Baiting with Game Theory

By joining our community you will have the ability to post topics and access other forums reserved for members. Registration is quick, simple and absolutely free. Join our community today by clicking here.

insults and more · Posted: Thu Dec 30, 2010 1:17 pm

This may take some explaining.

I'll start with the classic Prisoner's Dilemma:

Quote:

Two suspects are arrested by the police. The police have insufficient evidence for a conviction, and, having separated the prisoners, visit each of them to offer the same deal. If one testifies for the prosecution against the other (defects) and the other remains silent (cooperates), the defector goes free and the silent accomplice receives the full 10-year sentence. If both remain silent, both prisoners are sentenced to only six months in jail for a minor charge. If each betrays the other, each receives a five-year sentence. Each prisoner must choose to betray the other or to remain silent. Each one is assured that the other would not know about the betrayal before the end of the investigation. How should the prisoners act?

Each prisoner must take into account for his decision, the decision of the other. This produces the uncomfortable dilemma in that although silence on the part of both prisoners would yield the minimum penalty for each, both prisoners are led to defect lest the other betray him. thus receiving hefty penalties. Any rational prisoner will defect rather than risk the maximum penalty caused by being the only one to cooperate.

This is but one of many games studied in game theory. All of them have a devious, Machiavellian flavour to them. And all of them are ripe for being used in baiting.

Just look at a random game. The Ultimatum Game, for instance.

In the Ultimatum Game, two or more players must decide how to split a sum of money. One proposes a division, and the other has veto power. If the second player refuses this division, however, the money disappears.

Suppose that you are a minister of a church. You offer the lad a hefty sum of money in exchange for something. Your accountant has issued the ultimatum that all funds are to be dispensed after this task has been completed, and furthermore that upon any attempt to increase his compensation or receive it in advance, the lad will find that the church must undergo a lengthy regulatory audit as per new charitable donation mandates. The lad now must decide whether to take his money and run, or risk asking for a single cent in advance. Meanwhile, your poor minister implores the lad that should he ask for more, the money that otherwise would have gone toward charity will end up locked away for an indefinite period while church finances are reviewed. What will the lad do?

Or how about the Princess and Monster Game?

In the Princess and Monster game, one player is pursued by an enemy inside a defined area whose interior is dark. The player is captured when the enemy gets close enough.

Suppose that your lad is on safari. He has just reached his destination, when suddenly his email is hacked by a group of unsavory people. They know roughly where he is, and they have accomplices scattered all along the route back home. These accomplices know what the lad looks like, believe he has picked up your hefty sum of money, and are prowling each internet cafe. Your lad may be safe for an unknown time in any one cafe, but at some point they will find him. The optimal solution for your lad is to zigzag all over the place like a hummingbird. This solution will occur of its own accord, and need very little prompting. His journey home may now take much, much longer.

I want to do a series of game theoretic baits, exploring as many games and their variations as possible. In the course of this, I will need several lads, some helpful enemies, a few allies, etc. Anyone who's interested is free to work on this with me.

Posted: Thu Dec 30, 2010 3:57 pm

Sounds rather interesting, i personally have no knowledge in game theory, but i can always be a "clueless ally" Very Happy

, if you so wish

Posted: Thu Dec 30, 2010 5:19 pm

Good luck! I rather suspect your lads will just lurch after the low-hanging fruit each time - money now, even $1 against $1M tomorrow. Bottom-feeders Wink

insults and more · Posted: Thu Dec 30, 2010 6:11 pm

^^^^ Game theory allows one to systematically block off easy rewards. It is all a matter of making sure that for the victim, the expected utility of his action outweighs his other options given the expectations of the other players.

There are games in which the rational course of action calls for suicidal behaviour, and games which though all logic suggest one course, it is all but certain the average person will pick the losing maneuver.

If you want to see an example of how far from intuitive you can get, check out the Monty Hall Problem and then ask if a lad cannot be persuaded to avoid the low-hanging fruit.

Posted: Thu Dec 30, 2010 6:37 pm

In my opinion, lads are greedy, arrogant, stupid and have trouble following any specific rules other than their own. When they are 'forced' to think they become surly, agitated and quit easily. Which is inherent and why they have sought a criminal path in life rather than work hard for the meagre crust that most honest people do.
If you managed to source a lad with enough intelligence to follow this and make a cultured and intelligent decision, I would suppose they would opt for the solution that would give them the cash the soonest.
The term 'bottom feeders' is entirely apt.
I would be interested in any results that you might garner from this experiment Smile

insults and more · Posted: Thu Dec 30, 2010 6:57 pm

I'm thinking about turning this thread into a discussion of game theory as applied to baiting. If I have the time and there is interest, I'll add background information to make it more accessible.

Basically, I want to go through the basic games and the elements of game theory, explain them as best as possible in non-technical terms, and then describe how they may be incorporated into baits. Details will be kept generic enough to stay in front of the login.

As I start baits with game-theoretic elements, I'll report on them here.

Posted: Sat Jan 01, 2011 10:06 am

Hi leonsumbitches,

Thank you for raising the game theory idea. Very Happy

I am very interested in the game theory of scambaiting, so much so that it has inspired me to reactivate my account from a few years ago!

Unfortunately, I have been too busy in real life recently and haven't had the time and energy to devote to anything more than a mildly diverting straight bait Smile

or a random insane insult bait... Twisted Evil

I actually wrote the outlines of an essay on scambaiting about 18 months ago as part of a University level economics subject on the games theory.

I soon had to abandon my essay however, as it soon became apparent that I wasn't addressing the question (which was to describe a Prisoner's Dilemma game), and the game I was describing, whilst having some similar aspects, was not Prisoner's Dilemma.

My draft essay was designed more to explain scambaiting to an economist using game theory, (rather than what you are describing here, explaining game theory to scambaiters), but it might be of useful to you nonetheless.

I would be more than happy to send it to you via PM to edit as you like and possibly use for your explanations of the subject matter.

Leave the money on the fridge,

Grumblebum (aka Ted Bullpit).

insults and more · Posted: Sat Jan 01, 2011 4:48 pm

I'd love to read it!

Posted: Sat Jan 01, 2011 6:50 pm

So would i, recreational maths was once a hobby of mine, but then i found beer and girls. The beer's all dried up and so have the girls, so (sigh) back to maths then.

Posted: Wed Jan 05, 2011 12:05 am

Oooo, Game Theory. I used that back in my first years in college in some conflict research I was doing. I'd love to revisit it. Smile

insults and more · Posted: Wed Jan 05, 2011 10:07 am

Prisoner's Dilemma

This is perhaps the quintessential game. As Feynman said about the double slit experiment, every other game is like the Prisoner's Dilemma except more so.

I'll restate the classic example:

Quote:

Two suspects are arrested by the police. The police have insufficient evidence for a conviction, and, having separated the prisoners, visit each of them to offer the same deal. If one testifies for the prosecution against the other (defects) and the other remains silent (cooperates), the defector goes free and the silent accomplice receives the full 20-year sentence. If both remain silent, both prisoners are sentenced to only 3 months in jail for a minor charge. If each betrays the other, each receives a five-year sentence. Each prisoner must choose to betray the other or to remain silent. Each one is assured that the other would not know about the betrayal before the end of the investigation. How should the prisoners act?

In analyzing games, it is helpful to view them as a payoff matrix, like so:

This lets you see all the possible outcomes:

1. Both Cooperate by staying silent: both get 3 months
2. Dave Cooperates, Henry Defects: Henry goes free, Dave goes in for 20 years.
3. Henry Cooperates, Dave Defects: Dave goes free, Henry goes in for 20 years.
4. Both Defect: both get 5 years.

Now examine how each must decide, given the decision of the other. Dave knows that if he cooperates, he will go to jail for either three months or twenty years. If he defects, he can either go free or receive 5 years. No matter what Henry does, Dave will still receive a lesser sentence by Defecting than by Cooperating.

A strategy like this is called a strictly dominating one: no matter what the other player decides, one is always better off by following this strategy.

Henry is faced with the same decision, and finds that Defection is the strictly dominating strategy for him as well.

The result is that both will defect and receive 5 years in prison. Although they could have both cooperated and instead only gotten 3 months, it would have been irrational for them to do so. Hence the dilemma: both must make the wrong and more risky choice in order to get the best outcome. By rationally weighing the risks involved, they end up with a worse outcome than by being incautious and dumb.

This is the Prisoner's Dilemma in a nutshell.

By creating the right incentives, you can force someone to, by acting entirely rationally and cautiously, take much greater risks than they otherwise would by being irrational and risktaking. Or, alternatively, they find that the lazy way ends up being the most effort.

There are a ton of variants, and as Ted Bullpit noted, the line can be quite fuzzy at times.

In the Iterated Prisoner's Dilemma, the players have multiple opportunities to Cooperate or Betray. In this case, their strategies depend crucially on their knowledge of how many rounds there will be.

If both know that there will be, say, 20 rounds, then both can reason backwards from what the best decision is on the 20th round. Since in the last round it is optimal to defect without fear of retribution, both players will defect. In the second to last round, both players know the other will defect, so they both defect. The result is that both players will defect in each round.

If both do not know how many rounds there will be, interestingly, Cooperation becomes possible. In the 80's, Robert Axelrod ran a computer simulation of several competing strategies ranging from a childish Tit For Tat strategy (You defect on the last round, I defect this round) to devious Kissinger-esque strategies. As the number of rounds approached infinity (equivalent to neither player knowing how many rounds there will be), he discovered that these childishly forgiving strategies emerged as the winners.

If you wish to create a PD with your lads, therefore, you are advised to make sure both know how many rounds they will be playing. You want them to wisely take the worst course of action, not stumble into the best course (for them).

here · Posted: Wed Jan 05, 2011 11:19 am

^^^^^^^^
Feynman = all time hero