What it does
Our project follows a simple approach for making relatively low risk trades. It generally follows the principle given in the challenge description. The algorithm identifies opportunities to sell ETFs or the corresponding stocks and use the money to buy the other. We try to keep the hedgded position to a minimum.
How we built it
We take the best bids and asks of each triplet and calculate whether buying or selling in a 2:1 ratio allows us to generate profit. If that is the case we calculate the maximum volume for that the transaction can be carried out and they are sent to the exchange.
Challenges we ran into
One of the biggest challanges was keeping hedging at bay. For this we used linear algebra, the squared error and minimization to calculate the most efficient way to add or subtract the volume bought or sold of individual stocks or ETFs so that the ratio decreases.
At the same time we had a hard time generating consistent profit. This was solved by first executing the selling action and then buying the corresponding amount. It especially helped to set the price one is willing to pay for a stock to the value one used for selling. This allows the exchange of the most amount between stocks and ETF. It minimizes hedging and with that protects the investment from being affected by fluctuations of the market.
What we learned
Risk Management and that one can work through a night with enough energy drinks and pressure.
What's next for AT-AT: Algorithmic Trading at TUM
Selling iPhones?
