Selling systems I developed to hedge fund
- Thread starter mchugh
- Start date
This is just simplified description how institutions do it . Boss comes to a geek and tells him to make a system which will create these trades ( supplied by system seller for evaluation ) Geek goes to their system library and runs it on the market of a choice .Trades generated by systems library are matched against system seller trades and the identical ones are picked out .
At this point, all the library systems that produce at least 20% of the same trades are backtested on the same market as sellers system.
If any of the library systems has better backtested performance, sellers system is rejected.
If the sellers system has better performance then they take rules from library systems for all matching trades and make combined system and adjusting rules till they beat sellers performance. If they cannot beat it they will buy it .
So unless one has a personal ties to upper decision makers at that institution and they want you to work with them, seller's chances are not that great .
I do it little bit differently and not as precise as they do it but results are always the same.
As far as Gann goes , it does not matter at all how the signals are generated they will just generate identical signals using something else or they write their own formula .
Gann mechanical systems are not profitable though.
I agree. If you design a system that optimises itself based on some kind of volatility level or changing condition or cycle, trades signals will change with market dynamics. You can't predict that from a limited number of signals on a chart.
Anyone who thinks you could has no idea what they are talking about. Unless of course their idea of a sophisticated trading system is an MA crossover
Why would institutions wait until someone comes forth with a system before they attempt to combine their existing systems & adjust rules to produce the best possible system? It sounds like the person who brings forth the system to beat doesn't add any value to the process other than motivating the geeks to try harder (e.g. why weren't they doing this all along?). Are you saying that they could not have come up with this combined system from their existing libaries without the target signals to reverse engineer?
