We are developing the social individualist meta-context for the future. From the very serious to the extremely frivolous... lets see what is on the mind of the Samizdata people.

Samizdata, derived from Samizdat /n. - a system of clandestine publication of banned literature in the USSR [Russ.,= self-publishing house]

Why doesn’t the CIA tell them?

There has been some discussion on the Libertarian Alliance Forum about “if they know where the weapons are, then why don’t they just tell the inspectors where to go?” I will attempt to tackle this question from a tacticians’ point of view.

Iraq is big: about the size of France and a hell of a lot emptier. There are miles of underground facilities. We can’t possibly be one hundred percent certain we’ve found everything. No matter how long the inspectors take there is uncertainty for the Searchers. However there is also uncertainty for Saddam. He can’t know what our spies have found out, if anything.

So we have a mathematical “game” with two players that might be likened to “battleship”, but is far more complex. It’s also deadly serious. There are potentially hundreds of thousands of lives at stake.

One player has assets on his hidden board and the other player is trying to find them. The second player knows where some of the assets are but can’t even be sure what percentage they know of; the other side knows all of its’ assets but can’t be sure how many of them the other side knows. This gives us a matrix of four possibilities:

  1. Searcher knows of the asset : Owner believes the Searcher knows of it.
  2. Searcher does not know of the asset : Owner believes the Searcher knows of it.
  3. Searcher knows of the asset : Owner believes the Searcher does not know of it.
  4. Searcher does not know of the asset : Owner believes the Searcher does not know of it.

What is the best strategy for each player?

The owner will be as helpful as possible on all the sites they believe the Searcher knows of. They can clean them out in advance and pretend great surprise at the inspection. The pretense also assists them in their game playing over the other three categories. They know their assumptions about some sites are “false positives” but they can’t know the percentage or which ones are in categories 2 and 3. So their best strategy is to toss some proportion of the uncertain sites in with the first group. They can hope to gain information if they can learn how many of the sacrificed assets were “known”. It might be zero, it might be all. Either way, they gain information for the next round.

The Searcher wishes to give them as little information as possible. They know that categories 2 and 4 exist and the only lever they have is category 3. If the Owner doesn’t know what the Searcher doesn’t know, they may give up more assets they believe are compromised. Some of those assets will be in category 2 and others in 3. The Searcher proves the existence of category 4 this way.

Lets look at the actual situation with Iraq. The UN inspectors have a list of known sites; Saddam will have had them cleaned up in advance. However Saddam must have evidence that some other sites may be compromised. He can’t be sure. His best bet is to force the CIA to show its’ hand. That reduces his uncertainty and gives him a better idea of which assets are safe. He can admit to whatever the CIA finds, claim oversight, apologize and claim “that’s it, there ain’t no more.”

At this point the CIA will have used its’ information up but gained no knowledge in return. They can prove nothing at all and have lost. Saddam can keep his remaining hidden weapons, safe in the knowledge they are firmly in the “we know they don’t know” category. That is the game Saddam would like to see played. He wins and eventually… we lose a city.

A better game plan for the CIA is to give out nothing, or give out only information that might elicit a response which returns useful data. The inspectors will find nothing at the “we know you know” sites; Saddam may try to guess which sites the CIA know so as to preclude getting caught, but he cannot know if he has been successful until the endgame.

Endgame comes at the end of January. The inspectors will lay out what they have found. They may have gotten some “new” sites. The CIA can then estimate how honest Saddam has been. If the inspectors are given info on “everything we know”, then either Saddam has told the truth… or he has outplayed the CIA. In this game plan the odds are in the CIA’s favour.

The existence at endgame of “they didn’t know we know” assets pretty much proves there are also “they know we don’t know ” assets. QED as they say at the end of mathematical proofs. We win and Saddam receives an all expenses paid vacation in Valhalla.

16 comments to Why doesn’t the CIA tell them?

  • Kevin Connors

    Bravo: excellent analysis, Dale.

  • Geo

    Very good treatment, Dale.

    I’d like to add a possible improvement to the game strategy, if I might.

    In this gambit, the CIA plays it essentially as you have described, but tells the UN inspectors of one of the known weapons sites to ensure a positive inspection result. This gives the US better positioning as they head into the next round.

    They trade a small informational advantage (loss of one datum) in the inspection round for significant advantages in the later prosecution round (e.g., undermining Arab-world resistance, enhancing ally support, diminishing negative voter backlash, etc.).

    By safeguarding all of the other data to which you referred, with the exception of one, they might still gain the significant advantages of your game strategy, while adding the benefits of having the UN team actually find evidence of WMD’s.

    Happy New Year!

  • Ryan Waxx

    In this gambit, the CIA plays it essentially as you have described, but tells the UN inspectors of one of the known weapons sites to ensure a positive inspection result.

    Which of course presupposes that:

    1. We have more than one asset in category 3.

    2. We KNOW it is in category 3 rather than category 1… else we look like fools, Saddam GAINS rather than loses, and possibly he figures out how we obtained the site and shuts down our source and figures out more of category 3.

    3. This one relevation will be enough to give us any advantage at all in the U.N. Given Kofi Annan’s past behavior, he’d simply swallow whatever excuses Saddam made and say ‘this isn’t enough’.

    Perversely, if we give the U.N. evidence of a meterial breach, the standard for what is a material breach will be raised higher.

    In short, you assume perfect knowledge of Saddam’s mind (cat 1 vs cat 3), totally impartial judges (the U.N?), and the U.S. not merely having evidence of a meterial breach, but OVERABUNDANCE of such evidence such that we can throw it away with no guaranteed result.

    That’s 3 assumptions too many.

  • gray1

    Unfortunately there is also the problem of the UN inspectors themselves. As I recall, there were reports that some of the inspectors had phoned ahead to some of the sites to let the Iraqis know they were coming, and there are not wholly unwarranted suspicions on the part of some cynics that the inspectors (or some of them) don’t WANT to find any evidence of WMD. If this is true giving valuable intelligence about the whereabouts of WMD will (1) not result in the verification of their existence, and (2) will undermine the credibility of the US when the hot tip is said to be false.

    The matrix does not really reflect the entire dynamics of the situation because the search must be conducted by third parties who have their own agenda .

  • Dale Amon

    The matrix does reflect the whole situation. The issues you discuss have to do with the uncertainties involved and that the two players cannot fully trust their own sides.

    In terms of the game, this is just more uncertainty that must be dealt with. Conceptually the situation is simple: but in real world play we have wheels within wheels with cross and doublecross. An attempt on my part to follow all the possible complexity would not result in understanding, only confusion for most readers.

    Chess too has simple rules: I could describe it in a short Samizdata article. And yet a description of possible strategies and tactics of play would fill many shelves if not entire stacks in a library.

  • Ryan Waxx

    I agree. This matrix is a simplification by definition. That doesn’t mean its inaccurate.

    That’s why I tried to explain why that gambit wouldn’t work in TERMS of the matrix, not as an attack on it.

  • Geo

    Certainly, as with any gambit, it would not ensure victory. That still relies on competent play to deal with the opponent’s defensive tactics.

    As for the gambit requiring three assumptions, I’m not sure I agree. I might offer countervailing considerations:

    1. Perfect knowledge of Saddam’s mind is not required. If evidence of WMD exists and intelligence shows that it is still in place and findable, then the datum behaves as if it fit into category 3 regardless of whether Saddam “knows we know” about it.

    2. Total partiality of judges is not required, though perhaps the diciest problem to deal with in this (as in most) games that require judges. As long as there is a preponderance of impartiality among the judges OR at least one influentially impartial judge OR a judge “sympathetic” to Player 1 OR another mechanism by which Player 1 can force the revelation of judicial obfuscation to the spectators of the game, then the gambit works.

    In fact, demonstrable obfuscation by the judges actually might play into an even stronger position in Round 2 for Player 1, since the judges have a less influential role in that round than they do in this one, and material advantage and spectator moral support become much more important. Which makes the gambit perhaps even more appealing.

    3. The overabundance of evidence (in order to overwhelm UN balkiness) is not a prerequisite, either. Or, if it is, it is a prerequisite for all other gambits, as well. Player 1 will likely encounter judicial balkiness no matter when the evidence surfaces, whether within Round 1, as a result of the UN searches, or in the intermission between Round 1 and Round 2, as a result of unilateral revelation by Player 1. Since this is an obstacle on the game board which will have to be surmounted regardless, it does not actually count negatively toward this gambit.

    On the contrary, it has been shown that the judges in this case are demonstrably much more balky when presented with evidence which comes from unilateral revelation than they are with that which results from their own process. Which suggests that the best way to deal with this evidence is to engage the judges’ own process in its revelation, another point in favor of the gambit.

    Of course, I take the excellent point that this gambit could have any number of counters (just as does King’s Pawn in chess) which a competent player must deal with; I’m not sure that means it necessarily “wouldn’t work in TERMS of the matrix,” however.

  • gray1

    I am sorry if I gave offense, I did not mean to do so. Given the basic assumptions laid out in the problem as stated, I have no argument with it. I think the logic is impeccable. All I was trying to say was that the actual dynamics of the “game ” can be altered by the fact that the Searcher has good reason to suspect that a search for an asset will consistently result in a set of results which consist of negatives and false negatives, and that is highly unlikely to contain any positive results, false or not. I would like to point out that if this dynamic is taken into account the net result is the same: it is foolish to give out any information about the location of WMDs.

  • Ryan Waxx

    and intelligence shows that it is still in place and findable

    That in an additional condition that fits perfectly into the ‘perfect knowledge’ category. Logically, it is proving a negative… how do you know something hasn’t been moved again?

    Its not feasable, and if it were, then it would be much simpler for Saddam to figure out who leaked the info. We would burn the source.

    2. Total partiality of judges is not required

    Total impartiality is a claim I never made, and would have been a fool to try to make. Strawman.

    As long as there is a preponderance of impartiality among the judges OR at least one influentially impartial judge OR a judge “sympathetic” to Player 1 OR another mechanism by which Player 1 can force the revelation of judicial obfuscation to the spectators of the game, then the gambit works.

    You assume that even if a judge is convinced by logic, that that would change their minds. I disagree. I beleive it is possible and likely that even if every single judge were personally convinced that Saddam has WMD, they would vote “no” to disarming him.

    The only relevant scenario you put forth is the last one, revealing the judicial obfuscation to the spectators.

    Unfortuately, a straightforward accusation like that would effectively destroy the U.N., if proven. We have to dance around and pretend that France isn’t in Saddam’s pocket and that Kofi would support us if only we mustered enough ‘proof’.

    The point is, that your gambit would gain us nothing, because the truth does not matter one bit to the U.N. Not one shred, not one particle, not one molecule.

    The only thing that motivated them to write the resolution in the first place was the credible threat that we’d go ahead anyway if they didn’t.

    3. The overabundance of evidence (in order to overwhelm UN balkiness) is not a prerequisite, either.

    You completely and utterly miss my point. We need an overabundance of proof so that we have some left AFTER we flush our gambit down the U.N. toilet.

    That means we need at least double what it would take to convince an average Joe: Half for the U.N. to consult with Saddam on how they want to cover it up, the other half after its become clear that we didn’t get anything out of the revelation.

  • Geo

    Now, let me see if I can herd these little ducks back into some semblance of a row…

    1. I proposed a scenario.

    2. Ryan countered my scenario as inadequate, offering three putative assumptions he felt it made. One of these was “totally impartial judges.”

    “In short, you assume perfect knowledge of Saddam’s mind (cat 1 vs cat 3), totally impartial judges (the U.N?)…

    3. I then noted that, on the contrary, this assumption was never made in my scenario and in fact, is not required by it.

    Total partiality of judges is not required…

    4. Ryan countered my correction by saying:

    Total impartiality is a claim I never made, and would have been a fool to try to make. Strawman.

    Hmmm. As I understand it, a strawman argument is the attempt to replace an actual assertion with something more extreme and therefore more easily defeated in logical discourse.

    However, Ryan, your statement demonstrates that you DID explicitly say that my scenario assumed total impartiality of the judges. So my correction of your assertion was a response directed specifically and solely at a comment you explicitly and erroneously made. It was clearly not directed at some assertion which I invented for you.

    Therefore, your assertion of “strawman” is erroneous, QED.

    Cheers!

  • Geo

    In response to my qualifier, “and intelligence shows that it is still in place and findable”, Ryan notes:

    That in an additional condition that fits perfectly into the ‘perfect knowledge’ category. Logically, it is proving a negative… how do you know something hasn’t been moved again?

    RE perfect knowledge: It is not necessary to have perfect knowledge to have a high degree of certainty of a thing. One might, for example, dedicate an intelligence satellite to tracking this “something” before the inspectors are sent. Even if it is moved it can therefore be tracked and “perfect knowledge” that it has not been moved is moot. “Proving a negative” in this case is a red herring, since what we are actually doing is simply continuing to collect positive information.

    Its not feasable, and if it were, then it would be much simpler for Saddam to figure out who leaked the info. We would burn the source.

    Well, where to begin? This pithy statement has at least three errors of logic: it makes unsupported assumptions, it is a strawman and it is fundamentally an illogical construct based upon irrealis conditions.

    As for unsupported assumptions, you are presuming that you know clearly the feasibility of the premise, when you and I and everyone else in this forum clearly do not, as we are outsiders ignorant of the particulars. (Unless you happen to be with the CIA, in which case I swiftly withdraw my assertion and wish you a very pleasant day, sir, while urging you to read no more of my post.)

    Moreover, the statement is a strawman, in that it argues against a position which I never proposed. There are certainly sources of information which do not require “Saddam to figure out who leaked the info” and therefore would not necessarily require us to “burn the source.” Satellite imagery of site construction, weapons movements and so forth is an example of a possible source which is not considered in your rebuttal. And the world already knows the US possesses this technology, though perhaps not its degree of development. Producing a sat photo at a somewhat lower res than we are capable of, but still shows the evidence (if evidence there is) meets the criteria while preserving the source.

    And finally, the statement “It’s not feasible, and if it were, then it would be…” is fundamentally an illogical construct. If you are correct and it is indeed “not feasible,” then the qualifier “and if it were” speaks to an irrealis situation. That is, you are projecting an assertion of your certitude onto a condition which does not exist. If it did exist, one might conceivably have various other conditions altered as well, which could make the assumptions inherent in our discourse no longer valid. Note my conditional use of “might” and “could” which is the logically sound way to address irrealis conditions.

    Thus, your objection to my scenario on these bases seems to be completely illogical, q.e.d.

    There are other statements you make which are similarly questionable. Such as:

    We need an overabundance of proof so that we have some left AFTER we flush our gambit down the U.N. toilet.

    Which is pure circular reasoning: it makes the fallacy of assuming your conclusion (i.e., the failure of the gambit) as supporting argument for your conclusion (i.e., the gambit would fail).

    And:

    …the truth does not matter one bit to the U.N. Not one shred, not one particle, not one molecule.

    Which also sports three errors of logic in that it is grossly unsupported, an appeal to emotion and, most importantly, quite irrelevant, since I have already proposed scenarios in which obfuscation by the judges does not undermine the gambit.

    Well, it is enough. As enjoyable as wrangling with logic is to me, I simply must get on to other things. More critically, I don’t think the folks in this forum really care to wade through it all.

    Allow me, then, simply to respectfully disagree with you and retire from the field to lick my wounds in ignominy.

  • Ryan Waxx

    One might, for example, dedicate an intelligence satellite to tracking this “something” before the inspectors are sent.

    Yep. Just dedicate a satellite to watching it. Oh, and follow it if its moved. Uh Huh.

    Gee… you don’t think that maybe its kinda difficult to know everything from satellite? I mean, all we’d have to do to kill Osama would be have the satellite track him! Or better yet, we don’t need anyone in Itraq, we can just track all the WMD from orbit!

    You, sir, haven’t the faintest clue as to the limitations of satellites(most intelligence ones are not geostationary for example), and I rather suspect your ignorance extends to all other areas of intelligence-gathering.

    Moreover, the statement is a strawman, in that it argues against a position which I never proposed. There are certainly sources of information which do not require “Saddam to figure out who leaked the info” and therefore would not necessarily require us to “burn the source.”

    Horse frigging crap. You can catcall ‘strawman’ all you like, but you CANNOT disguise the fact that knowing where something is AND that it has not been moved is a MUCH more formidable feat than knowing where it was moved to initially.

    And the people watching the WMD is necessarily a much smaller group than the people who moved it. Which also means that the source is part of a smaller group (unless we use one source to locate, another to watch… which assumes a huge overabundance of intel assets)… A smaller group of possible moles means its much easier to catch them. And THAT is why revealing that we know where something is, AND that we know it hasn’t been moved since, will burn the source.

    Think it through, and things become that much clearer.

    Thus, your objection to my scenario on these bases seems to be completely illogical, q.e.d.

    OIC… so my sentence contruction makes the arguement invalid? Too many “certainly’s” prove that the opposite is true?

    Fine. You are certainly, unquestionably, proven by the rising of the sun in the south, NOT a frigging moron.

    When you figure out that statement call me. We’ll up you to Gerber instead of breast milk.

    Which also sports three errors of logic in that it is grossly unsupported, an appeal to emotion and, most importantly, quite irrelevant, since I have already proposed scenarios in which obfuscation by the judges does not undermine the gambit.

    You proposed exactly ONE scenario in which obfuscation by the judges does not undermine the gambit, and you know it. And as I said, it involves directly attacking the U.N. and therefore won’t happen.

    Proving a negative” in this case is a red herring, since what we are actually doing is simply continuing to collect positive information.

    That is a lie, and I suspect you KNOW it’s a lie. Knowing for sure something hasn’t been moved is not “simply continuing to collect positive information”, unless you plan on having our spies drop in once a week and see if the stuff is still there.

    “Hey… Ahmed… still got those 20 canisters of mustard gas in the basement?”

    *BANG!*

    It is the assumption that Saddam CANNOT move anything once found without our noticing, in the middle of a hostile country… and franky that assumption is deeply ignorant.

  • Dale Amon

    STOP.

    Discourse on this forum will *not* include ad hominem, name calling, or anything other than collegial disagreement and honest debate.

    Ryan, you have some good points. You have no need for this.

  • Ryan Waxx

    Fine. Here’s a non-namecalling, ad hominem, collegial thingamabob:

    How exactly will your ‘gambit’ handle WMD that’s moved into Syria? I’d like to see the spy that crosses the borders back and forth without comment, but I suspect he only lives in 007 movie clones.

    We are fairly certain that Saddam is moving normal weapons to Syria… why not WMD, as well?

    And remember: Satellites are not the eye of God himself…

  • Ryan Waxx

    Dale Amon put forth a game-theory analysis on weather it makes sense to UPS our intel to the inspectors. He put forth a matrix of possilbilities, which you can see above. This matrix seems to have been created from Saddam’s perspective, as it does not deal with what we believe, only what saddam believes we know.

    Geo suggested that we trade one ‘datum’ from category 3 (Searcher knows of the asset : Owner believes the Searcher does not know of it.), for diplomatic advantage.

    Part of my difficulty with this idea is that the definition of category 3 is not sufficient in and of itself for this purpose, since the matrix does not take into account what the searcher believes.

    The datum that may be used for this gambit I will call category 3a, for lack of a better term:

    Searcher knows of the asset : Owner believes the Searcher does not know of it. Searcher knows that the Owner believes the Searcher does not know of it, and Searcher is relatively certain that the asset has not been moved.

    Because if the asset we are are prepared to sacrifice is really 1 rather than 3 and therefore covered up, or it has been moved for any other reason, then we lose diplomatic advantage AND we quite possibly lose the intel source if it is human (we do not lose the WMD location itself, because we never had it… or rather we were mistaken).

    May I humbly submit that 3a is a much more restrictive category than 3, because it requires not only that we have intel regarding WMD locations, but that that intel is of such high quality that we know which category the WMD is actually in. Its still possible, but much tougher, and tougher-to-get intel is nearly ALWAYS much easier to trace back and burn.

    Logical?

  • Geo

    If you will permit me, Ryan, I would tend to agree with your statement:

    “May I humbly submit that 3a is a much more restrictive category than 3, because it requires not only that we have intel regarding WMD locations, but that that intel is of such high quality that we know which category the WMD is actually in.”

    I also believe it is quite possible to have data in that category. I do not know that we do so, in fact. Presumably, neither do you.

    But others might and, if they do, I tend to disagree with you on how they might use that information to their best advantage. Of course, as neither of us know the relevant factors as insiders, both of our positions are unsupported by the best information.

    In that spirit, my proposed gambit was meant only to recognize both the cleverness of the original matrix proposed by Dale, while suggesting one possible wrinkle in game play. It certainly was not meant to proselytize on the best political strategy nor to suggest that I am an expert in such matters. In this, you see, I readily admit I am on “Gerber.”

    I regret that somewhere along the line I may have indicated to you otherwise and encouraged some invective in the process.

    Truce?