Detecting Bots: Uncovering Insights from GGPoker Data Analysis
05-04-2024
, 01:52 PM
05-27-2024
, 10:13 PM
Bot Analysis Report: Tight Cash Bots for Rake Races
We continue our analysis of various bot groups. This time, we focus on tight cash bots that play primarily for rake races and rakeback. Out of this group, 10 nicknames were previously reported by y2da, and we've identified 2 additional accounts.
General Gameplay Patterns:
1. 0% Cold-Call (CC) from HJ, CO, and SB: Only 3% CC from the BTN.
2. 70% Fold vs. 3-Bet: Significantly overfolding.
3. BB Isolation vs. SB 30%: Solver frequency is 40-42%.
4. C-Betting Ranges: 63/49/47 - tight c-betting ranges, almost similar to solver frequencies.
5. Raiser Out of Position (OOP) and Raiser SBvsBB Freqs slightly higher and value-oriented than GTO.
6. Delay C-Bet Frequencies and BXB: Solver or below solver frequencies.
7. Caller OOP: Probe Turn, Probe River, XXB also at solver or below solver frequencies.
8. C-Bets Flop into 3-Way Single Raised Pot (SRP) Freqs.
C-bet Flop into 3-Way SRP vs IP/OOP is very tight.
9. C-Bet Flop into 4-Way SRP: Even tighter, hypervalue-oriented.
I could elaborate on their characteristics extensively, but the main point is that they employ a mix of solver-oriented and tighter strategies. All accounts use a single symmetrical strategy. Instead of delving into more specifics here, I invite you to examine their hands yourself for a more detailed study.
https://mega.nz/file/G9EgRIoR#5_XTzi...MC5ZPO8LgxADgg
Results with This Strategy:
Despite their structured approach, their results were not impressive. They couldn't beat the rake and played at -3.7 EVbb/100, but remained profitable or breakeven due to rakeback and promotions.
You might wonder, if they don't win much, why does it matter? The issue is that poker is meant to be a game of people competing against each other, not bots versus humans. A bot at the table kills the enjoyment, even if it isn't a particularly dangerous opponent. Botfarm also start with cautious strategies in low-stakes games and through years of evolution, develop exploits to gain significant advantages. Other botnets like this could follow a similar path. The only effective way to combat them is by studying their strategies and monitoring their evolution. It's impossible to counter them effectively without transparent hand histories.
Donation Wallets:
BTC: 3NCGQ6kiKVRVf2HCdXPGGsdvwTNZcnXpNX
ETH: 0x32fAa44bf82B8565c67bF551F801BB1feA7eC5F0
USDT (TRC20): TNxEzYgoYMcoyHmLuJYtxSKNsXDwHtThu5
TON: UQDEMfu7YIOp8di0rN__G_elC6SdZLERJvAcU3z0nUwRNVti
We continue our analysis of various bot groups. This time, we focus on tight cash bots that play primarily for rake races and rakeback. Out of this group, 10 nicknames were previously reported by y2da, and we've identified 2 additional accounts.
General Gameplay Patterns:
1. 0% Cold-Call (CC) from HJ, CO, and SB: Only 3% CC from the BTN.
2. 70% Fold vs. 3-Bet: Significantly overfolding.
3. BB Isolation vs. SB 30%: Solver frequency is 40-42%.
4. C-Betting Ranges: 63/49/47 - tight c-betting ranges, almost similar to solver frequencies.
5. Raiser Out of Position (OOP) and Raiser SBvsBB Freqs slightly higher and value-oriented than GTO.
6. Delay C-Bet Frequencies and BXB: Solver or below solver frequencies.
7. Caller OOP: Probe Turn, Probe River, XXB also at solver or below solver frequencies.
8. C-Bets Flop into 3-Way Single Raised Pot (SRP) Freqs.
C-bet Flop into 3-Way SRP vs IP/OOP is very tight.
9. C-Bet Flop into 4-Way SRP: Even tighter, hypervalue-oriented.
I could elaborate on their characteristics extensively, but the main point is that they employ a mix of solver-oriented and tighter strategies. All accounts use a single symmetrical strategy. Instead of delving into more specifics here, I invite you to examine their hands yourself for a more detailed study.
https://mega.nz/file/G9EgRIoR#5_XTzi...MC5ZPO8LgxADgg
Results with This Strategy:
Despite their structured approach, their results were not impressive. They couldn't beat the rake and played at -3.7 EVbb/100, but remained profitable or breakeven due to rakeback and promotions.
You might wonder, if they don't win much, why does it matter? The issue is that poker is meant to be a game of people competing against each other, not bots versus humans. A bot at the table kills the enjoyment, even if it isn't a particularly dangerous opponent. Botfarm also start with cautious strategies in low-stakes games and through years of evolution, develop exploits to gain significant advantages. Other botnets like this could follow a similar path. The only effective way to combat them is by studying their strategies and monitoring their evolution. It's impossible to counter them effectively without transparent hand histories.
Donation Wallets:
BTC: 3NCGQ6kiKVRVf2HCdXPGGsdvwTNZcnXpNX
ETH: 0x32fAa44bf82B8565c67bF551F801BB1feA7eC5F0
USDT (TRC20): TNxEzYgoYMcoyHmLuJYtxSKNsXDwHtThu5
TON: UQDEMfu7YIOp8di0rN__G_elC6SdZLERJvAcU3z0nUwRNVti
05-28-2024
, 12:36 AM
Today
, 05:51 AM
Hey guys, I messaged Tyler about this one, but wanted to share suspicions about the influx of new players in the NL100-500 regular tables player pool.
Although we have a small sample (due to Smarthand having issues tracking GGPoker the last couple of months), we can clearly see very many similarities between these new accounts, they share some characteristics from the original bots that Tyler pointed out in this thread, which I've played thousands of hands against them (noteworthy GG probably banned them and there was no email from them about the issue, nor the compensations).
This is Excel sheet for some stat comparison of these players :
All these accounts are new (2-4 months old), and they are very active currently, but the small sample is just because GG is trying to combat data tracking. There are more of these players, but I encountered these ones in the past 24 hours and decided to do some digging. Like I said I already wrote to Tyler, maybe he could dig deeper and get some more insights.
Although we have a small sample (due to Smarthand having issues tracking GGPoker the last couple of months), we can clearly see very many similarities between these new accounts, they share some characteristics from the original bots that Tyler pointed out in this thread, which I've played thousands of hands against them (noteworthy GG probably banned them and there was no email from them about the issue, nor the compensations).
This is Excel sheet for some stat comparison of these players :
All these accounts are new (2-4 months old), and they are very active currently, but the small sample is just because GG is trying to combat data tracking. There are more of these players, but I encountered these ones in the past 24 hours and decided to do some digging. Like I said I already wrote to Tyler, maybe he could dig deeper and get some more insights.