Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss
Optime Backgammon for iPhone possibly cheating (with evidence)
1,070 views
Skip to first unread message
tthoma...@gmail.com
Nov 20, 2011, 5:33:49 PM11/20/11
to
There's been a long history of people saying they think various backgammon programs cheat. Much (all?) of it is people's impressions, suspicions, etc. When asking for actual evidence, however, the arguments typically come up wanting. In the situation with Optime's (free) backgammon program for the iPhone, the situation is a bit different...
This version of backgammon allows you to undo plays, either playing the checkers differently or re-rolling the dice. You can back up moves presumably all the way back to the beginning of the game (although I haven't tried to go back that far). Messing about with this feature, I discovered something rather disturbing.
At any point in the game, if the player rolls a non-double, the computer always rolls a particular value, but if the player rolls a double, the computer always rolls a different value. So, for example...
Let's say at some point the player rolls 5-3, or 6-3, or 4-1, or..., the computer will always roll a 3-1 for its subsequent roll. However, if the player rolls a d6, or d5, or ..., the computer always rolls a 6-4. No matter what the player rolls at that point in the game, if they don't roll a double the computer rolls a 3-1, but if they do roll a double it rolls a 6-4.
These values are just for illustrative purposes. The point is, the computer's rolls are influenced by whether the player is rolling a double or not. I've played this through multiple moves, and found the computer's rolls mostly consistent even across multiple plays.
Now why would someone code a RNG for rolls that are dependent upon the player rolling doubles? If the computer always rolled the same value after backing up, that would make some sort of sense. It would mean for example the player RNG and computer RNG are independent, and the programmer may have wanted that as a "feature" to test the AI for various player moves.
However, the computer rolls are clearly dependent upon the player rolls. This is infinitely repeatable, so others can try this out. In the "real world" this would be evidence enough for investigation, and maybe even a search warrant. :) However, in our "virtual world", I'm not sure anyone really cares. I post this simply because I thought others might find the analysis interesting. Also, because of how easy it is to test, others can verify these findings if they wish.
This version of backgammon allows you to undo plays, either playing the checkers differently or re-rolling the dice. You can back up moves presumably all the way back to the beginning of the game (although I haven't tried to go back that far). Messing about with this feature, I discovered something rather disturbing.
At any point in the game, if the player rolls a non-double, the computer always rolls a particular value, but if the player rolls a double, the computer always rolls a different value. So, for example...
Let's say at some point the player rolls 5-3, or 6-3, or 4-1, or..., the computer will always roll a 3-1 for its subsequent roll. However, if the player rolls a d6, or d5, or ..., the computer always rolls a 6-4. No matter what the player rolls at that point in the game, if they don't roll a double the computer rolls a 3-1, but if they do roll a double it rolls a 6-4.
These values are just for illustrative purposes. The point is, the computer's rolls are influenced by whether the player is rolling a double or not. I've played this through multiple moves, and found the computer's rolls mostly consistent even across multiple plays.
Now why would someone code a RNG for rolls that are dependent upon the player rolling doubles? If the computer always rolled the same value after backing up, that would make some sort of sense. It would mean for example the player RNG and computer RNG are independent, and the programmer may have wanted that as a "feature" to test the AI for various player moves.
However, the computer rolls are clearly dependent upon the player rolls. This is infinitely repeatable, so others can try this out. In the "real world" this would be evidence enough for investigation, and maybe even a search warrant. :) However, in our "virtual world", I'm not sure anyone really cares. I post this simply because I thought others might find the analysis interesting. Also, because of how easy it is to test, others can verify these findings if they wish.
Tim Chow
Nov 21, 2011, 10:52:45 AM11/21/11
to
Even assuming that your observations are correct, this sounds like a
bug report rather than evidence of cheating.
---
Tim Chow
bug report rather than evidence of cheating.
---
Tim Chow
sully...@gmail.com
May 1, 2015, 7:46:23 AM5/1/15
to
This game is the biggest piece of shit online! Optime should be ashes that they have done nothing about it. I intend to tarnish the Optime name as long as possible. Worst game producer in the world. Please join me in boycotting Optime and finish this company forever for the lack of treating buyers with respect. Not cheat on them.
davidbl...@gmail.com
Oct 6, 2015, 8:32:10 PM10/6/15
to
On Friday, May 1, 2015 at 7:46:23 AM UTC-4, sully...@gmail.com wrote:
> This game is the biggest piece of shit online! Optime should be ashes that they have done nothing about it. I intend to tarnish the Optime name as long as possible. Worst game producer in the world. Please join me in boycotting Optime and finish this company forever for the lack of treating buyers with respect. Not cheat on them.
SORE LOSER!!!!!!
> This game is the biggest piece of shit online! Optime should be ashes that they have done nothing about it. I intend to tarnish the Optime name as long as possible. Worst game producer in the world. Please join me in boycotting Optime and finish this company forever for the lack of treating buyers with respect. Not cheat on them.
Grunty
Oct 6, 2015, 9:41:56 PM10/6/15
to
davidbl...@gmail.com escribió:
<nothing>
Wonder whatever happened to David Ulbricht? Paul, Walt would recall him. Nice guy.
<nothing>
Wonder whatever happened to David Ulbricht? Paul, Walt would recall him. Nice guy.
Paul
Oct 8, 2015, 8:28:07 AM10/8/15
to
Re "Nice guy" I'm not sure what you mean. Can you back that up a bit? I don't even know whether you're being sarcastic or serious. Not that I'm arguing with you. In fact, given that I don't know whether you're being sarcastic or serious, I couldn't argue with you even if I wanted to.
Paul
Grunty
Oct 8, 2015, 5:38:17 PM10/8/15
to
That's a 'nice guy' to me.
Tim Chow
Oct 8, 2015, 8:27:05 PM10/8/15
to
On Thursday, October 8, 2015 at 8:28:07 AM UTC-4, Paul wrote:
> I recall David Ullrich, yes. He stopped posting here, but continued posting
> in sci.math. Now, I think (but am not sure) that he stopped posting in
> sci.math
I think he's still active on math.stackexchange.com. As far as I know he's still an active math professor at Oklahoma State.
> I recall David Ullrich, yes. He stopped posting here, but continued posting
> in sci.math. Now, I think (but am not sure) that he stopped posting in
> sci.math
I also liked his posts here on rec.games.backgammon. In other contexts he could be rather...argumentative, let's say. I give him credit for helping me break my addiction to sci.math.
---
Tim Chow
Paul
Oct 9, 2015, 4:45:23 AM10/9/15
to
Paul
Justin Thyme
Oct 9, 2015, 9:11:48 AM10/9/15
to
Paul wrote:
> On Wednesday, October 7, 2015 at 2:41:56 AM UTC+1, Grunty wrote:
>> davidbl...@gmail.com escribió:
>> <nothing>
>>
>> Wonder whatever happened to David Ulbricht? Paul, Walt would recall him. Nice guy.
>
> I recall David Ullrich, yes. He stopped posting here, but continued posting in sci.math. Now, I think (but am not sure) that he stopped posting in sci.math
He posted less than one month ago in the thread "convex hull". He--how
> On Wednesday, October 7, 2015 at 2:41:56 AM UTC+1, Grunty wrote:
>> davidbl...@gmail.com escribió:
>> <nothing>
>>
>> Wonder whatever happened to David Ulbricht? Paul, Walt would recall him. Nice guy.
>
> I recall David Ullrich, yes. He stopped posting here, but continued posting in sci.math. Now, I think (but am not sure) that he stopped posting in sci.math
shall we say?--doesn't suffer fools gladly.
nickm...@gmail.com
Jan 8, 2017, 9:14:34 AM1/8/17
to
The obvious argument that the game never cheats is that the player is just a poor player, and does not use probabilities of rolls in their decisions. Thus creating the appearance that the game isn't cheating they are just making poor moves. Both skeptics and shills alike have been making that argument for years. However while it is true that my game can always stand improvement, I have been a competitive player for over 40 years, and while there are many players better than me, I'm not a poor player. Here are some observations about optime I have observed and this has been repeated. The beds a specific roll and no other roll to jump me and gets b it, I back up the move so that the computer needs a different specific roll to jump.me and gets that roll, I back nip again to the original position and the computer gets the original roll it needed in the first place. I even backed up to a third position and the computer gets that move also. Clearly the computers dice are tracking what it needs. This is repeatable in optime on hard setting. Also the number of doi n les the computer gets, is also suspicious. Additionally I have also observed the reverse where the b player consistently gets cold dice. Specific to what the player needs. I'm not sure why, but the evidence seems to support that the dice are not rand but rather calculated by the AI.
Michael
Jan 8, 2017, 9:56:05 AM1/8/17
to
On Sunday, January 8, 2017 at 4:14:34 PM UTC+2, nickm...@gmail.com wrote:
> The obvious argument that the game never cheats is that the player is just a poor player, and does not use probabilities of rolls in their decisions. Thus creating the appearance that the game isn't cheating they are just making poor moves. Both skeptics and shills alike have been making that argument for years. However while it is true that my game can always stand improvement, I have been a competitive player for over 40 years, and while there are many players better than me, I'm not a poor player. Here are some observations about optime I have observed and this has been repeated. The beds a specific roll and no other roll to jump me and gets b it, I back up the move so that the computer needs a different specific roll to jump.me and gets that roll, I back nip again to the original position and the computer gets the original roll it needed in the first place. I even backed up to a third position and the computer gets that move also. Clearly the computers dice are tracking what it needs. This is repeatable in optime on hard setting. Also the number of doi n les the computer gets, is also suspicious. Additionally I have also observed the reverse where the b player consistently gets cold dice. Specific to what the player needs. I'm not sure why, but the evidence seems to support that the dice are not rand but rather calculated by the AI.
Well, that's clear evidence the bot is cheating. Could you post those 3 positions and the rolls the bot was getting on each one?
> The obvious argument that the game never cheats is that the player is just a poor player, and does not use probabilities of rolls in their decisions. Thus creating the appearance that the game isn't cheating they are just making poor moves. Both skeptics and shills alike have been making that argument for years. However while it is true that my game can always stand improvement, I have been a competitive player for over 40 years, and while there are many players better than me, I'm not a poor player. Here are some observations about optime I have observed and this has been repeated. The beds a specific roll and no other roll to jump me and gets b it, I back up the move so that the computer needs a different specific roll to jump.me and gets that roll, I back nip again to the original position and the computer gets the original roll it needed in the first place. I even backed up to a third position and the computer gets that move also. Clearly the computers dice are tracking what it needs. This is repeatable in optime on hard setting. Also the number of doi n les the computer gets, is also suspicious. Additionally I have also observed the reverse where the b player consistently gets cold dice. Specific to what the player needs. I'm not sure why, but the evidence seems to support that the dice are not rand but rather calculated by the AI.
0 new messages