Exploring Population Underraises on River

Posted by

You’re watching:

Exploring Population Underraises on River

user avatar

Max Lacerda

Essential Pro

Video Player is loading.
Current Time 0:00
Loaded: 0%
Duration -:-
Remaining Time 0:00
  • descriptions off, selected

Resume Video

Start from Beginning

Watch Video

Replay Video

10

You’re watching:

Exploring Population Underraises on River

user avatar

Max Lacerda

POSTED Jul 22, 2021

In his first video for Run it Once, Max Lacerda introduces himself to the community and then delves into the topic of river play and the populations tendency to have far fewer raises than needed to maintain balance.

33 Comments

Loading 33 Comments...

Shaun Pauwels 3 years, 7 months ago

Your MDA shows people not blockbetting river OOP enough as well as IP not raising enough.
You consider IP play to be the thing you focus on, pool playing wrong compared to GTO.
Isn't it so that when OOP blockbets less often that IP raises at a lower frequency? In a way the pools lower IP raise is a correct adjustment to OOP not blockbetting as frequently as they should.

How does SPR affect this? In the example you are in a 3bet pot. What is it like for SB vs BB in a bigger SPR? My assumption would be that there are even more blockbets going on for OOP.

Max Lacerda 3 years, 7 months ago

Hey, thanks for the comment Shaun! It all comes down to how the ranges are constructed. So yeah, it's true that population don't block bet enough OOP OTR, but that information alone doesn't provide us awareness about how weak/strong their bet is when they do blockbet.

From my researches even tho they underblockbet in frequency, they still overbluff when doing so. So in practice pool has even more incentive to raise blockbets OTR

Max Lacerda 3 years, 7 months ago

And about your second question:

Unfortunately we (BrPC) haven't build the reports for river spots on SBvsBB SRP yet, we only have it till the turn, so I won't be able to give you the exact number. But I do agree that it makes a lot of sense, the spot that we usually want to depolarize the most is SBvsBB SRP.

Shaun Pauwels 3 years, 7 months ago

It all comes down to how the ranges are constructed. So yeah, it's true that population don't block bet enough OOP OTR, but that information alone doesn't provide us awareness about how weak/strong their bet is when they do blockbet.

It does to some degree. Let's assume that their value:bluff ratio stays the same no matter which frequency OOP blockbets.
60:20
45:15
30:10

Because the absolute amount of value combo's changes we can assume that the average strength of a value combo is stronger. Due to this IP can not raise as light for value. More hands in IP's range are considered solid value raises vs the 60:20 compared to the 30:10. Even with the right value:bluff ratio IP's raising frequency becomes smaller as OOP's frequency decreases. In your example we'd start seeing K8s to become calls instead of value jams when OOP's blockbet frequency decreases.

I do want to state that this is what I think happens theoretical. It's possible (for pool) the lowered blockbet frequency is still made up off the weaker value hands and the stronger value hands are being used in a larger betsize, keeping the average strength of value the same or even weaker.

EDIT: The more I think about it, the more unsure I am about my logic.

From my researches even tho they underblockbet in frequency, they still overbluff when doing so. So in practice pool has even more incentive to raise blockbets OTR

Ok. That's an MDA I didn't check out yet.

Max Lacerda 3 years, 7 months ago

Alright, so here's 2 sims, 1st GTO and 2nd node lock for half of the betting range. As you said, I didn't change anything about the range construction, I just copied and paste the checking range to change the freq. OOP's line is Bs-X-Bs
1st:
OOP
IP

2nd:
OOP
IP

So, on the GTO sim, IP raises 22.5% vs the small bet, while on the node locked sim, IP raises 21.94%. So as I expected, the raising frequencie won't be very different. The equity of the small size betting range remained pretty much the same (left locked, right GTO with the Bs selected). But the equity of the bigger size decreased, and the equity of the checking range increased. So even tho I tried to make the different ranges for the different actions remain with the same equity, that wasn't possible.

I'd say the conclusion is that the frequency of blockbetting doesn't influence the frequency of the raise response, but it does influence on the sizing, mostly because of blocker effects I guess.

Another thing that it influences is the calling response, and that's pretty interesting. I guess that once again this is blocker effects. So if you take AQ for example, on the GTO sim OOP is blockebetting a lot with those combos, while on the node locked one it's much less. In response you see IP calling much more with AQ on the node locked one.

So yeah, I guess my theory is still correct, maybe the only way we can confirm that 100% is by making some toy games in which blockers have zero effect over results.

notalexwong 3 years, 7 months ago

Nice video. One thing I wished you'd discussed was the population tendencies for the OOP player. I would imagine the reason IP doesn't find the correct raise frequency is because they are just always up against JJ+ when facing a small bet. What would the IP response look like if we node locked the OOP river play?

Thanks,

Max Lacerda 3 years, 7 months ago

Hey! Your doubt is very similar to the question Shaun asked above, there's a misconception of correlating frequency and range construction. It's possible for a player to only raise 1% on certain node, but when they do they are bluffing much more than they should.

So as I said above, even tho the reg population underblockbets, when they do it's still overbluffed, so we have incentive to raise. I don't have this info for every single possible spot in the game tree, but from the most common ones (XC-XX-Bs on SRP OOP as PFC for ex) they still overbluff when betting small.

Doblou07 3 years, 7 months ago

Video brabo! I think this could be your specialty (mixing theory and practice). There's not a lot of this type of content om RIO and you could become a unique coach if you keep bringing up some content mixing GTO and MDA.

Worm 3 years, 7 months ago

Great video and overall concept. Would love to see many more videos in the future following the same theme: here's what GTO does, here's what pool actually does, here's our response to maximally exploit. I'm actually surprised by the lack of this type of content as this 'style' of play is clearly the most +EV when implemented correctly. It's almost as if there's a negative connotation to being labeled an exploitative player by some in the online poker world, and it's trendy to try to be some GTO bot. The reality is, humans play vastly different than solver in many predictable ways and spots and to not adjust is simply leaving money on the table.

Max Lacerda 3 years, 7 months ago

Thank you! Yeah, GTO is nothing more than a set of strategies completely exploiting each other to a point where there's no incentive for any of them to change anything. It makes no sense to play perfectly GTO on an environment where deviations are huge (assuming someone could actually play perfect GTO, which is impossible). But it's important to always keep GTO at the back of your mind, cause at the end of the day, those are still very important concepts and guidelines of how poker works, and without knowing GTO frequencies and trends, there's no way to know if (and how much) someone is deviating from it.

RunItTw1ce 3 years, 5 months ago

From my researches even tho they underblockbet in frequency, they still overbluff when doing so. So in practice pool has even more incentive to raise blockbets OTR

Max Lacerda Where do you see data that pool over bluffs? It's very common for almost every pool micro to high stakes to under bluff. Bluffing frequency is usually 25-30% in general on the river.

I do want to state that this is what I think happens theoretical. It's possible (for pool) the lowered blockbet frequency is still made up off the weaker value hands and the stronger value hands are being used in a larger betsize, keeping the average strength of value the same or even weaker.

Shaun Pauwels This seems correct to me.

I have seen data from H2N that shows across multiple bet sizes the bluffing frequency for ignition pool stays roughly the same from small size 20% pot to big size 150% pot. The smaller size will be closer to 70% value and the bigger size closer to 75% value. The small size is being bluffed more often, but it is not being over bluffed.

Going to watch the video now though LOL. Just wanted to go through all the comments first. I did find the decrease in raise frequency discussion interesting as the betting frequency went from 61% to 47% but less than 1% raise frequency for IP. Looking at the bet sizes, looks like the half pot frequency was cut in half, but the block bet and big size only cut down in frequency by a little bit. That could be why the raise frequency barely changes. Makes a little more sense when we are discussing why the raise vs block frequency barely changes, but it wasn't a huge decrease in the betting frequency either for that size.

Steve Paul 3 years, 5 months ago

If you bet 1/3 pot on the river then 25% bluffs is an overbluff. Opponent is getting 4:1 so only needs to be good 20% to call bluffcatchers.

RunItTw1ce 3 years, 5 months ago

28:40 idk if this was node locked correctly to show 100% range bet on the river. With A7 folding or raising but hand like 55 calling? Also ATs never raising? Wouldn't there he a frequency of nut flush here? Your video makes sense but the node lock seemed weird to me.

RunItTw1ce 3 years, 5 months ago

One last comment for the examples you gave with value betting 2nd pair for 1\3 pot on the river. Based on your Max Lacerda node locks do you still like the 1\3 block bet or would you size up with these 2nd pairs to half pot given the calling range and lack of raising range?

Be the first to add a comment

You must upgrade your account to leave a comment.

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