I am wondering how it work in big firms like Citadel Securities in market-making and HFT teams.
You have a lot of different pods, and the number of markets where you can operate is small. Hence multiple pods are going to operate on the same exchange.
Isn't that a problem?
Let's say team A starts running there new algo on NYSE and takes the order quicker than the team B that has been trading on NYSE for a long time. So team A takes all the alpha of team B.
Hence my question is: how do they prevent this from happening?
My guess would be that there is a central execution team that takes all the alpha, so the prediction of each team and combine them to at the end get only one order. But then the problem is how the PnL attribution is done for each team? If the alpha are combined then how to know which alpha contributed the most to the PnL?
