Assorted links

1. Markets in everything: condoms to get you pregnant.  Bonus prize for those who can guess the context without reading the article.

2. The popularity of in-game markets.

3. Employment at Apple stores, a good piece.

4.  What predicts Muslim anti-Americanism?, and Madonna markets in everything?

5. Jerry Brito on Stuxnet, though it is stopping nukes that could encourage bioweapons.

6. Forget the September result!


I suspect the difficulty in setting up in game markets, is one of balance. It's quite difficult to set up a reward system that's random and rewarding enough to trigger the Skinner Box effect (which is the source of the major enjoyment of economy type games) while balancing it so that every portion of the game is challenging at appropriate levels, and offers rewards that scale with the players skills without changing in the rewards/time.

Not to mention that information about any imbalance in your system will quickly propogate to the biggest exploiters of the imbalance first.

The claims in the article about the Diablo 3 economy are pretty much false.
You can buy gear of any level for any budget in the AH, not just "starting gear", which is also generally not overpriced (at least on EU AH, but I doubt US is much different).

Of course there's a $1.25 dollar minimum price on real-money AH, so if you try to buy starting or very low-end gear on it it will be indeed somewhat overpriced, but there's no need to do so, since items found by the player during gameplay and the gold AH work fine to get such gear.

There have been gold inflation issues indeed, which they attempted to fight by making people lose 6x as much gold when they die; it's still early to say how successful that change was.

At least in MMOs, inflation is the main problem with game economies:

Players earn currency through playing the game (rewards from doing quests, looted directly from monsters or selling items looted from monsters,) think of this as their wages, and income through this is directly related to time spent playing. As players get higher leveled, the amount of currency they earn for each monster killed or quest completed increases.

This money is coined fresh when the player earns it, just like a government paying wages through printing money.

Players also sell items to other players, through things like the auction house in Diablo or WoW, or through their own "stores" like in Aion or Ragnarok Online. (These items are both acquired by killing monsters and created by players themselves, using items acquired by killing monsters as inputs.)

The game takes money out of the economy through different mechanics, like causing players to lose money when they die or requiring items bought by NPC (non-player character, ingame agents) vendors to continue playing. Examples go from archer-type characters needing to buy arrows in order to do damage, to consumables required to cast spells or increase your character's abilities, to WoW requiring you to repair your armor after you die or else it will break, to simply taking money away when your character dies.

There are two problems: 1.) The net outflow of money has to be lower than the net inflow or players will not be able to sell items to each other, liquidity is necessary. 2.) As the game grows in population and as the average character level increases, more currency will be created through the playing process. However, prices of required purchases must remain stable or you will foreclose new players, so this money will flow into the player to player transactions, increasing prices in those items.

I'm not sure how to deal with this. As far as increasing monetary returns to higher level monsters, even if every monster in the game was worth a constant amount of money, higher level characters would be able to more low level monsters in the same period of time, leading them to making more money, so more money will enter the economy as the player population progresses. Perhaps the game could create a central bank with both assets and liabilities, limiting the amount of currency available to players: Items would sell for less to vendors and quests would pay less as more currency collects in players' hands, but then wouldn't you have liquidity problems, especially for player to player transactions? Perhaps you can provide ingame banks that provide loans to provide non-inflationary liquidity? (The fact that player-created banks do not exist in most MMOs is a strong argument against anarcho-capitalist optimism about the state of the world absent any government: There are no ways to enforce contracts outside of social pressure in a game, and lending to anyone outside your circle of friends is impossibly risky.)

I actually expect pretty strong deflation in the Diablo III auction house for durable equipment since, unlike most other MMOs, gear does not bind to characters. In Blizzard's other big game, World of Warcraft, any non-trivial equipment will become "soulbound" to the first character that uses it and therefore cannot be traded or sold. This limits the supply to newly-found equipment. From what I understand, gear doesn't bind to characters in Diablo III, so equipment that will be replaced in a few levels can be resold, so the supply will be ever-increasing, and my guess is that it will out pace the gold supply.

I think the most interesting aspect is the stimulus package of Diablo III known as Inferno Difficulty. By the time you reach it your characters' skills are already long ago maxed out, you've run though the story line three times, and done everything there is to do. They've made an extra level that is so hard you'll feel obligated to go shopping to be able to compete even though there is nothing new there to do. Very clever way to make money, but how could they be certain people would be interested in dying so much that they would get frustrated enough to pay up?

1. I was sure this was going to be the NBA.

I, too, failed to win the bonus prize.

I'm still not sure how these condoms are to get me pregnant.

Why does anti-American rhetoric resonate with the publics these elites appeal to, though? Perhaps countries that are firmly Islamist see no threat to their arrangements and thus have no reason to be anti-American, while countries that are firmly secular don't mind being secular and have no reason for anti-Americanism. Secularists in a divided country might prefer to divert people's anti-secularism against America so they have a freer hand at home.

The article is limited to anti-Americanism in Muslim countries, pushing aside the anti-Americanism prevalent in Latin America and Asia. Perhaps the reason it resonates with the publics should be more mundane and generic than simply secular vs islamist. And by comparing with anti-British sentiments during 19th century, this feeling is even more generic.

Regarding the September result :

"I think a lot of academic papers are just low-value rent-seeking data-massaging whose only real purpose is to justify more low-value rent-seeking data-massaging."

The blogger killed the paper in three paragraphs. Brilliant and very simple. I wish people looked at data with common sense before theorizing. Also, if they new these facts and hid them, then the authors have committed a crime worse than simply lacking common sense.

you really buried this retraction. these guys should be shamed.
(and you should be ashamed)

I am an Orthodox Jew and I follow the rules, but I also guessed wrong about #1.


Why does he assume nobody died from Stuxnet? Obviously details are thin on the ground, but isn't it thought that Stuxnet makes centrifuges break violently? They're high-speed rotating machinery processing radioactive materials - I wouldn't want to be around when one came apart. Who got the job of cleaning up the radioactive mess? Surely Iran would hide this information, as well.

Also, it's difficult to believe there aren't periodic executions in the Iranian nuclear program. One expects they would have tried very hard to determine who brought the malware in. It's not completely obvious that the Iranian scientists getting assassinated are being killed by a _foreign_ government. Again, Iran would have reason to hide this information, or mislead.

It's fair to say that the death toll was likely small, and probably consisted of technicians and scientists working on the program, and that it may have delayed deaths from a more thorough attack by as much as a couple of years.

Stuxnet attempted to make centrifuges break invisibly, not violently.

Is there not a video of one of the national labs testing malware that caused a large electrical generator to effectively explode (it flies apart at high speed) by performing the sort of resonance tricks Stuxnet employs?

Obviously it's not public exactly what effect Stuxnet had. My understanding was it physically broke the centrifuges, it didn't just pollute the output or turn them off.

That said, if I knew exactly what Stuxnet did, I wouldn't be out posting about it on the internet.

"My understanding was it physically broke the centrifuges"

No, it burned out parts invisibly. It was very insidious.

I just don't think that's physically plausible in a device like this. If parts come into contact with each other, all the energy in the device will dissipate in a fraction of a second. That will tear metal and throw it around.

See, for example,

"The resulting excessive centrifugal force caused the aluminum tubes to expand, increasing the risk of parts coming into contact with one another and thereby destroying the centrifuges."

Clearly, it wasn't invisible breaking.

"Increasing the risk" is not the same as breaking. From what I read the goal was to make it simply not work while falsely reporting that it was working.

It's a big, heavy, metal object spinning at kilohertz. If parts come into contact with one-another, the result will be violent. In the video (which admittedly, I can't find), an electrical generator - a similar device, although lower precision and hence with greater clearances - comes violently apart when this happens. A layman would call it an explosion. The centrifuges are also filled with radioactive and chemically nasty material, so it's not just shrapnel to worry about.

The article I linked to called the centrifuges destroyed, not silently nonperforming. Wikipedia calls the centrifuges "destroyed" and refers to a "serious nuclear accident." Besides, the mechanism for silent non-performance is not obvious, while the mechanism for destruction is straightforward.

See also:

"Those first small-scale tests were surprisingly successful: the bug invaded the computers, lurking for days or weeks, before sending instructions to speed them up or slow them down so suddenly that their delicate parts, spinning at supersonic speeds, self-destructed. After several false starts, it worked. One day, toward the end of Mr. Bush’s term, the rubble of a centrifuge was spread out on the conference table in the Situation Room, proof of the potential power of a cyberweapon."

One thing I'm unclear on is exactly how big the centrifuges were. My impression is they are usually large - in photos I've seen of US facilities, they are large. But I imagine that's a design choice.

"Clearly, it wasn’t invisible breaking."

Stuxnet was specifically designed to be undetectable for a long period of time. The centrifuges didn't explode. They were run at non-optimal frequencies which wore out the ball bearings and other components. It's highly doubtful anybody was killed. Life != Hollywood

What ball bearings? These centrifuges float on magnets don't they? The point is not to wear out bearings but to ruin the centrifuges. They are spinning very fast. Even if you wore out the bearings, if you do not notice they are about to fail, the failure would be spectacular.

We are talking about aluminum tubes that are spinning very fast indeed. They contain uranium hexafluride - an extremely nasty gas that will eat its way through most things - including most metals. It is not good for human health.

Any unexpected failure would be catastrophic. You would have an explosion (or extremely rapid disassembly if you like) that spread a highly toxic corrosive gas all over the facility. Which reacts violently with water by the way. And throws pieces of metal at, I would guess, something like the speed of sound, all over the place.

It is hard for me to believe that this virus has not killed respectable numbers of people - anyone in the facility when one of them failed for instance.

I think So Much For Subtlety has it right. Think "hand grenade going off leaving poisonous radioactive mess behind." If I am interpreting the publicly available photos correctly, the devices that exploded are man-sized, which is a little smaller than photos I've seen of similar devices at Oak Ridge, which were dated.

Stuxnet was designed to be undetectable because it kept no record of the RPM excursions that caused the explosion, not because it didn't cause the centrifuges to explode.

Why do people have the crazy idea that it just subtly wore things out? Does that make any mechanical sense? Why do all the news stories talk about them ripping themselves to pieces? They showed Bush rubble, not worn bearings. For what it's worth, I compute about 50 kilojoules of energy in a 250 kg cylinder of radius 20 centimeters spinning at 1500 RPM. I'm guessing at mass and dimensions from the photos. A hand grenade is around 200 kilojoules of energy. Look, if people are trying to nitpick the explanation, I know a centrifuge won't detonate and distribute a fine pattern of fragments - it will tear itself into big, fast moving pieces. It will look a little different from a classic explosion, but you really, really wouldn't want to be standing near it.

The best argument for nobody being killed is that there probably aren't people standing next to the devices regularly. There's still the cleanup crews, however, which surely was a life-threatening job.

Here's an article on the technology that talks a bit about fragment risks to nearby devices in the event of failure:

One reason to expect big chunks of flying metal rather than small fragments is that most articles say the Iranians were using aluminum centrifuges, and aluminum tends to tear rather than fragment. Strength of materials is the limiting factor in the performance of these devices - they operate highly stressed.

One comment about the cleanup, though, is that assuming they have rapid shutoff valves, there's not actually that much uranium hexafluoride in one of these devices at one time. IIRC, for reasons of balance they need to work hard to make sure the gas doesn't condense, so they need to keep the gas pressure very low.

I may have made an error in my calculation. Although most articles about Stuxnet reference it attacking machines and rotating them in the range of about 1400 Hz, I used rpm in my calculation of energy. My numbers should be scaled up by a factor of 60. I'm really guessing at the mass and radius. If I had to bet I'd say I might be low on the mass and high on the radius.

So there's a great deal more energy than a hand grenade in one of these devices, then.

> IIRC, for reasons of balance they need to work hard to make sure the gas doesn’t condense, so they need
> to keep the gas pressure very low.

This suggests the mechanism for inducing the wobble. Turn the speed up to get the gas closer to condensing, and then pulse the power to induce a standing wave in the gas. The resulting few grams sloshing around (and maybe condensing at the high pressure point of the wave) would upset the dynamic balance of the rotating section and cause a whirl. This pattern is just what Stuxnet applies: it periodically turns the motor controller to 1410 Hz and then to 2 Hz and then to 1064 Hz (which is around normal operating speed). It then erases any record of having done so. So you have to figure this doesn't work every time. But when it does, it's like a giant washing machine, running a thousand times as fast, when the clothes load gets unbalanced. The centrifuge disassembles itself and ruins the machines in the immediate vicinity with fragments and shock, and releases a (probably small) cloud of toxic gas.

"ultra-Orthodox men aren’t allowed to masturbate"

Is this right?!

Yes, Stuxnet is an act of war. However, it won't matter. Iran has officially been at war with the US since the 70's.

Also, China commits cyber-acts of war against the US, constantly.

"Sabotage I’m less certain about, but I think it inhabits a similar space as espionage: frowned up, prosecuted criminally, but not an act of war per se." --- I am not sure what to make of this statement. If there is sabotage of a public transport system or of the water distribution system in a city of a major western power (and assuming there is no or very little loss of life), would the said power say, ho hum just another thing that we need to frown upon or would it go all guns blazing at the perpetrators? I suspect it would be the latter.

I would definitely like someone with some idea of International law to weigh in on this discussion. It just seems that Stuxnet is indeed an act of war, at least to my untutored mind.

"International law" is an even more self-conflicting term than "military intelligence". In the end, these treaties are gentleman's agreements. They only work so long as everyone in the room decides to follow them. These Islamic groups have a different set of rules.

I would argue that Stuxnet did far less damage to Iran than Russia's assault on Latvia did, and what was done?

The argument over "no one was killed" is silly. At the national level, countries often overlook small numbers of killed, even if deliberate. After all, in a war, tens of thousands WILL die.

But, as has been noted, Iran declared war on the US in 1979. Under the mores of Western "International Law", a state of war still exists. By my lights, our President retains the constitutional authority to order a full-scale invasion at any time. Not that I'm happy about that.

Comments for this post are closed