Hacker News new | comments | ask | show | jobs | submit login

I think you might be selling Carreyrou short. The "corporate scandals" you're saying he wrote about are, in many cases, scandals because he broke the story about them. There might still be a Theranos as a going concern if it weren't for him. He's not like Andrew Ross Sorkin writing about the collapse of investment banks in 2008 --- his reporting had a causal impact on the collapse he wrote about.

Read the book! It is a excellent read. There are technical details in it a-plenty.

_When Genius Failed_ isn't simply a description of a scandal, but also a pop-grade exploration of the technical factors that led to the scandal. The book is impossible to write without conversance with its technical subject matter. Which is my point: Lowenstein wasn't a quant, but he was able to accurately and effectively write on them.






I cannot speak on Carreyrou, but parts of _When Genius Failed_ make me think that Lowenstein did not have a good grasp of the subject he was writing about.

For example, when discussing LTCM volatility strategies: >The stock market, for instance, typically varies by about 15 percent to 20 percent a year. Now and then, the market might be more volatile, but it will always revert to form—or so the mathematicians in Greenwich believed. It was guided by the unseen law of large numbers, which assured the world of a normal distribution of brown cows and spotted cows and quiet trading days and market crashes. For Long-Term’s professors, with their supreme faith in markets, this was written in stone. It flowed from their Mertonian view of markets as efficient machines that spit out new prices with all the random logic of heat molecules dispersing through a cloud. From this quote it looks like Lowenstein believed that 1. Volatility is not mean reverting 2. The reason people at LTCM believed volatility was mean reverting was due to complex mathematical models rooted in market efficiency. Now, volatility mean reversion is something that can be easily seen by looking at a graph of volatility over time [1].

Also, Lowenstein does not do a good job at describing the way LTCM modeled risk, and the way it failed. In fact, he barely even tries. He just hints here and there about "correlations going to one", but nothing more. For example, which assumptions failed? Did they assume that different kind of bets (relative value between bonds, merger arbitrage, arbitrage between double-listed equities, etc) were uncorrelated? Did they get hurt more by different kind of bets being correlated, or by some bets going particularly wrong? Did they stress test their risk measures in any way?

[1] E.g. https://a.c-dn.net/b/4tKv1V/Forex-Trading-Video-SPX-and-VIX-...


Ask quants if they feel that it’s a balanced view of quants though. It might seem accurate to you, but you’re not a quant now, nor were you one during the era of LTCM.

You are in no position to judge how accurately or effectively it was written (unless you judge effectiveness as entertaining you).




Applications are open for YC Summer 2019

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: