Get Better at Poker Overnight

S

stil370

Visionary
Joined
Mar 19, 2010
Total posts
674
Awards
1
Chips
0
I want to share two FREE tools that I have recently come across that will no doubt make you a better poker player. In addition, unlike some other tools people use to get better at Poker, they are far from boring or tedious. So without further ado, they are:

1. Poker Alfie
This is an Artificial Intelligence 6max ring game. There are 3 settings, so it is good for beginners up to intermediate players. The competition ( I set it at the highest level ) will surprise you. I couldn't believe it the first couple times around because usually it doesnt take me long to figure out how to beat these pre-programmed opponents and ulitimately I walk way not having learned much. If anything after playing poker alfie, you will become disciplined, if not, you will end up losing. It is available on google play.

2. OBS Project
Some of you are probably already familiar with this tool, but Im pretty sure a great majority of forum members are not. OBS allows you to tape all of your sessions from beginning to end. In addition as you play, there is a audio component that allows you to talk about your decision making as you go along.
When you go back and watch the session and listen to what you were thinking, it will no doubt make you think twice about whether you made the right play or not. This also completely free. Google OBS Project and it will appear immediately.

Hope you all find these tools as useful as I did.
 
NWPatriot

NWPatriot

Rock Star
Joined
Jun 18, 2018
Total posts
480
Chips
0
Thanks for the Poker Alfie tip.

There was a time when there was a similar free poker AI called PokerNeo (I think). It had a few more options for play, but their playable simulator is no longer available.

A good AI can be very helpful to play against. We can test out how sound our strategy really is. We can try out specific strategy changes to see how they work (or don't work).

I am not sure how this AI ranks among the myriad of AI's created in the last 5-10 yrs - this would be a very interesting Cardschat research project.

It would also be quite interesting to know how some of the Cardschat members might fair against this AI. I would be especially interested in having some of the Cardschat experts play a 5000 hand sample and see what their bb/100 rate might be. This would be a fantastic yardstick for the Cardschat community to compare to.

You in?
 
S

stil370

Visionary
Joined
Mar 19, 2010
Total posts
674
Awards
1
Chips
0
Thanks for the Poker Alfie tip.

There was a time when there was a similar free poker AI called PokerNeo (I think). It had a few more options for play, but their playable simulator is no longer available.

A good AI can be very helpful to play against. We can test out how sound our strategy really is. We can try out specific strategy changes to see how they work (or don't work).

I am not sure how this AI ranks among the myriad of AI's created in the last 5-10 yrs - this would be a very interesting Cardschat research project.

It would also be quite interesting to know how some of the Cardschat members might fair against this AI. I would be especially interested in having some of the Cardschat experts play a 5000 hand sample and see what their bb/100 rate might be. This would be a fantastic yardstick for the Cardschat community to compare to.

You in?


Im definitely in.
I think it would be a great idea.
 
venycyos

venycyos

Visionary
Joined
Jan 6, 2017
Total posts
958
Awards
2
Chips
1
Thanks for sharing, I will test both tools.
 
S

stil370

Visionary
Joined
Mar 19, 2010
Total posts
674
Awards
1
Chips
0
Thanks for sharing, I will test both tools.

Great venycyos,
I hope you benefit from one or even both.
if you are tech savvy the OBS is pretty easy. If you are like me (not tech savvy), then you probably will have to look at a few youtube videos-but still pretty easy.
 
NWPatriot

NWPatriot

Rock Star
Joined
Jun 18, 2018
Total posts
480
Chips
0
Im definitely in.
I think it would be a great idea.


OK, so i have done 1000 hands and i have done fairly well. I was doing 100 hands per session, and I had a couple horrible sessions and thought it was really going to kick my butt.

After 1000 hands I am at +0.23bb per hand or +22.6bb/100 hands. I think this is fairly solid results, but again not knowing how others have really done, I don't know how good it is. I plan to put in 5000 hands here. Jonathan Little and another pro faired far worse , but that may very well be because of the learning that the App is accomplishing and I am doomed to have some more negative sessions - I have no idea.

I am also very interested in knowing how this compares with other AI, and even some of the pros here at Cardschat. This would just tell us whether this is a good tool or not.

Good luck and God bless.
 
ventrolloquist

ventrolloquist

Visionary
Joined
Mar 20, 2019
Total posts
647
Chips
0
Thanks for the Poker Alfie tip.

There was a time when there was a similar free poker AI called PokerNeo (I think). It had a few more options for play, but their playable simulator is no longer available.

A good AI can be very helpful to play against. We can test out how sound our strategy really is. We can try out specific strategy changes to see how they work (or don't work).

I am not sure how this AI ranks among the myriad of AI's created in the last 5-10 yrs - this would be a very interesting Cardschat research project.

It would also be quite interesting to know how some of the Cardschat members might fair against this AI. I would be especially interested in having some of the Cardschat experts play a 5000 hand sample and see what their bb/100 rate might be. This would be a fantastic yardstick for the Cardschat community to compare to.

You in?
I'm actually in the process of testing it. 2000 hands in so far and winning at 50bb/100. There were many small prior sessions where I hit high winrates too. This is all a small sample though.

Unfortunately I have to play at the highest playspeed on my phone or else it takes a minute to make each decision. And the high playspeed sacrifices the strength of play. But I'm also testing it on pc in parallel using the strongest level.

So far I see it do wierd stuff like 3bet quite small (mixed with typical 3 bet sizes) and occasionally open limp. It also calls down very light making it easy to stack it by valuebetting nutted hands. Donks very often from the blinds which I still can't understand. It also has a very aggressive postflop 3bet strategy with thin value types of hands, relying on getting you to fold, this is mixed in with 3betting nutted hands.

Other than that the lines it takes are fairly logical and its a lot better than many other AI's, definitely not snowy level, I think there was another AI I tested called pokergenius which I would consider superior but I didn't play a big sample with that one.

Overall it's still a good practice tool to get in the groove before playing a real game and despite it's quirky plays I find myself getting higher winrates at micros after doing practice runs with Alfie. I use it as a tool to help me see if my mindset is in the right place before jumping into real games.

It's too early to tell how good it is but I'm gonna come back here periodically with progress updates on my winrate.

Disclaimer: I wouldn't consider myself as skilled as some other players on cardschat but I know a decent bit.
 
Last edited:
NWPatriot

NWPatriot

Rock Star
Joined
Jun 18, 2018
Total posts
480
Chips
0
....Unfortunately I have to play at the highest playspeed on my phone or else it takes a minute to make each decision. And the high playspeed sacrifices the strength of play. But I'm also testing it on pc in parallel using the strongest level.


I am definitely only trying to beat the best it has to give: slowest speed and optimal playing style. Even on a PC, i have to wait sometimes for it to make its decision, so I assume it is running some algorithms. I really want to know if I can beat its best. I am less interested in beating its "easier" levels.

I agree that it's play looks a little quirky at times. I assume that this is just due to some GTO type of solution where it plays a very mixed strategy to prevent us from predicting its moves. Because it is called AI, i have to assume it is doing some "learning" on us. Otherwise it is just an algorithm/program and not really AI. I don't know how long it takes (hand count) to pin us down, but it may still be "testing" us. Maybe i am giving it too much credit. I do intend to push through to the 5000 hand threshold.

Good luck and God bless.
 
ventrolloquist

ventrolloquist

Visionary
Joined
Mar 20, 2019
Total posts
647
Chips
0
I am definitely only trying to beat the best it has to give: slowest speed and optimal playing style. Even on a PC, i have to wait sometimes for it to make its decision, so I assume it is running some algorithms. I really want to know if I can beat its best. I am less interested in beating its "easier" levels.

I agree that it's play looks a little quirky at times. I assume that this is just due to some GTO type of solution where it plays a very mixed strategy to prevent us from predicting its moves. Because it is called AI, i have to assume it is doing some "learning" on us. Otherwise it is just an algorithm/program and not really AI. I don't know how long it takes (hand count) to pin us down, but it may still be "testing" us. Maybe i am giving it too much credit. I do intend to push through to the 5000 hand threshold.

Good luck and God bless.
Okay, I will keep my winrate reports to optimal slow speed games only then.

It doesn't learn, I asked the guy who created it once.

Has yours ever open limped on maximum playing strength? That's non-existent in nearly any GTO strategy as far as I know. (small blind flats are fine, I'm talking about open calling from the non blind positions)
 
S

stil370

Visionary
Joined
Mar 19, 2010
Total posts
674
Awards
1
Chips
0
I'm actually in the process of testing it. 2000 hands in so far and winning at 50bb/100. There were many small prior sessions where I hit high winrates too. This is all a small sample though.

Unfortunately I have to play at the highest playspeed on my phone or else it takes a minute to make each decision. And the high playspeed sacrifices the strength of play. But I'm also testing it on pc in parallel using the strongest level.

So far I see it do wierd stuff like 3bet quite small (mixed with typical 3 bet sizes) and occasionally open limp. It also calls down very light making it easy to stack it by valuebetting nutted hands. Donks very often from the blinds which I still can't understand. It also has a very aggressive postflop 3bet strategy with thin value types of hands, relying on getting you to fold, this is mixed in with 3betting nutted hands.

Other than that the lines it takes are fairly logical and its a lot better than many other AI's, definitely not snowy level, I think there was another AI I tested called pokergenius which I would consider superior but I didn't play a big sample with that one.

Overall it's still a good practice tool to get in the groove before playing a real game and despite it's quirky plays I find myself getting higher winrates at micros after doing practice runs with Alfie. I use it as a tool to help me see if my mindset is in the right place before jumping into real games.

It's too early to tell how good it is but I'm gonna come back here periodically with progress updates on my winrate.

Disclaimer: I wouldn't consider myself as skilled as some other players on cardschat but I know a decent bit.

Im glad you gave it a try ventrolloquist. I've played a similar amount of hands on it and I experienced many of the things that you did. I wrote the original quote after I had lost $100 in 2 consecutive sessions. For me at first, it seemed I would get 3 bet at a very high rate. So for example if raised it 3x vs a raise of $9-my re-raise would be $27 and then the AI would 4 bet to $81. All of a sudden Im all in. I had KK and of course the AI was holding Aces. Session over. That was sort of how my 1st couple sessions went..

When I first started playing against poker alfie I thought I would win easily as I do with most of the AI's, but that it didnt happen, that is what prompted me to write a little review here on it. I'm sure beginners could learn alot about the game of poker if they played against Alfie.

Plus the fact that it was free. Its definitely a good warm up tool. Putting you in the proper mindset for a real money session.
 
S

stil370

Visionary
Joined
Mar 19, 2010
Total posts
674
Awards
1
Chips
0
I am definitely only trying to beat the best it has to give: slowest speed and optimal playing style. Even on a PC, i have to wait sometimes for it to make its decision, so I assume it is running some algorithms. I really want to know if I can beat its best. I am less interested in beating its "easier" levels.

I agree that it's play looks a little quirky at times. I assume that this is just due to some GTO type of solution where it plays a very mixed strategy to prevent us from predicting its moves. Because it is called AI, i have to assume it is doing some "learning" on us. Otherwise it is just an algorithm/program and not really AI. I don't know how long it takes (hand count) to pin us down, but it may still be "testing" us. Maybe i am giving it too much credit. I do intend to push through to the 5000 hand threshold.

Good luck and God bless.

Hey Patriot, From the beginning, I had it set for optimal. Alfie is definitely set to not have a pattern. It does things that make you think a little before your next move, which is why I think it is a great little to use. I believe it an above average simulator to go up against. I use now just to maintain my game if Im not able to play a real moey session for a few days.
 
ventrolloquist

ventrolloquist

Visionary
Joined
Mar 20, 2019
Total posts
647
Chips
0
Im glad you gave it a try ventrolloquist. I've played a similar amount of hands on it and I experienced many of the things that you did. I wrote the original quote after I had lost $100 in 2 consecutive sessions. For me at first, it seemed I would get 3 bet at a very high rate. So for example if raised it 3x vs a raise of $9-my re-raise would be $27 and then the AI would 4 bet to $81. All of a sudden Im all in. I had KK and of course the AI was holding Aces. Session over. That was sort of how my 1st couple sessions went..

When I first started playing against poker alfie I thought I would win easily as I do with most of the AI's, but that it didnt happen, that is what prompted me to write a little review here on it. I'm sure beginners could learn alot about the game of poker if they played against Alfie.

Plus the fact that it was free. Its definitely a good warm up tool. Putting you in the proper mindset for a real money session.
It's definitely an above average AI. But I've also caught 4 betting and calling a jam with pocket 9's, it seems to often make very light plays, this could possibly be the range of hands it uses as bluffs and bluff catchers as wierd as it sounds.

Definitely top notch for a free tool though.
 
NWPatriot

NWPatriot

Rock Star
Joined
Jun 18, 2018
Total posts
480
Chips
0
Okay, I will keep my winrate reports to optimal slow speed games only then.

It doesn't learn, I asked the guy who created it once.

Has yours ever open limped on maximum playing strength? That's non-existent in nearly any GTO strategy as far as I know. (small blind flats are fine, I'm talking about open calling from the non blind positions)

So if it doesn't learn, then what about this is AI. To me then, this is just a program claiming to be AI. A bunch of pre-defined algorithms and calculations is not AI, right?

As for open limping, i am looking a the BTN open limping with me in the BB. So yes, this is part f its repertoire.

I have noticed the large betting at times, and it almost always seems to have me beat, though, not always.
 
ventrolloquist

ventrolloquist

Visionary
Joined
Mar 20, 2019
Total posts
647
Chips
0
So if it doesn't learn, then what about this is AI. To me then, this is just a program claiming to be AI. A bunch of pre-defined algorithms and calculations is not AI, right?

As for open limping, i am looking a the BTN open limping with me in the BB. So yes, this is part f its repertoire.

I have noticed the large betting at times, and it almost always seems to have me beat, though, not always.
I think I remember reading somewhere that Alfie uses similar logical deduction to what a chess computer uses, whatever that means, and in that sense it's not a preprogrammed algorithm since there are far too many unique situations which would require thousands of pages of code. So it's not quite your typical bot. So I guess what that means is it can still come up with decent solutions to atypical plays from human opponents. But I contacted the creator and he definitely said it doesn't adapt, that's the main thing I was curious about when I asked. Perhaps what he meant is it doesn't adapt to the player in the sense that it knows if you are a nit or a lag etc, but it does have some kind of deductive reasoning that reacts to a specific play.

I'm definitely having a lot of fun picking apart it's stategy and wierd plays, a few more thousand hands and I think I'll get the hang of it's logic :)

Programming a bot that adapts to a player is supposed to be a difficult feat anyways and not even snowie/pluribus do that.
 
Last edited:
manzanillo53

manzanillo53

Legend
Joined
Mar 13, 2017
Total posts
1,097
Awards
2
Chips
1
Thank very much, I need all the help I can get.
 
S

stil370

Visionary
Joined
Mar 19, 2010
Total posts
674
Awards
1
Chips
0
I think I remember reading somewhere that Alfie uses similar logical deduction to what a chess computer uses, whatever that means, and in that sense it's not a preprogrammed algorithm since there are far too many unique situations which would require thousands of pages of code. So it's not quite your typical bot. So I guess what that means is it can still come up with decent solutions to atypical plays from human opponents. But I contacted the creator and he definitely said it doesn't adapt, that's the main thing I was curious about when I asked. Perhaps what he meant is it doesn't adapt to the player in the sense that it knows if you are a nit or a lag etc, but it does have some kind of deductive reasoning that reacts to a specific play.

I'm definitely having a lot of fun picking apart it's stategy and wierd plays, a few more thousand hands and I think I'll get the hang of it's logic :)

Programming a bot that adapts to a player is supposed to be a difficult feat anyways and not even snowie/pluribus do that.

I am glad to hear that you are finding it to be useful ventroloquist. Im not afrraid to say that it kicked my ass the first couple of sessions. The more I use, the more I can anticipate what Alfie is going to do or at least take an educated guess but still after about 10 sessions or so, i still find it challenging. The first couple time I played against it, I was really surprised by the 3 and 4 bets it threw at me.
 
S

stil370

Visionary
Joined
Mar 19, 2010
Total posts
674
Awards
1
Chips
0
Thank very much, I need all the help I can get.

Have fun with it Manzanillo. Im sure you will learn a few things.
 
ventrolloquist

ventrolloquist

Visionary
Joined
Mar 20, 2019
Total posts
647
Chips
0
After playing 5000 hands against Alfie I managed a 27bb/1000 winrate. This was at the highest playspeed on optimal and therefore the lowest strength. Unfortunately I'm not around my computer often enough these days and therefore have to play on my phone which can't handle the more cpu intensive modes.

I played strict preflop ranges without flatting much. Played fairly balanced from an aggression standpoint but maybe didn't floaat as often as I should.

Noticed some major leaks such as donking from BB with a big size on flops that in my opinion don't make sense to do it on, and also min 3bets pre and calling from SB.

Hopefully the slower playspeeds are stronger.

Also I think I could have hit a higher winrate by flatting more unexpected hands.

Overall it's much better for learning than playchip games. Just has a few bizzare leaks here and there.
 
Igor Popadyk

Igor Popadyk

Visionary
Joined
May 7, 2015
Total posts
726
Awards
6
Chips
1
a very interesting topic, how much do these programs load the computer? are they safe?
 
ventrolloquist

ventrolloquist

Visionary
Joined
Mar 20, 2019
Total posts
647
Chips
0
After playing 5000 hands against Alfie I managed a 27bb/1000 winrate. This was at the highest playspeed on optimal and therefore the lowest strength. Unfortunately I'm not around my computer often enough these days and therefore have to play on my phone which can't handle the more cpu intensive modes.

I played strict preflop ranges without flatting much. Played fairly balanced from an aggression standpoint but maybe didn't floaat as often as I should.

Noticed some major leaks such as donking from BB with a big size on flops that in my opinion don't make sense to do it on, and also min 3bets pre and calling from SB.

Hopefully the slower playspeeds are stronger.

Also I think I could have hit a higher winrate by flatting more unexpected hands.

Overall it's much better for learning than playchip games. Just has a few bizzare leaks here and there.
Will post a screenshot of the stats page as soon as I can manage to log in on my pc
 
ventrolloquist

ventrolloquist

Visionary
Joined
Mar 20, 2019
Total posts
647
Chips
0
a very interesting topic, how much do these programs load the computer? are they safe?
Computer can handle it very easily. I think Alfie is basically a solver that takes a lot of computing shortcuts.

My phone on the other hand has a meltdown on the most cpu intensive difficulty and it takes each opponent in excess of a minute to make a move
 
V

vittopio

Visionary
Joined
Nov 13, 2019
Total posts
961
Awards
2
Chips
0
I want to share two FREE tools that I have recently come across that will no doubt make you a better poker player. In addition, unlike some other tools people use to get better at Poker, they are far from boring or tedious. So without further ado, they are:

1. Poker Alfie
This is an Artificial Intelligence 6max ring game. There are 3 settings, so it is good for beginners up to intermediate players. The competition ( I set it at the highest level ) will surprise you. I couldn't believe it the first couple times around because usually it doesnt take me long to figure out how to beat these pre-programmed opponents and ulitimately I walk way not having learned much. If anything after playing poker alfie, you will become disciplined, if not, you will end up losing. It is available on google play.

2. OBS Project
Some of you are probably already familiar with this tool, but Im pretty sure a great majority of forum members are not. OBS allows you to tape all of your sessions from beginning to end. In addition as you play, there is a audio component that allows you to talk about your decision making as you go along.
When you go back and watch the session and listen to what you were thinking, it will no doubt make you think twice about whether you made the right play or not. This also completely free. Google OBS Project and it will appear immediately.

Hope you all find these tools as useful as I did.
Thank you very much for the useful information (especially for 2 tools). That's why I love card chat! The only question is how much memory is needed - otherwise I'm afraid my old laptop might not pull?
 
ventrolloquist

ventrolloquist

Visionary
Joined
Mar 20, 2019
Total posts
647
Chips
0
Thank you very much for the useful information (especially for 2 tools). That's why I love card chat! The only question is how much memory is needed - otherwise I'm afraid my old laptop might not pull?
There are different playspeeds that require different computing power. I can run the most intense one cpmfortably with an average CPU and 8G RAM, with almost no hangups.

If you have a good phone you can run it comfortably on the middle level in terms of system requirments so an old laptop should handle it just fine I think.
 
ventrolloquist

ventrolloquist

Visionary
Joined
Mar 20, 2019
Total posts
647
Chips
0
If we're gonna be talking about sweet training tools that require 0 money invested it's also worth mentioning advanced poker training which lets you play the easy level tables for free (but you'll never find such easy games in real life except for in a home game).

Then there is slumbot, hibiscus biscuit, and cleverpiggy if you want to practice heads up. All of which are world class opponents.

Plenty of tools for people starting out who don't wanna risk a bankroll which are better than playchip games.
 
ventrolloquist

ventrolloquist

Visionary
Joined
Mar 20, 2019
Total posts
647
Chips
0
Playing on slow playspeed now and 700 hands in and I'm getting absolutely wrecked at -30bb/100. Hit a few massive bad beats though. Feels like it is playing a stronger preflop game but this could be variance.

Developer said the loss in AI strength at the fastest calculation speed is minimal.

Anybody else want to give this a shot? I'm genuinely curious if it is indeed a world class opponent like the developer claims.
 
Top