Welcome! These forums will be deactivated by the end of this year. The conversation continues in a new morph over on Discord! Please join us there for a more active conversation and the occasional opportunity to ask developers questions directly! Go to the PS+ Discord Server.

The theory of blackmail

9 posts / 0 new
Last post
Arenamontanus Arenamontanus's picture
The theory of blackmail
I came across this nice little article in Slate http://www.slate.com/articles/arts/the_undercover_economist/2008/10/the_... on why it is hard to make blackmail work. It cites this classic paper on the topic by Daniel Ellsberg http://www.rand.org/pubs/papers/P3883.html (and it all began when I passed by tvtropes page http://tvtropes.org/pmwiki/pmwiki.php/Main/Blackmail of course.) In Eclipse Phase the combination of surveillance, a reputation economy and long lives make blackmail a really interesting activity.
Extropian
Smokeskin Smokeskin's picture
Re: The theory of blackmail
Another problem with blackmail is that if it is based on information (as opposed to say a physical piece of forensic evidence), you can't really buy it. So what you have is a repeat game, and at each step the payoff matrix looks exactly the same. What is very interesting is that typically, if the victim refuses to pay, the blackmailer either has the same payoff from revealing or hiding the secret, or even a cost associated with revealing (black mail being illegal, involvement in the secret, social repercussions from being a tell, having kept it secret for long, etc.). The two types you don't want to deal with are professional black mailers (who'll have a positive payoff from a reputation of revealing when he deals with future victims) or emotional black mailers (who would gain pleasure from revealing your secret). If you want to black mail, you should try to demonstrate you are this kind of black mailer, true or not. The other thing to consider is that if you pay, the black mailers pay off matrix still looks the same afterwards, so you'll have to pay again. This is actually what the victim wants - the black mailer's expectation of future payments is what keeps him from revealing your secret. As a victim, just be sure to only pay in a way that the net present value of your future payments doesn't exceed the value of the secret. The math becomes interesting for secrets whose value change over time, where you'll as always have to look forward and reason backwards. The simplest case is a secret of constant worth until it expires at a certain date. Say you have a secret that becomes irrelevant in 10 years (say its a crime with a statute of limitations). Just before the secret expires, you'll be willing to pay the full worth, say 1,000 credits. A year before that, your blackmailer will approach you and ask for payment. You know you'll pay 1,000 credits in a year if your secret isn't revealed, and with say a 10% discount rate, that makes it worth ~900 credits now. At this point in time, your secret is also worth 1,000 credits to you, so you can be blackmailed for maxium 100 credits that year. And so on for previous years - what it works out to is that maximum amount you can be blackmailed is exactly like a loan with out amortization. Of course, when negotiating the black mail price, you need to consider the blackmailer's payoffs - no one says you should accept the worst possible situation for you. If black mailing is legal and you can set up binding agreements (or if there are illegal elements willing to enforce such an agreement), things become much simpler, as you can simply change the payoff matrices with contracts to avoid the repeat game complications. Provoking idea: Black mailing is a cost on secrecy. Black mailing being illegal and immoral is a cost on black mailing, meaning it is less profitable and so fewer will engage in it and victims will be squeezed for less. Thus opposition to black mailing makes secrecy CHEAPER, which in turn makes activities that depend on secrecy more profitable, like crime and socially unacceptable activities (and for activities that are socially unacceptable due to bias, legal black mailing would mean it would be comparatively more attractive to step out into the open and combat the bias - so laws against black mailing help perpetuate prejudice). Imagine that the cost of commiting crimes with a partner, cheating on your spouse and being a closet atheist or homosexual (including in the US) went dramatically up - would that be a good thing?
Decivre Decivre's picture
Re: The theory of blackmail
The inherent problem with blackmail is that it's a contradiction; a business transaction built on distrust. Neither party has any reason to assume the other will keep their end of the bargain, which means that they have no reason to keep their end of the bargain. It's like a prisoner's dilemma in inverse. So the real question becomes this: how do you violate a person's trust to blackmail them, then get them to trust you enough that they will be willing to honor their end of the coercion?
Transhumans will one day be the Luddites of the posthuman age. [url=http://bit.ly/2p3wk7c]Help me get my gaming fix, if you want.[/url]
Smokeskin Smokeskin's picture
Re: The theory of blackmail
Decivre wrote:
The inherent problem with blackmail is that it's a contradiction; a business transaction built on distrust. Neither party has any reason to assume the other will keep their end of the bargain, which means that they have no reason to keep their end of the bargain. It's like a prisoner's dilemma in inverse.
If it is about trust, of course it'll fail as everyone knows the blackmailer will come back with more demands, but since you can simply use game theory and look at it as a repeat game, trust isn't an issue. It actually comes down to the opposite, the blackmailer has a challenge making his threat convincing. What you describe sounds like the prisoner's dilemma, not its inverse, and certainly not a blackmail situation.
Decivre Decivre's picture
Re: The theory of blackmail
Smokeskin wrote:
If it is about trust, of course it'll fail as everyone knows the blackmailer will come back with more demands, but since you can simply use game theory and look at it as a repeat game, trust isn't an issue. It actually comes down to the opposite, the blackmailer has a challenge making his threat convincing. What you describe sounds like the prisoner's dilemma, not its inverse, and certainly not a blackmail situation.
It really depends on the blackmail in question. Trust is still important because if the mark has no reason to assume you will honor the ransom, they have no incentive to give in to your demands. A convincing threat is worthless if the victim knows that the carrot and stick is an illusion. I suppose it really is comparable to the prisoner's dilemma. The optimal scenario for both parties is that both will honor their end in the bargain. The optimal scenario for either party individually is that the other will honor their end of the bargain, while they can renege on their end. The likely scenario is that neither party will trust the other, and no one will end up happy with the outcome.
Transhumans will one day be the Luddites of the posthuman age. [url=http://bit.ly/2p3wk7c]Help me get my gaming fix, if you want.[/url]
Smokeskin Smokeskin's picture
Re: The theory of blackmail
Decivre wrote:
It really depends on the blackmail in question. Trust is still important because if the mark has no reason to assume you will honor the ransom, they have no incentive to give in to your demands. A convincing threat is worthless if the victim knows that the carrot and stick is an illusion.
I'll try to say it again. If it depends on trust it will fail. But as a repeat game with continued payments, you don't need trust. The blackmailer will come back for more money rather than reveal the secret. The real issue that the blackmailer's payoff from revealing the secret is mostly very small or negative, so he'll have a challenge convincing the victim to pay much.
Quote:
I suppose it really is comparable to the prisoner's dilemma. The optimal scenario for both parties is that both will honor their end in the bargain. The optimal scenario for either party individually is that the other will honor their end of the bargain, while they can renege on their end. The likely scenario is that neither party will trust the other, and no one will end up happy with the outcome.
The best play, no matter what the other prisoner does, is to defect. The game is extremely simple with obvious dominant strategies. Blackmail is a repeat game most often with posturing as the deciding element and the payoff matrices for the players are very dissimilar. I don't see how they're the inverse of eachother.
Decivre Decivre's picture
Re: The theory of blackmail
Smokeskin wrote:
I'll try to say it again. If it depends on trust it will fail. But as a repeat game with continued payments, you don't need trust. The blackmailer will come back for more money rather than reveal the secret. The real issue that the blackmailer's payoff from revealing the secret is mostly very small or negative, so he'll have a challenge convincing the victim to pay much.
And how exactly will you get that small payoff if your victim has no reason to assume you will honor your end of the bargain? Hell, let's even assume that there's no ransom. How will a person who has a secret to maintain not work in every functional way to destroy someone to protect that secret, should they have every reason to assume that person is guaranteed to reveal it? Blackmail (for that matter every transaction, criminal or otherwise) requires some degree of trust. Otherwise, you're just baiting your victim into coming after you.
Smokeskin wrote:
The best play, no matter what the other prisoner does, is to defect. The game is extremely simple with obvious dominant strategies. Blackmail is a repeat game most often with posturing as the deciding element and the payoff matrices for the players are very dissimilar. I don't see how they're the inverse of eachother.
I corrected myself in that last statement, and instead compared it to the prisoner's dilemma directly rather than an inversion of it. I'm sorry if I did not make that clear.
Transhumans will one day be the Luddites of the posthuman age. [url=http://bit.ly/2p3wk7c]Help me get my gaming fix, if you want.[/url]
Smokeskin Smokeskin's picture
Re: The theory of blackmail
Decivre wrote:
Smokeskin wrote:
I'll try to say it again. If it depends on trust it will fail. But as a repeat game with continued payments, you don't need trust. The blackmailer will come back for more money rather than reveal the secret. The real issue that the blackmailer's payoff from revealing the secret is mostly very small or negative, so he'll have a challenge convincing the victim to pay much.
And how exactly will you get that small payoff if your victim has no reason to assume you will honor your end of the bargain?
Why do I have to say the same thing over and over? The blackmailer keeps coming back for more money, so he's not interested in the secret getting out.
Quote:
Hell, let's even assume that there's no ransom. How will a person who has a secret to maintain not work in every functional way to destroy someone to protect that secret, should they have every reason to assume that person is guaranteed to reveal it?
destroying someone will often be more costly than the secret getting out, or the blackmailer has a dead man switch
Quote:
Blackmail (for that matter every transaction, criminal or otherwise) requires some degree of trust. Otherwise, you're just baiting your victim into coming after you.
It is very difficult to discuss game theory when you don't understand it. Trust doesn't have to factor into it. The whole analysis is done assuming no trust (and if there were trust, you'd just factor it into the payoff values anyway).
Quote:
Smokeskin wrote:
The best play, no matter what the other prisoner does, is to defect. The game is extremely simple with obvious dominant strategies. Blackmail is a repeat game most often with posturing as the deciding element and the payoff matrices for the players are very dissimilar. I don't see how they're the inverse of eachother.
I corrected myself in that last statement, and instead compared it to the prisoner's dilemma directly rather than an inversion of it. I'm sorry if I did not make that clear.
They don't have similarities either. Draw up the normal form matrices and you'll see. I use game theory regularly and I still have to draw them, you can't skip it and expect to get it right. Look at how you went from inverse to same - intuition sucks for this.
Decivre Decivre's picture
Re: The theory of blackmail
Smokeskin wrote:
Why do I have to say the same thing over and over? The blackmailer keeps coming back for more money, so he's not interested in the secret getting out.
What does motive have to do with perception? The question is whether the victim [i]perceives[/i] the blackmailer as honoring their end, not whether they actually will or not. That takes trust, whether you want to admit it or not. It's the same issue in kidnapping. A kidnapper does not necessarily want to kill the victim until they have received the ransom. They still have to gain enough trust from the payer that they will believe the kidnapper can honor their end of the deal and return the victim safely... otherwise, no payment will be made.
Smokeskin wrote:
It is very difficult to discuss game theory when you don't understand it. Trust doesn't have to factor into it. The whole analysis is done assuming no trust (and if there were trust, you'd just factor it into the payoff values anyway).
Game theory isn't necessarily an accurate way to gauge strategy in this scenario. Blackmail isn't a game or even battlefield for victims... it's an emotional issue. Psychology plays a severe factor. Most people, when under the emotional duress of blackmail, are probably not going to think of the whole scenario in terms of game theory. Therein is your error. Humans are not cold and emotionless machines (they are warm and emotional machines), and your assumptions fail to take that into account. Trust might not be necessary when talking game theory, but we're talking blackmail... game theory can be used to analyze the scenario, but game theory isn't the focus of this conversation nor should it be assumed the only means of looking at the dilemma.
Smokeskin wrote:
They don't have similarities either. Draw up the normal form matrices and you'll see. I use game theory regularly and I still have to draw them, you can't skip it and expect to get it right. Look at how you went from inverse to same - intuition sucks for this.
Yes, I see what you mean now. The decision matrix is a bit wider for blackmail, and there are different optimal strategies.
Transhumans will one day be the Luddites of the posthuman age. [url=http://bit.ly/2p3wk7c]Help me get my gaming fix, if you want.[/url]