IT in banks - a theory

I'm coming to a tentative theory that in the ecosystem of bank-on-bank-on-hedgefund trading (consenting adults, very limited requirements on required disclosure) the majority of bank profits are not made by clever trading, hedging, sub-millisecond HFT trades and so on. Those are just the symptoms. Principally, profits are made over time by simply not screwing up as massively or as frequently as the competition; unplanned network or power outages, bad software pushes, fat fingers without a checking system, misplaced decimals, they all contribute to a steady drain on the bottom line.
If you can engineer a trading system which is designed to be robust, and as long as it's not managed by the all-too-prevalent idiots in banks' IT departments, you're at least half way towards being ahead of the competition. What you actually decide to trade isn't too germane as long as you trade a lot of it around the market price.