Foreign Exchange Arbitrage with Haskell

>> Assuming that I've understood the format properly (anyone know any good books to get this kind of basic information?), then running this code through the 21 million (!) ticker updates in January 2010 yields absolutely ZERO arbitrage opportunities. With my limited knowledge, I think this is because of the efficient-market hypothesis.

This doesn't even support an extremely weak version of the efficient-market hypothesis. It's very easy to believe that simple, first-order arbitrage opportunities like this effectively don't exist; the weak version of efficient-market hypothesis implies that the price of significantly more complex assets reflect all publicly available information. Possible currency arbitrage opportunities like he tried to find don't even require any information beyond the prices.

That said, it's very cool that he did this, even if the results are unsurprising.

>> Back to the drawing board with my money making schemes!

Is anyone surprised with that result? Nice reading though, i appreciate his time and willing to share the result. I would be to lazy to prove that myself. ))

You can also do this using standard Bellman-Ford or Floyd-Warshall if you set edge weights to be -log(exchange rate).