand and it further supported that idea given that this is one of the prime check-back candidates on the flop. I have seen this increasingly often from regs, especially those playing higher stakes. From my sim, with a min cbet sizing, the solver is only betting 35% frequency on this board. It seems a pretty wild stretch to go from 35% to 100%. IP EV drops from 2.64 to 2.58. So I'm quite confused as to why I'm see this type of strategy so often. Does the player pool respond to this bet size so poorly in a bunch of spots that it makes for a very easy way to exploit player tendencies? Even if I nodelock in BB folding more and raising less than they should, I still only get a cbet frequency of around 55%.

Congrats on hitting the jackpot! Funny that you knew exactly what the player had. I lost with AAATT yesterday and was sat there trying to remember whether the minimum hand was AAATT or AAAJJ. Also was trying to work out whether other criteria were met - both players using both hole cards etc. Nothing appearing on screen. Then bang, $5700 in my account. I got slowrolled by the GG server lag which caused the animation to be delayed! Certainly added to the suspense, would recommend! Do you have any idea how often a typical reg will hit these jackpots? They are of course extremely rare but do in fact account for a huge amount in terms of "rakeback".