Useful frameworks or things to factor

[mjws00]

Poor players are anything but random. A wide range is something we exploit with position and on later streets. At low levels there are tons of holes that should be easy to spot and quantify. I don't think a simple or even complex framework will be able to find the small edges, but we should be able to spot large ones.

I think this could be a brilliant project, and it is definately something I've considered a lot. My thought was to go street by street considering their range, past actions, and bet size. We are creatures of habit, and poor players are especially inept as recognizing how much information they give away.

Basics I'd consider Preflop:
Range on button, Range in bb, Difference in range sb bb, Calling range bb vs (limp, min, 2x, 3x, 4x+), 3bet range (bb) vs (l,min,2x,3x 4x), rrai freq and cards, limp rr freq and cards, exact hands bet/limped (1x,2x,3x 4x+), shove freq and cards (lots show), was bet higher than or lower than avg and cards what other hands were played like this.

On the flop:
We consider all the same betting raising patterns, but we have to add frequency based on hands and draws hit. In addition we need to consider pf action and related patterns. Things like bet size as it relates to semi-bluffs, in relation to made hand strength, betsize as it relates to board texture, cbet frequency and size trends in comparison to hand strength, etc... But there are still basic essentials that could easily be analyzed by a player in concert with common tools.

Obviously as we consider more detail and we must on later streets. It becomes more difficult for a human to keep up, but a HUD could show us a lot and break it out realtime easily. Next Gen tracking like Holdem Manager, PT3, capture and store much of this additional information. Even if we could only spot 10 major leaks or confirm a players subjective 'feel' and analysis of common HUD stats it would be a powerful tool.

A few programs attempt to show holes and fundamental errors in play (chasing bad odds, loose calls etc) but I haven't seen one tailored for HU. It's tough to build a framework to win HUNL autonomously but it isn't tough to give a decent player better information to increase thier edge. For beginners it provides fundamentals and a framework for what must be studied.

Mike

[daveT]

A large portion of playing 5/10/20 SNGs is seeing people call with A-high, middle pair, a draw, a set, two pair, etc. It is mandatory to understand the base-line exploitation against these kind of calling plays. If you play incorrectly, you will force your opponent's mistakes to become profitable.

I don't really know what to expect at the 100s, but I imagine that a good player learns the basics of these plays, and how to best exploit them. The learning curve would seem, to me at least, be dictated by how well you can play versus a broad assumption of tendencies and adjust from there.

By "random" I mean that many players will call pf with ATC, then about 60% of the time, they will call on the flop. Regardless, it is impossible to put these players on a hand.

I think we all agree with the goal, but disagree with on the foundation.

[mjws00]

They are only calling with ATC in response to your current bet size. If it is truly ATC then just shove a strong hand... guess what. They will fold. So it isn't ATC. Same on the flop. Your Cbet just hasn't hit their fold threshold, or they are overvaluing their holding. Both scenarios are exploitable, both allow us to play optimally and put opponents on a range. We can further narrow that range in response to our flop action and we still have two more streets.

Obviously you apply the framework to the population as a whole to determine baselines, and a particular opponents variance from it. In addition you look for errors and suboptimal plays.

You sound like a .01/.02 player that believes it is unbeatable because they 'call with anything'. Hand reading is definately an art, but it is NOT more difficult at these levels. These players seldom mix up their play, they telegraph with their bet sizes, and they are extremely predictable. People find them difficult because they don't adjust and end up over playing hands while behind. In HU our ranges are so much wider this can lead to huge postflop errors. El loose passive (call with any piece) donk makes comparitively small errors (at the cost of value), and simply allows you to hang yourself, especially if you are hyper aggressive and you can't read their hand.

Where is the disagreement on foundation? We are tailoring a solution against a specific set of opponents and building a basic beginners framework. Profiling, and exploiting specific weaknesses are pretty much all there is.

The challenge is that our edge (assuming you can actually read a hand) is increased dramatically on later streets where we can utilize additional information better than our opponents. Unfortunately factoring these scenarios and relating them to prior action, opponents variance from the mean, and our current strength involves considering thousands and thousands of variables which makes it tough to code and difficult to put into a framework a human could manage in realtime. Additionally if our opponent can adjust, or tilts, our framework must allow for that, and there is no statistical clue for a concrete framework to use.

We know how to do this intuitively, it is a matter of quantifying and expressing it. At this point while other forms of poker have standard plays, and fundamentals well documented HUNL is largely left up to 'feel'. Point blank the better players on here could sit down, play ABC, and absolutely crush the low levels. Documenting that play and thought process is the intent.

Mike

[daveT]

Please familiarize yourself with the posters before you write this kind of stuff:

"You sound like a .01/.02 player that believes it is unbeatable because they 'call with anything'"

[ QUOTE ]
They are only calling with ATC in response to your current bet size. If it is truly ATC then just shove a strong hand... guess what. They will fold

[/ QUOTE ]

This is plainly obvious, except in extreme cases. But then you continue with this:

[ QUOTE ]
Your Cbet just hasn't hit their fold threshold

[/ QUOTE ]

And poker is all about bluffing? I have never won a game on a bluff. It is about extracting value.

[ QUOTE ]
These players seldom mix up their play, they telegraph with their bet sizes, and they are extremely predictable

[/ QUOTE ]

This is why people are getting so confused. There is so much talk about "hand reading" when in reality the game is about making decisions.

You agree in one sentence that if you want to play, you must force them to call with bad hands, yet in the next, you want them to fold these bad hands so they can become more predictable.

You can't have it both ways.

If you are going to extract the most moneys from these opponents, you must be willing to force them to make bad calls with random hands, and learn how to extract the most based on these assumptions. There is no hand reading here.

[ QUOTE ]
Where is the disagreement on foundation?

[/ QUOTE ]

Two sentences later, you answer your own question.

[ QUOTE ]
Profiling, and exploiting specific weaknesses are pretty much all there is.

[/ QUOTE ]

And then:

[ QUOTE ]
fundamentals well documented HUNL is largely left up to 'feel'.

[/ QUOTE ]

Do experienced poker players expect a new player to have this ability?

[ QUOTE ]
Point blank the better players on here could sit down, play ABC....

[/ QUOTE ]

And yet, your entire post is based on NOT playing ABC poker.

[mjws00]

I've read your posts, Dave. I said 'sounds like' specifically not 'you are'. SAGE sucks, shove junk early in a match, 'Learn to play RANDOM'. WTF? They play ATC, call flop 60%, so you can't read them is complete and utter BS and counter-productive. We read them all the time that is how we win. I always have a read, I'm sure you do as well. We are not exploiting a 'random hand' we are exploiting a PLAYER.

Any good framework must address both profiling and specific weaknesses. That is not disagreement simply two parts of a complex system.

I haven't discussed how to play or bluff, only how to document and spot trends. That would be how you develop a framework. At somepoint we can apply an analysis for optimal play against, but we don't even have a system or even a list of what we are looking for specifically.

Btw, my intent certainly isn't to attack you. I do happen to disagree with a chunk of what you write, but I believe that is mostly due to the weak medium.

Regards,

Mike

[daveT]

[ QUOTE ]
so you can't read them is complete and utter BS and counter-productive.

[/ QUOTE ]

I don't feel like tossing daggers either, but this statement has been disproved time and time again on this forum.

How many posts have been on this forum dealing with this very problem with the new posters?

Look at the difference between my posts and a newbies.

Me: My read is that he is likely to call with k-high.
or: My read is that he is on a flush draw.

A newbie is: My opponent is passive.

What does that us? Is that all you have at the 25/50 levels?

We can't expect them to do better than this right now. These players at the lower levels are beatable. People are playing 4 tables and walking with a 10% roi, and I can guarantee you that they are playing ABC poker, meaning that they are betting with the best of it, checking with the worst of it, and not bluffing. Your final statement is why I think your post falls apart.

It is for the quoted sentence that SAGE is used and favored by many posters. I feel like I can do better than SAGE, because I have a better sense of my opponent's tendencies, although I may not know what to do with this information at all times. For me, I feel that learning to use this information will be far more valuable than learning SAGE, and my WR dropped by nearly 5% since I started using it.

How do you do at the "x,y" posts? Are you able to think of what the villain is doing without some sort of knowledge about the OP's hand? That the betting makes sense in some sort of fashion?

What do I mean about RANDOM?

Range merging, polarizing, etc. All these concepts are talked about on this forum, and it is a vital piece of play for any player.

There are two different ways to approach this problem. You can either start with your opponent, disregarding what you have in your hand, or you can start with what is in your hand, disregarding your opponent.

While I understand that the base of this OP is to figure out hand reading, I believe it is mandatory to understand how to play your own hand, so that you are able to identify what is wrong with your opponent's play easier.

My reads usually go along the lines of "What is my opponent doing wrong?" Then I ask, "What can my opponent do wrong?"

Mastering those two sentences will get you very far, and that, IMO, is all the read you need at lower levels. But, if you are under the impression that your opponent is making a huge mistake, when in reality, he is playing correctly, then there is a problem.

[daveT]

[ QUOTE ]
shove junk early in a match

[/ QUOTE ]

This accusation is bothering me.

[mjws00]

Where was it proven that HU players are unreadable? I missed that lesson. Every beginner forum is full of newbs that can't make reads. That doesn't mean it is difficult or impossible only that lots of people haven't developed the skill.

The question at hand is how do you quantify that 'feel' and intuition that excellent players use to their advantage. Most of us got there by playing a few 100K hands. But there is no doubt it could be simplified and distilled.

The challenge is there is little incentive for good players to do this. In addition successful players tend to move up so sample sizes down low with winning strategies tend to be small, and considerations in their current higher games are often not applicable to low levels.

I'm not sure if you're saying, "just play your hand" or what. To me ABC does not equate to playing our hand in a vacuum, it is best done in context. Determining what is important in that context would seem to me to be a useful beginnig.

Mike