Betting on Black Swans(immadsnewworld.com)
immadsnewworld.com
Betting on Black Swans
http://www.immadsnewworld.com/2012/12/betting-on-black-swans.html
5 comments
Malcolm Gladwell outlined the Black Swan strategy Taleb used at Empirica in this New Yorker article from way back in 2002:
http://www.gladwell.com/2002/2002_04_29_a_blowingup.htm
Here's the part that resonated with me:
"We cannot blow up, we can only bleed to death," Taleb says, and bleeding to death, absorbing the pain of steady losses, is precisely what human beings are hardwired to avoid. "Say you've got a guy who is long on Russian bonds," Savery says. "He's making money every day. One day, lightning strikes and he loses five times what he made. Still, on three hundred and sixty-four out of three hundred and sixty-five days he was very happily making money. It's much harder to be the other guy, the guy losing money three hundred and sixty-four days out of three hundred and sixty-five, because you start questioning yourself. Am I ever going to make it back? Am I really right? What if it takes ten years? Will I even be sane ten years from now?" What the normal trader gets from his daily winnings is feedback, the pleasing illusion of progress.
[Hmmm. Can't get italics to work on the quotation.]
http://www.gladwell.com/2002/2002_04_29_a_blowingup.htm
Here's the part that resonated with me:
"We cannot blow up, we can only bleed to death," Taleb says, and bleeding to death, absorbing the pain of steady losses, is precisely what human beings are hardwired to avoid. "Say you've got a guy who is long on Russian bonds," Savery says. "He's making money every day. One day, lightning strikes and he loses five times what he made. Still, on three hundred and sixty-four out of three hundred and sixty-five days he was very happily making money. It's much harder to be the other guy, the guy losing money three hundred and sixty-four days out of three hundred and sixty-five, because you start questioning yourself. Am I ever going to make it back? Am I really right? What if it takes ten years? Will I even be sane ten years from now?" What the normal trader gets from his daily winnings is feedback, the pleasing illusion of progress.
[Hmmm. Can't get italics to work on the quotation.]
Ten years on Empirica is closed, and its record of returns is interestingly vague, given all the media attention at the onset. Universa Investments L.P. sprung up from the ashes in 2007. Again, the record of return is "secret", except for careful leaks in the press when things happen to pay off. But wait, there's more! They have a brand-new "Black Swan ETF" that hedges the S&P 500 and charges you 0.95% for the privilege.
It looks a whole lot like the usual hedge fund bullshit play to me. Make lots of noise for a few years, get the punters to play, close down, then open up under a new name. Lather, rinse, repeat.
It looks a whole lot like the usual hedge fund bullshit play to me. Make lots of noise for a few years, get the punters to play, close down, then open up under a new name. Lather, rinse, repeat.
I don't agree with this completely. The idea that Nassim Taleb talks about in Antifragility relates to the allocation across assets of 1/n, i.e. all have a uniform probability of 'being huge', since it's impossible to know at the time of investment. You don't know which ones are going to take off, but you put yourself in a position to capture as much positive upside as possible.
there is a chance he read Nassim Talebs book and tried to apply a technique of betting on very rare and unpredictable events as a risk mitigation measure or profiting when everyone else loses. but i'm not sure how that applies to his examples.. It's seem like more of a "numbers game" with some general advice on how to be successful. Kudos on making a catchy title.
I don't think he understands what a black swan means. You can't 'shoot for' a black swan because it's by definition unknowable. A specific 'sale or product idea' isn't a black swan. Meeting the perfect cofounder is arguably one; it's as likely to happen at a rave as at a tech meetup. An even better example is the side project that you never expected to turn into a startup.
"Although Black Swans are risky by their nature you can put yourself in a position where you maximize the chance of success. For example; You want to have the best team to maximize the chance of success."
There's something wrong with this paragraph; I don't see an example.
"Although Black Swans are risky by their nature you can put yourself in a position where you maximize the chance of success. For example; You want to have the best team to maximize the chance of success."
There's something wrong with this paragraph; I don't see an example.
The basic idea behind "the black swan" is that given a limited sample it's foolish to try and estimate the probability of tail events, and thus in particular the expected value of convex functions.
Say for instance you have a random variable which follows the following mixture: draw from a standard normal 99.99% of the time and from a standard normal 10 standards deviations away 0.01% of the time.
If you collect 1,000 sample, 90% of the time you won't see a single instance where the value was drawn from the second mode. You plot your data and find that it perfectly fits a normal distribution. Neat!
I then ask you, what the probability to get an event >10 standard deviation away. You punch in the numbers and come up with 1.31e-23. The real answer was about 1/20000. You're wrong by about 19 orders of magnitude.
Moral of the story: if your model spits a probability much lower than the inverse of the sample size, don't trust it.
Many financial instrument, like options, display convexity. This places greater emphasis on the tails of the distribution than on the center. If the option seller underestimates tail risk, it would be profitable on average to buy such options.
There are many reasons why they might underestimate tail risk. They could ignore the theoretical argument I've exposed above. It could be that in a marketplace, sellers who ignore tail risk can stay in business long enough to put those who don't out of business.
Say for instance you have a random variable which follows the following mixture: draw from a standard normal 99.99% of the time and from a standard normal 10 standards deviations away 0.01% of the time.
If you collect 1,000 sample, 90% of the time you won't see a single instance where the value was drawn from the second mode. You plot your data and find that it perfectly fits a normal distribution. Neat!
I then ask you, what the probability to get an event >10 standard deviation away. You punch in the numbers and come up with 1.31e-23. The real answer was about 1/20000. You're wrong by about 19 orders of magnitude.
Moral of the story: if your model spits a probability much lower than the inverse of the sample size, don't trust it.
Many financial instrument, like options, display convexity. This places greater emphasis on the tails of the distribution than on the center. If the option seller underestimates tail risk, it would be profitable on average to buy such options.
There are many reasons why they might underestimate tail risk. They could ignore the theoretical argument I've exposed above. It could be that in a marketplace, sellers who ignore tail risk can stay in business long enough to put those who don't out of business.
VC is by definition not investing in black swans, since there is a long history of exits and failures to the point that PG & co can make claims like '1 company invested in per YC will pay for all the rest'.
VC is merely extremely risky but so far positive expected value.
VC is merely extremely risky but so far positive expected value.
I'm not sure if you're entirely correct here. True, VC has had a long history of exits and failures. However, PG & co have generally had a better history of Black Swan-type events in their portfolio (e.g. Dropbox).
Taleb's Black Swan theory and approach to hedge fund management is similar to VC.
From his Wikipedia entry: "As a trader, his strategy has been to safeguard investors against crises while reaping rewards from rare events, and thus his trading career has included several jackpots followed by lengthy dry spells."
His fund steadily and intentionally loses money or breaks even on a normal day, but vastly outperforms other funds on days when the market swings wildly in either direction. The strategy is not for the faint of heart.
Taleb's Black Swan theory and approach to hedge fund management is similar to VC.
From his Wikipedia entry: "As a trader, his strategy has been to safeguard investors against crises while reaping rewards from rare events, and thus his trading career has included several jackpots followed by lengthy dry spells."
His fund steadily and intentionally loses money or breaks even on a normal day, but vastly outperforms other funds on days when the market swings wildly in either direction. The strategy is not for the faint of heart.
> I'm not sure if you're entirely correct here. True, VC has had a long history of exits and failures. However, PG & co have generally had a better history of Black Swan-type events in their portfolio (e.g. Dropbox).
I don't see how Dropbox is a black swan. 'Wow, another Internet company went to billion-dollar valuation? Who could possibly have seen that coming?' Pretty much anyone who has been breathing since the Netscape IPO...
> From his Wikipedia entry: "As a trader, his strategy has been to safeguard investors against crises while reaping rewards from rare events, and thus his trading career has included several jackpots followed by lengthy dry spells."
Yes, that's his strategy for exploiting convexity, but Taleb is the one insisting on 'Black Swan' as a term for the unpredictable and unpredicted, not me. I am merely pointing out that by Taleb's lights - the inventor of the term - VC is not a black swan.
> His fund steadily and intentionally loses money or breaks even on a normal day, but vastly outperforms other funds on days when the market swings wildly in either direction. The strategy is not for the faint of heart.
They also failed, BTW. Even in 2001 his Empirica fund turned in a crappy performance; and personally, if you set up a fund to exploit sudden market shocks and you can't profit handsomely off 9/11, you have failed and partially discredited your overall thesis.
I don't see how Dropbox is a black swan. 'Wow, another Internet company went to billion-dollar valuation? Who could possibly have seen that coming?' Pretty much anyone who has been breathing since the Netscape IPO...
> From his Wikipedia entry: "As a trader, his strategy has been to safeguard investors against crises while reaping rewards from rare events, and thus his trading career has included several jackpots followed by lengthy dry spells."
Yes, that's his strategy for exploiting convexity, but Taleb is the one insisting on 'Black Swan' as a term for the unpredictable and unpredicted, not me. I am merely pointing out that by Taleb's lights - the inventor of the term - VC is not a black swan.
> His fund steadily and intentionally loses money or breaks even on a normal day, but vastly outperforms other funds on days when the market swings wildly in either direction. The strategy is not for the faint of heart.
They also failed, BTW. Even in 2001 his Empirica fund turned in a crappy performance; and personally, if you set up a fund to exploit sudden market shocks and you can't profit handsomely off 9/11, you have failed and partially discredited your overall thesis.
This reads like a badly-explained piece of Taleb's most recent book "Antifragility". Taleb specifically mentions most of the points mentioned, "the focus on teams" and "learning from past mistakes". Granted, these are obviously a huge part of startups in general, but if you pretend that you knew what a Black Swan was before Taleb popularized the term in 2001, you're a philosophy of science PhD or a liar. The fact that our young blogger gives absolutely NO credit to Taleb's book is disingenuous at best.
Most people who write about Black Swans nowadays would acknowledge the phrase comes from the book "The Black Swan". It's not disingenuous to not mention Taleb in every piece of writing which uses the phrase "Black Swan" precisely because Taleb is famous for creating/popularizing the phrase.
Attempt to achieve many Black Swans in parallel so that the overall probability of hitting one is higher
This can be misunderstood. One of the most valuable lessons I learned in life so far, is keeping focus. At least I am unable to do more than a few things with full force. And a half baked attitude will be counter productive towards reaching black swan moments.
This can be misunderstood. One of the most valuable lessons I learned in life so far, is keeping focus. At least I am unable to do more than a few things with full force. And a half baked attitude will be counter productive towards reaching black swan moments.
"The disproportionate role of high-profile, hard-to-predict, and rare events that are beyond the realm of normal expectations in history, science, finance, and technology"[1]
[1] http://en.wikipedia.org/wiki/Black_swan_theory