Odin poker training tool under fire for removing feature to prevent cheating with it

[Dan Druff]

There's a bit of a "scandal" going on right now involving poker training tools, but at the moment it's mostly confined to arguing between rival training services and poker lab nerds.

However, I feel it's interesting enough to delve into, so I'll do my best to explain what's going on here, in simple language.

There are various pieces of poker training software out there, which have the capability of analyzing actual hands and advising what is the recommended play. This is actually quite useful, and is a much better learning experience than simply watching others' hands and attempting to apply it to your actual play later. With this, you get a GTO (game theory optimal) software product critiquing your actual play.

This eventually evolved into "GTO calculator" software which can analyze your REAL TIME play, so you don't have to go back to it hours later to remember the hand and your thought process. Unfortunately, this has the tremendous potential for abuse. If the software is analyzing your hands as you play them live, then it can advise you the best way to play them before you act, thus becoming equivalent to using a bot to play for you!

The developers of these types of software came up with a fair compromise to the situation of potential cheating. Many software packages instituted a delay in the analysis of hands entered in "real time", thus making it impossible to use the software to make live decisions for you. This has been the case for several years.

Rory Young is the founder of Prometheus Poker and Odin Poker. Odin's website now brags, "Instant Solutions! No delay. NEW!"





Some in the poker community are understandably furious with Rory Young, and his blatant promotion of a real-time solver, which even amateurs can use to play like super-experts. It wasn't always like this. Rory once had a delay in Odin, but it has recently been removed.

What was Rory's response? Read on...

[Dan Druff]

Rory defended his decision to remove Odin's delay by trying to claim that similar training products also have no delay, and his software was struggling as a result. Thus, Odin's delay was also removed.

Embedded Tweet

https://twitter.com/RoryYou59455440/status/1632926941270425600

Embedded Tweet

https://twitter.com/RoryYou59455440/status/1632926943568965634

Embedded Tweet

https://twitter.com/RoryYou59455440/status/1632926945942929412


It is important to understand that these "calculators" do not scrape the screen, but rather the player enters in all of the data, such as the cards he holds, the makeup of the board (if any), his position, his opponents' positions, and the respective number of big blinds of each person in the hand.

You see in the first tweet above that Rory mentions GTOWizard. That is a different software package which also doesn't have a delay (and apparently never did), though according to Odin's own promotional material, Odin allows for quick selection of a lot of less-common situations, whereas GTOWizard supposedly does not.

Rory's point was that his company's "10 developers have families to provide for", and his software was simply not selling as well when it had a 20-second delay.

Interestingly, in a July 2021 video featuring promoter Fedor Holz, both of them talked about how "important" it was to "keep the games safe" from cheating. Apparently not anymore. Go to the 6:54 mark below to hear that discussion.





Laughably, the end of Rory's defense blamed the online poker sites for not being good enough at catching the type of cheating that his tool provides! LOL!!


What were some of the community's reactions to this? Read on...

[Dan Druff]

KL Cleeton, who has his own training company, called it straight up cheating. However, Rory reiterated that he attempted to get an agreement where everyone would have an agreed-upon delay, industry-wide, and that nobody would formally agree:

Embedded Tweet

https://twitter.com/RoryYou59455440/status/1633194309666078725


Ryan Laplante called on Phil Galfond (Run It Once) and Dominc Nitsche (DTO Poker) to agree to this, and said that his company (Learn Pro Poker) would be glad to join in:

Embedded Tweet

https://twitter.com/DominikNitsche/status/1633242116536057857


As Nitche pointed out, even if Rory and the other come to an agreement, there will still be tools like GTO Wizard with no delay, which can be used for cheating. Ryan stated that perhaps a formal agreement between most industry training companies would pressure all of the others to do the right thing, but this is not likely. Note that GTO Wizard's owner is unknown, and thus will not be easy to pressure into such an agreement.

Embedded Tweet

https://twitter.com/DominikNitsche/status/1633244726601064450

Embedded Tweet

https://twitter.com/DominikNitsche/status/1633246404473651202


Nitche also mentioned that Run It Once's "Vision" software also allows such real time strategy to be calculated with no delay, and wondered why Galfond hasn't been given any shit for it. That's a good question.


Unfortunately, the idea of an industry-wide agreement to have a delay broke down, when Rory claimed it would be unfair to his company, as long as GTO Wizard is not on board for the same thing. Additionally, while Nitche's tool does not have the same capabilities to calculate in real time, he also refused to get on board for the moment:

Embedded Tweet

https://twitter.com/RoryYou59455440/status/1633250345747984384

Embedded Tweet

https://twitter.com/Protentialmn/status/1633254278793150469

Embedded Tweet

https://twitter.com/Protentialmn/status/1633255861060796416

Embedded Tweet

https://twitter.com/Protentialmn/status/1633256313064128512

Embedded Tweet

https://twitter.com/Protentialmn/status/1633261114707574787



Vanessa Kade, always one who enjoys drama, got this jab in against Rory:

Embedded Tweet

https://twitter.com/VanessaKade/status/1633175804611014656


I'll give you my opinion in the next post...

[Dan Druff]

First off, I want to say that what Rory did with Odin was fairly scummy. It's one thing to have a tool which never had any delay, where the developer takes the attitude of, "It's not our problem if people use it to cheat."

It's another if the developer actually foresees that it might be used for cheating, builds a delay in to prevent that, promotes it as responsible software, and then removes the delay when they realize it's not selling as well as the cheat-enabling competitors! It's even worse when they advertise "INSTANT SOLUTIONS, NO DELAY" upon making that change, which is basically beckoning cheaters to come buy it.

However, I will say that I understand Rory's position here. He was tired of getting his company's lunch eaten by less responsible companies which simply chose to ignore the potential cheating implications of their tool, including the highly respected Phil Galfond. If squeaky-clean Galfond can sell such software with no delay, why is Rory Young such a bad guy for doing the same thing? That's a fair point.

In response, though, I will say that "other guys are doing it too" isn't the best excuse to be selling a piece of cheating software, especially if you're promoting the "no delay" aspect of it, presumably to make more sales with those with nefarious plans. It should also be noted that Odin seems to have more options for quick input of less-common situations than the other pieces of software, and may actually be the best choice for would-be cheaters! (I'm going by their own marketing video, not personal experience.)


Rory awkwardly made the point that poker sites should be the ones to police cheating, rather than the makers of training tools, but that argument doesn't really fly. The top responsibility for cheating falls upon the cheater himself, the second-tier responsibility falls upon the maker of the tool to do it, and the third-tier responsibility falls upon the poker site for failing to catch it. He can't really pass the buck on this one, by essentially saying, "Don't blame my cheating tool, blame the sites for not catching cheaters!"

At the same time, these are all businesses, and it sucks when not everyone is on the same page regarding the "power" of the tools being provided to end users. You absolutely should not be selling a tool which can be used for cheating, even if that isn't its stated purpose. Even if others are doing it, that doesn't mean you should. I heard this same argument in the 1980s, when drug dealers would say, "Sure, I sold drugs to those junior high kids, but if they don't get it from me, they'll just get it from someone else!" You can't use the "everybody's doing it" excuse when doing something unethical.

Could Rory perhaps use the "guns don't kill people, people kill people" argument, when it comes to these solvers? Could he simply say that it's not his responsibility if adults misuse his tool to cheat, and you should blame the person misusing it? Not really. This is because a delay is a practical compromise to prevent abuse, whereas there would be no way to sell a gun with a "firing delay"! A gun by itself is a destructive tool, and it's up to the user to be responsible with it. Poker solvers are not supposed to be destructive at all, and it's up to the creators to ensure that.


I don't want to completely vilify Rory here. It does look like he made a legitimate effort to get these other packages to agree to the delay in 2021, and even after failing to secure such an agreement, he instituted a 20-second delay anyway. It should be noted that 20 seconds isn't quite long enough to fully prevent cheating, but it does put a big crimp in the ability to do so. I will give Rory credit for going ahead with this delay for two years, even when his competition refused. I also give him credit for being willing to reinstate it, as long as everyone else agrees to, as well.

However, I cannot give him credit for refusing to reinstate it, so long as even one tool remains out there without such a delay. It sounds like he's okay with his tool being used for cheating, as long as he's not the only one guilty of providing such a thing.

After all, "WE GOTTA EAT!"

[Dan Druff]

Postscript:

Rory is pointing out that there is a lot of selective outrage here. He mentioned that another tool with no delay, Ruse, was promoted on Berkey's show, without anyone acting concerned about cheating implications:

Embedded Tweet

https://twitter.com/RoryYou59455440/status/1633450161044180994


He has a good point, but it's not just Berkey. It seems that nobody was concerned with this until right now, even though everyone should have been.

However, I still think Odin/Rory is the most egregious of the bunch, as they DID have this concern in 2021, only to shrug their shoulders and remove the delay anyway. Still, it's worth noting that this has been an ongoing issue which nobody seems to have called out until Rory did it.

[SlowRoll]

so...
1. poker coach has students install their custom solvers on the players' devices
2. shows them how to enter in their hands in real time
3. solver software is relaying hole card information somewhere to be processed.

i wouldn't recommend providing real time hole card information to a 'custom solution' built and promoted by people who lack any depth of knowledge as to how the data is being secured.

rory looks like an opportunist scumbag here, but that is what you can expect from the poker world.

those rallying against him should dig into the security risks of sending hole card information to an unknown 3rd party to be processed.
probably somewhere in aws or another cloud service without proper security measures around it.
this is what will make poker player skeptical of the software.... security risks to THEM. not 'the health of the game' - that is not a priority to most poker players
____
also, since he was mentioned.... berkey is the epitome of 'fake it til you make it' in every endeavor that he publicizes. his words are big (thanks to his thesaurus) and often have very little substance.
not surprising that he seems to have promoted the exact type of software he rallies against.
im sure he'll have a long convoluted explanation, when a simple 'i should have done better' is what is warranted.

[highhands89]

so...
1. poker coach has students install their custom solvers on the players' devices
2. shows them how to enter in their hands in real time
3. solver software is relaying hole card information somewhere to be processed.

i wouldn't recommend providing real time hole card information to a 'custom solution' built and promoted by people who lack any depth of knowledge as to how the data is being secured.

rory looks like an opportunist scumbag here, but that is what you can expect from the poker world.

those rallying against him should dig into the security risks of sending hole card information to an unknown 3rd party to be processed.
probably somewhere in aws or another cloud service without proper security measures around it.
this is what will make poker player skeptical of the software.... security risks to THEM. not 'the health of the game' - that is not a priority to most poker players

Yeah, that's not how the software works. The user has to manually input all of the information about a hand in order to see the solution. No publicly offered post-flop database service has any sort of card/data scraping happening.

[SlowRoll]

so...
1. poker coach has students install their custom solvers on the players' devices
2. shows them how to enter in their hands in real time
3. solver software is relaying hole card information somewhere to be processed.

i wouldn't recommend providing real time hole card information to a 'custom solution' built and promoted by people who lack any depth of knowledge as to how the data is being secured.

rory looks like an opportunist scumbag here, but that is what you can expect from the poker world.

those rallying against him should dig into the security risks of sending hole card information to an unknown 3rd party to be processed.
probably somewhere in aws or another cloud service without proper security measures around it.
this is what will make poker player skeptical of the software.... security risks to THEM. not 'the health of the game' - that is not a priority to most poker players

Yeah, that's not how the software works. The user has to manually input all of the information about a hand in order to see the solution. No publicly offered post-flop database service has any sort of card/data scraping happening.

I didn't mentioned any scraping.
The user enters the hand information, are you saying that this solver software is processing the hand on the users' PC?

[SlowRoll]

Yeah, that's not how the software works. The user has to manually input all of the information about a hand in order to see the solution. No publicly offered post-flop database service has any sort of card/data scraping happening.

Based on Druff's post, the user of the solver is sending their hand information to the Odin software. No scraping.
User submits the information in real time via Odin, which likely sends that data to an external service for processing.

Sending information about my hand to an unknown recipient in real time is not something that is appealing.

[Online Veteran]

I don't play much online poker anymore. Won't the poker site catch someone using this, ban them, and confiscate all their funds?

[SlowRoll]

I don't play much online poker anymore. Won't the poker site catch someone using this, ban them, and confiscate all their funds?

As high hands stated - they aren't scraping the screen so I'm guessing its not against TOS.

I believe the question is about the ethics in providing "training tools" which can easily be used to gain a huge advantage over opponents who are not "in the know".

I don't have a strong opinion either way on the tools themselves - they will eventually ruin the game, no matter what.

The advertising of Odin with "Now No Delay!" certainly seems like the "training tool" is being marketed toward those looking to use it as a real time assistance tool, rather than a training tool.

Gambling is always going to be filled with shadiness, grey areas. It's off putting to see someone operating in / taking advantage of these grey areas while making a ton of excuses for the behavior and trying to gaslight people to believing that they are not operating in that grey area.

[BCR]

Yeah, that's not how the software works. The user has to manually input all of the information about a hand in order to see the solution. No publicly offered post-flop database service has any sort of card/data scraping happening.

Based on Druff's post, the user of the solver is sending their hand information to the Odin software. No scraping.
User submits the information in real time via Odin, which likely sends that data to an external service for processing.

Sending information about my hand to an unknown recipient in real time is not something that is appealing.

I understand the concept, but how does one input an entire scenario in like 30 seconds if it’s not scraping the screen? Manually? Does anyone have a video of someone doing it? I can’t imagine you can do it at any volume.

[SlowRoll]

Based on Druff's post, the user of the solver is sending their hand information to the Odin software. No scraping.
User submits the information in real time via Odin, which likely sends that data to an external service for processing.

Sending information about my hand to an unknown recipient in real time is not something that is appealing.

I understand the concept, but how does one input an entire scenario in like 30 seconds if it’s not scraping the screen? Manually? Does anyone have a video of someone doing it? I can’t imagine you can do it at any volume.

I also cannot imagine doing it; but there is mention of shortcuts for common scenarios. Probably some sort of scripting.

If they go so far as to add some basic ML; one could identify those situation and generate scripts for them very easily.

Scary to think how these tools could be used (and probably are being used) by the people with a particular set of skills.

[PositiveVariance]

I understand the concept, but how does one input an entire scenario in like 30 seconds if it’s not scraping the screen? Manually? Does anyone have a video of someone doing it? I can’t imagine you can do it at any volume.

I also cannot imagine doing it; but there is mention of shortcuts for common scenarios. Probably some sort of scripting.

If they go so far as to add some basic ML; one could identify those situation and generate scripts for them very easily.

Scary to think how these tools could be used (and probably are being used) by the people with a particular set of skills.

I have never used one, but this is how I would set it up:

It would only take about 3 seconds per round. When you first sit down at the table, you will need to enter some info that the program will keep track of and update on its own after each round.

Information recorded:
Number of Big Blinds for each player
Position of each player
Your hole cards
Run out of board
Amount Bet/Raised/Called by other players and yourself

I'm assuming the best way that this could be done is by having the training tool set up on a large tablet where all info is imputed via touch screen buttons.

The following would only be done ONCE when you first enter the table: (10 SECONDS)
On the training tool you would have an outline of a table with 9 players. First question it ask is how many players? Press "9 ENTER".

It sets up 9 spots at the table on the trainer.

Next it ask you for location of button?
Let's say seat 4)
Press "4 ENTER"

After pressing "9 ENTER, 4 ENTER" the program automatically highlights seat number #4 and ask for "Number of Big Blinds seat 4 has?" Let's say it's 100.
The best way would be to have a keypad on the screen displayed with large "buttons" that read:
5-10-15-20-25-30-35-40-45-50 and so on..
You press the "100 button" there would be no enter. After pressing "100", it now highlights seat #5, you do the same thing. It goes in position seat by seat.
To enter everyone's stack you would be pressing only 9 buttons. So technically you don't need "Enter" as it would jump to what info is needed next. You can even learn the touchscreen by "touch" and look at stack sizes while your hand presses the 9 correct "buttons".

The above process is only done once at the beginning of the session. This keeps track of:
1. Player position 2. Stack size. The program will update with each hand, you no longer need to enter this information.

Once you start playing:

You see your hole cards. You will have a similar touchscreen display:

DIAMONDS A-K-Q-J-10-9-8-7-6-5-4-3-2
And a row for every suite. A touch screen keypad is displayed on the entire screen (or whatever size is needed).

After you enter your hole cards, it would remove the touchscreen display and highlight the first player to act. Let's say the First player "Bets", you would press that seat number on the table. Then it would automatically go to the next prompt of how much bet. Let's say $50. Press the "$50" button. It jumps to the next action for you to "tap" the next person to act. You can "skip" checking as it is assumed players will check if there is no bet yet that round. Let's say you call the $50. Tap "yourself" then "Call button". After you had entered hole cards, this is only 3 more buttons pressed.

Keep in mind it will always know what the next desicion is that needs to be made, so there would not be a need to "press Enter" each time.

Flop comes, same thing it displays 4 rows of "tap buttons" on the screen. After seeing the flop you enter the 3 cards. After pressing 3 "buttons" (2-3 seconds) it jumps to the next action and highlights next player to act which would be "Check, Bet, Fold".

Rinse and repeat for the Turn and River.

Once hand completes you simply tap which seat won the hand. The Program AUTOMATICALLY UPDATES stack sizes since this information was imputed each round.

So yes, when initially sitting down at the table it may take up to 10 seconds (ideally less than 5 seconds). Beyond this, once you are familiar with the program, it should not take more than 2-3 seconds to enter info for each round. There should never be a "holdup" as it will take other players far more time to decide before it's your time to act.

I can see guys playing 5 tables at the same time.


And this is with 5 mins of brainstorming, and someone that has never played poker. Imagine what a set of engineers and real poker players can come up with?




To be honest, I'm surprised people still play online poker.


One more thing, for "Solved decisions", answers should be instant. However, if it's running "Non-solved" questions through a simulator, or utilizing previous hand histories to determine the optimal play, I am not sure how quickly it will provide an answer.

[PositiveVariance]

I don't play much online poker anymore. Won't the poker site catch someone using this, ban them, and confiscate all their funds?

Ideally you will run the training program on a separate computer.

It's not hard to get away with this. These companies claim they can catch these people using solvers. The ones with half a brain they may catch.

How not to get caught:

1. Run Solver on separate device

2. On the bottom 10% of +EV hands where the solver tells you to "Call", put this in a random number generator where you would only actually call 50% of the bottom 10% of hands. Your bottom 10% may be +.01% to 0.5% in EV earned. The reason for this is so when they run you hand histories against optimal play that the solver says is the correct play, you will only be 90% overlapping, but give up very little EV. Your doing the bottom 10% to give up as minimal EV as possible, while being viewed as a chump. This is similar to how a BJ card counter will do cover plays to not get heat, while only costing them a minimal amount of money (EV).

3. Same thing as in step 2, but instead of "call/no call" you will be betting 10% or so of your overall bets/raises slightly OVER or slightly UNDER the optimal amount given by the solver. Use a random number generator to pick 10% at random, then vary those 10% of bets between say 1% and 10% of the optimal number the solver generates. So if the solver says bet $1,000, you would bet at the least $900, and at the most $1,100, all determined by a separate RNG. Steps 2 and 3 will only overlap with "perfect play" by about 80%. Overall probably only giving up a few percent in EV. This is more than some tip the dealer.

4. Vary the time it takes you to act. Responding in the same amount of time, everytime is an indication of using a solver or a bot.

5. Cash out frequently. If they do nail you it's for very little money.


6. Use numerous accounts. This way there is not enough hand histories to overlap with perfect play. Kill an account after X number Of hands. How much is that number? I am not sure. Probabilities will have to be solved to determine likleyhood of particular indicators matching up after a set number of hands.


There are more, but this shall do for now

Even if they happen to take your money every now and then, you will be playing with a large enough edge for that to be ok. Their "Standard of Proof" is whatever they want it to be. Your at their mercy.

[PositiveVariance]

I personally do not necessarily disagree with what Rory did with Odin. He tried to do the right thing by implementing the delay. You can not expect a guy to compete in an industry where everyone has a delay in their product. The other option is to sell his his business. This wouldn't be fair to him.

Here is a question that someone could ask that MANY WILL HAVE A PROBLEM WITH:

I'm not saying the following is my option, but it does bring up an interesting point...

-"Perfect Play" is allowed at the Blackjack tables via basic strategy charts. "Perfect Play" via basic stratagy didn't come for some time until after Blackjack was a common table game in casinos.

Poker has traditionally always been a game where it is possible to profit in both the short and long term.

IF EVERYONE AT THE TABLE PLAYS EXACTLY THE SAME, POKER IS A -EV GAME DUE TO THE RAKE!

As we know, everyone at the table doesn't play exactly the same and this is what makes it profitable for some and unprofitable to others.


Let's say a trainer is utilized in real time by everyone at the table. Depending on the rake, let's say everyone would be at a 1% disadvantage. Every $100 they bet they get $99 back.

If a basic stratagy chart is used in blackjack, the player has a 1% disadvantage.

Poker 1% disadvantage with everyone using same solver
Blackjack 1% disadvantage using Basic Stratagy that was created with math and verified via simulations and solvers

Someone could ask, "What is the difference?"

Once again, this is not necessarily my opinion, just a point of view that could be brought up.


Blackjack can go from -EV of 1% to +EV with Card Counting.

Poker can go from -EV with Solvers to +EV for some people without solvers.

The difference?

With Blackjack there is a casino to tell you "NO" you can't count cards while playing. In poker there is no one to say "NO" you can't do it.

People who historically have made money playing poker want to keep it that way. They don't want solvers. People who have historically lost money playing poker want solvers.

It comes down to poker traditionally has been profitable to the best of players, while the house makes a rake on the game.

An interesting discussion would be if solvers were allowed in LIVE POKER games. It would turn the game -EV but this may actually INCREASE the number of players. -EV games are the most popular in the casino. Look at Bacarrat, Blackjack, and Craps. Many people are intimidated by poker since they know the best players will beat them. If they are allowed a "Solver" they will think they actually have a shot. This is similar to how when Thorp released "Beat the Dealer", everyone tried to count cards and the casinos Blackjack business exploded.

Conclusion: Solvers are a bad idea for winning poker players but possibly a good idea for casinos to create rake. Poker would essentialy become a "pit game" if solvers were allowed, they would also be able to GREATLY increase rake since it would attract "Poppy" type players not players looking for an edge.

Another idea? Have a separate live table where you can use whatever solver you want and it would get very interesting .