Call of Duty: Modern Bot Warfare

Small world! ��. Hopefully you were able to pick up some good tidbits from my lifting content back in the day.

I hear you on the bots. The “good” ones make the future seem bleak. The bad ones exist in a space somewhere between outright fish and bad reg. They’re lovely to have at the table.

I have my chat disabled which, honestly, might be a mistake, but bots typically don’t engage there. However, just as soon as the pool starts sharing around info like “bots never throw objects”… bots will start throwing objects, lol.

Thanks for the kind words! Glgl

Ha, maybe I shouldn't give them ideas! Although I'm sure there are much better ways to detect bots than "doesn't throw objects" lol

And yes, your content definitely helped me in the past. Looking forward to following your journey in poker as well!

My ‘Murican Thanksgiving holiday time is over and I am sitting at the airport waiting on a flight that has already been delayed twice… and I have a connection I am very much in danger of missing. Might as well try to do something productive instead of sitting around pissed off, right?

Tomorrow, I need to prepare materials for the Nacho’s Bot Strategy group study, but I did not bring a laptop so any real work building out and investigating stats in H2N4 is impossible. I won’t be playing my first hands of December until the second.

Still, post-holiday seems as good of a time as any for a bit of reflection and goal setting.

December Goals
Personal
*

• 1. Play 50k hands / 150hrs+ (125hr and 40k hands if I travel for Christmas)
  
• 2. Switch to a playing schedule with fixed hours: 12am-5am
  
• 3. Lose some weight
  

CFP Business
*

• 1. Complete the final major improvements to the Nacho’s Fish algorithm
  
• 2. Complete at least four major formation groupings in the new Nacho’s H2N4 that I am developing for the team
  
• 3. Start a public discord for anti-Bot strategy discussion and HH exchange for Ignition/Bovada/Bodog players
  

1) Volume Targets
I don’t believe in any type of leadership that comes in the form of “do as I say” and not “do as I do”. You should never ask someone to do something you’re not willing to do yourself. As such, even with ever-increasing responsibilities at Nacho’s, in my mind, it is imperative that I continue to lead from the front by maintaining a high volume of play. I don’t expect anyone to hold themselves to the same standards that I do, but, with my personal situation, anything less than 40k/m hands is unacceptable. With this being a holiday month, all-time volume records aren’t feasible, but that doesn’t mean I have an excuse to stop logging in, sitting in the chair, and clicking the buttons. Needs must.

2) Fixed Playing Schedule
As I alluded to earlier, with increasing CFP responsibilities, I can’t afford to have 13 hour marathon sessions that completely disrupt my sleeping schedule. If I give myself soft stopping rules, I inevitably abuse them chasing “Fish”. Because I have non-optional off-table responsibilities, and because I refuse to put out inferior materials/content to people who trust me with 1-2 hours of their time, I need to have a consistent, reliable amount of time to complete my off-table duties each week. Giving myself flexibility on when to end my session is not conducive to maintaining a steady, consistent stream of off-table working hours. I am going to give a hard stop a shot this month. I will only play between the hours of 12am to 5am EST.

3) Lose Some Weight
For my entire adult life I have been involved in some kind of competitive “sport”: powerlifting, bodybuilding, and, more recently, jiujitsu. In the last five months, I’ve stopped working out completely and I have paid zero attention to what I eat. The results have been predictable. While I am not willing to commit to a gym schedule right now, dieting takes no extra time and there is no excuse to continue with a poor diet. The initial goal will be to lose approximately 1% of my body weight per week. Currently, I have no idea what I weigh but I’ll find out as soon as I get and I’ll work from there.

4) Complete Nacho’s Fish algorithm
When working with H2N4, in order to be able to parse through the data and come to accurate conclusions, you have to define player types properly. If you’re trying to study regs or bots, you want as few fish as possible showing up in that data set. Fish lose 35-40bb/100 in our sample and they can quickly pollute reg&bot data when they sneak in the sample because they play a disproportionate amount of hands compared to a regular player (50+ vpip isn’t that uncommon). There are still a few important tells that I’d like to incorporate into that algorithm before tackling improving our bot algorithms in more depth.

5) Complete x4 Formation Groupings for the Nacho’s H2N4 Config
My goal here is to build a product similar to Mobius 2.0, but completely optimized for vs. Bot play. I believe my H2N4 config is the biggest source of edge I have right now and I am excited to get started on creating a licensable version for all Nacho’s players. I think it’s going to make a huge difference for team wide WRs.

6) Start a public Bot Strategy Discord server
In my blog, I’ve started to get regular questions about the bots. How many are there? Are they beatable? Is there more than one kind? Are they more prevalent at certain stakes? I’ve realized that we are in need of a spot where Ignition/Bodog/Bovada players can come together and discuss their experiences with the bots. Additionally, most players lack the data oversight to properly study the bots. One of the goals of this group will be to create a HH exchange so that individual contributors can gain access to a larger data set and thus be able to more productively research the bots themselves. Nacho’s possesses ~3mil Ignition/Bodog/Bovada hands for 2025 alone. I think we can potentially do some really cool things with an anti-bot group.

Dear Diary…
All in all, I’m hopeful that December will be just as productive as November in terms of on table results because I still have my eyes on moving up to 1k. However, even if that isn’t the case, I’ll make sure I control the controllables and make significant progress regardless.

Glgl in December fellas.
WAGMI

I mentioned before how dumb it is to keep going on about bot exploitation publicly

Your coaches were already part of an invite-only anti bot group, why on earth do you want to go make a public one. The bots will just find out about it, have someone pose as a player, and find out all the info you have on them

I get that you like to talk but have some common sense. The only explanation that makes sense for why you want to keep going like this is that you are trying to sell either your cfp or some product, which is what seems more and more likely

Again, why is it dumb? I asked you the first time and you didn’t reply.

A couple of things:
1) There is fundamentally no difference between an “invite-only” group where all you have to do is ask to be invited and a public group where you need an invite link.
2) In that “invite-only” group, no one is just giving away their exploits. I am in the group, I am one of the coaches, and we still discuss important information about the bots without giving away specific strategy: how many bots are there?, are they beatable?, what are the main stakes they play?, are there more than one type?, etc etc. If individual players want to share more than that, that’s up to them. Last but not least, that group requires the use of SnapMDA which, at this time, is not sufficiently powered to compete with H2N4 in terms of fighting the bots on Ignition.
3) Data won’t be exchanged without a vetting process. Frankly, even if bots had ALL our data, they would not be able to “find out everything we know”. Even actual members of Nacho’s CFP do not have access to “everything we know”. We don’t have a monolithic group-wide strategy. It takes hours of building your own stats and doing your own research in H2N4 to optimize anti-bot strategies. It’s not as simple as hurr durr huge river overbets are overbluffed. Fixing the leaks present in some of the bots is not as simple as bot devs reading a forum with a couple of hints here and there.

Of course we want dedicated new members to join the CFP at Nacho’s. More members means more quality data and more people collaborating on strategy to increase our edge. Surely you realize SnapMDA and Darius are directly selling products and coaching services?

I am confident I can protect my edge while still exchanging information and value with other Bodog players. You aren’t. You’re entitled to your opinion.

# 12/3: +108bb, -$334.41

Tracker Results:

Month-to-Day and Daily Graph:

Session Debrief:

• 1. I stuck to my hard time stop which was a positive. I still had good tables at 200nl. Normally, I find it hard to quit when that is the case.
  
• 2. I was a bit rusty and missed a few obvious bot exploits that I wouldn't pass up when fully "on".
  
• 3. The gdamned software kept me off of 500nl tables for at least 1 hour out of the 5 due to client issues. Standard Iggy experience, of course.
  

Sat in the seat. Captured some EV. We'll do it again tomorrow... and the next day... and the next day. At the end of the month, the results will be there.

GL with your grind!
WAGMI

Damn, you can actually beat the bots? Maybe there is hope for online poker after all now.

Are the bots on Iggy card sharing collusion bots or more basic?

Rough estimate from Nacho's CFP Data (not including November):
~750k total profit on the year for the Iggy team

~125k from Bots
~250k from Fish
~375k from "Unknown" (usually players that we weren't sure what they were because we only played 5-20 hands against them and it is anonymous pool -- players remaining unclassified after 5-6 orbits are usually just regs)

I would say with 95%+ confidence that they are colluding via card sharing based on the HHs I've seen. Lots of situations where one bot folded relevant blockers and another bot hero called/folded. On Iggy, we get the clairvoyant hand histories so we can see exactly what they're doing.

Oh right so they are like the GG ones then that we’ve observed. How do you guys use data to exploit them if they’re playing Balatro against your 2 cards?

It’s great news to hear that they are beatable though!

There aren't always two of them on the same table and that helps.

Even when there is, though, they play a hyperexploitative, MDA style of game. With enough data, you can gain vision on the spots where they're deviating to try to exploit the pool. Because they're incapable of adjusting in-game, you can kind of just keep funneling them towards the spots where you know what the data incentives are and they never, ever adjust in a given session. All adjustments they make are macro and only really happen when a critical mass of sorts is reached by the pool at large. I get the sense their strategy adjustments are AI-driven and that there isn't an actual human behind the wheel making the decisions. I don't think they (the bot devs) care what individual players are doing to counter-exploit them so long as they are winning against the pool as a whole.

We're basically like mosquitoes on an elephant's back hitching a ride, I guess. They beat the pool as a whole pretty soundly and we take a chunk of that back while they do.

Yeah that’s fair enough: so basically using MDA to find multi street MES on a ‘large scale’ whilst going under the radar?

The ones who made the bots on GG / WPN were reported to be operated by engineers in Eastern Europe / West Asia iirc. It’s great news that they only macro adjust though as it means you gan get enough hands to adapt your lines again and hopefully stay ahead of them.

Online Poker still being alive is good news at least in time for Christmas 😀

Yeah, you've basically nailed it. There's a bit of cat and mouse that goes on, for sure. To stay ahead, personally, I employ a combination of MDA and manual review of CV HHs. The 'ole human noggin' can sometimes spot patterns before they become obvious in the data and this generally lets us run an exploit for quite a while before they adjust (if they adjust). It is important to add "if" because there are a small number of things that they seem to never change no matter how hard they get exploited.

Fun thread, might be one of the most unique blogs this forum has seen. Hope you don't stop

1) how much does the stake you play really matter? Since you say there is on avg 2 bots per table that you guys presumably doing well vs from what you say. Shouldnt the skill gap between 500 and 2k be very small (non existent?) I assume rake is lower as well higher. Do the bots play the same on all stakes?
2) Has there been any refunds coming your way (team included) in last 18 months?
3) what about playing the bots HU, is that something you guys have figured out?
4) Have freenachos team suffered any more confiscated money/bans since that public case.
5) whats your rough guess how much the bots take from the pool in a year, do you guys got a idea in regards to their winrates?

Best of luck

WeÂ’re going to keep this train rolling. Thanks!

The stakes matter quite a bit for two big reasons. The first very noticeable drop in Fish ratio occurs between 200 and 500. At 1k, you can (occasionally) go entire nights without encountering a single big fish.

Second, there is more than one bot profile. We can just call them good bots and bad bots for simplicity. The bad bots play higher stakes, but tend to lose very badly and quickly withdraw for weeks at a time. They never leave 200nl and below. They are modest winners: 1-2 EV bb/100.

Good bots main stake is 1k with 500 and 2k being close behind in total volume. While they have some presence at 200 and below, they’re the main profile at 500+. They are huge winners with 5-6 EV bb/100. They’ve actually won even more than that in the last quarter after what appears to be an update to their strategy.

I am aware of zero.

Even bad bots are winning 9-10 EV bb/100 HU. I advise the guys to stay away from this spot. We only have 5k hands or so as a result. Even if you wanted to grind HU vs them, other players join the table quickly enough in most cases that it’s hard to build a sample.

Just logically speaking, it makes sense that a computer would have the biggest advantage in wide range spots where it is easier to make EV mistakes. I personally play a very cheap strategy HU against them. I take my button and leave. Every time. If they are at the table first and get the first button, I leave before a hand is dealt. I have a huge WR against them HU because of this (in a minuscule sample). My general recommendation is to do the same. Forgive me for not feeling bad about using these kinds of unfair tactics against bots, lol.

In every case, confiscated money has been returned including the public case you mention with Peter. However, they have banned at least one other account for winning too much. How much is too much? We don’t really know. One coach won 100k in a month and was not banned, but they play primarily 2k and their volume was enormous that month. Another player once won 20bb/100 at 1k for a full month and their account was closed with funds returned. To be clear, this isn’t a frequent occurrence but, yes, they ban people for winning too much. I’ll let the conspiracy people do what they want with that information.

Good bots? 1mil+ in 2025, 5.5 EV bb/100 (main stake 1k)
Bad bots? 200-300k, 1.5 EV bb/100 (main stake 200nl)
Total earnings could be way off because it’s hard to estimate their total volume. This is really just with Nacho’s at the table.

Keep the questions coming people. Happy to answer if you all find it interesting.

Thanks for the Interesting replies. Def noticed this as well that the bots moved up over time. Think for a long time 500 zone was the highest where they were a issue (excl 500 reg tables)

Will get back with more questions, find it very interesting as someone who played/backed hs for a decade in all formats basically. It's competitive as f what u guys do but nothing I got insight into/can relate to

12/4: -314bb, -$1494

Track Results:

Month-to-Day Graph:

Today's Graph:

Session Debrief:
Not a great start to the month, but what can you do? Zero concerns on my end. Not going to keep a negative blue line all month and my red-line is right where it always is.

After some feedback from my grinding partner, @MartinVee, I have decided that the hard stop at 5am is not a good idea in the long run. The best time to stop playing is when I start getting sat alone at 500nl tables. However, I may employ something like soft stop at 4:30-5:30am and a hard stop around 6:30-7:00am. I really need to be in bed by about 8am to not still be asleep when my wife gets home from work. It's hard on her when I am sleeping until 8pm and then she has to go to sleep at 9:30pm (she is a high school teacher and has to wake up at 6am every day). We don't get enough time together when I play too long and there are some things I will not sacrifice for poker (not many though, lol).

The bot group is off to a good start with 74 total members so far. As suspected, people have a desire to hear about this stuff but very few are willing to talk about it. I think that's mostly because they're misguided with regards to how the bot devs operate. They really think they're scouring 2+2 journals / small Discord for strategies clues when in reality they have millions upon millions of HHs from every site in existence and they don't need to do that. That huge mass of data gets fed to their AI and there likely isn't any human input whatsoever, lol.

As always, GL with your grind, fellas!
WAGMI

12/5: +$1694, +119bb

Tracker Results:

Month-to-Day:

Today:

Session Debrief:
Games were still pretty good but I had to invoke the hard stop. It's already 7:30am and I'm just now winding down. It's too late.

Funny enough, the **** run to start the month has bothered me exactly zero... BUT my mouse getting stuck going from screen to screen had me just about ready to punt. Not my stacks -- the mouse, though, yes. ��

Another solid day of grind in the books. Happy with my process today (besides the mouse...).

GLGL with your grind!
WAGMI

You have to be an incredibly smart person to go from starting poker in 2023 to beating tough bot games in a span of 2 years. Would you consider yourself extremely high IQ?

At the risk of making myself sound like a pompous ******* by talking about my own intelligence, my IQ is definitely above-average. How far above average? I really don't know. Maybe somewhere between 1-2 standard deviations from the norm, if I had to guess? However, when it comes to hours dedicated to the game, I am probably 5-6 standard deviations above the norm. I regularly put a combined 300 hours into study and play each month. Not sure which is more important (almost certainly the hours), but neither hurts my cause.

In August, I almost played 300 hours as part of a volume challenge:

I try to keep a more reasonable balance between study and play these days (results were pretty **** in August -- 2 ev bb/100)

12/6 Results: +28bb, -$168

Tracker:

December's Graph:

Today's Graph:

Session Debrief:
Played way past even my hard stop today, but my wife does not teach on weekends so the scheduling considerations are different. I was entirely disciplined with my stop. I had at least one very high quality Fish at 500 for the duration of the session.

Obviously running a little poorly to start the month, but capturing 11 EV bb/100 is not anything to be upset about. That's quality play, IMO.

We'll plug another 7-9 hour session tonight and make sure we soak up as much weekend EV as possible.

GLGL with your grind, gents!
WAGMI

12/7: -179bb, +$126

Tracker:

December's Graph:

Today's Graph:

Session Debrief:
No one cares. Work harder.

GLGL with your grind, men.
WAGMI

# 12/8: -41bb, -$635

Tracker:

December Graph:

Today's Graph:

Session Debrief:
Frustrating session. Frustrating week.

I will likely have to take tomorrow off to complete preparations for the weekly Nacho's Bot Strategy Call, but, if I can manage to get to sleep quickly, and I usually can't when I lose at... anything, I will hopefully get it done before prime time playing hours and get a session in as well.

I keep feeling like I burn win rate at the end of my sessions due to worsening game quality. I decided to fact check emotions and they were, basically, incorrect:

Oddly, I have lost the absolute most at the very beginning of sessions just after midnight. The drop off I perceive starting around 5am does seem to be real, but the hours are still profitable -- especially from 6am-8am.

If I was going to make a truly data-based decision here, I wouldn't implement a hard stop at 5am as I previously thought. Rather, I'd start my sessions one hour later, take a break at 5am to walk the dogs, do a Fish check, and play for another few hours if I find some. If I don't, quit.

I have a hard time believing the 12am-1am losing is "real" and not a sample issue. I also checked my entire Iggy DB instead of just the last six months and the results weren't fundamentally different:

I was also curious to look at the day of the week:

Of course, all of this suffers from the issue of insufficient sample, but it is interesting nonetheless. However, in the sample I do have, it looks like, if I want to target days off, they should be in the middle of the week.

In any case, back to work.

GLGL with your grind, boys.
WAGMI

As a working man I can tell you the fish come on when I do. Nights and weekends are always the best time for poker, live or online. Weekend nights (Friday,Saturday) are the very best for obvious reasons.

yoyo guys, also a member of Nachos for a few years. Long time member on this forum aswell. Izzy is a machine and huge asset to our team. Happy to also help anybody keen to learn more about beating bots.
Cheers
Max