DominiConnor
Quant Headhunter
- Joined
- 9/6/06
- Messages
- 1,051
- Points
- 93
I see on another thread a Quant Programming competition, and although we could debate how "quant" a matching engine problem might be, it's a good idea, and since I've personally written that sort of code, see it as a perfectly valid test.
I have something of a following these days, so for the right competition I believe I could attract a good number of entries, and if could act as one of the judges.
A good competition problem has the following characteristics:
1: You can't do well merely by googling a standard algorithm and turning pseudo code into C
Conversely, it should not require high levels of knowledge in a highly specialist domain.
2: It must be something like the work quant devs do in real life.
This partly conflicts is (1), I didn't say it would be easy...
3: It must have some objective measure of success, like speed or accuracy
I like the idea of X% for objective quality and Y% for being clever and innovative
4: Absolutely all criteria for success must be publicly available, including the formula for (3)
5: The problem should be small enough in volume that one smart person can do it, but we allow teams. I allow teams because there will be teams anyway, this just keeps it honest.
6: The problem should be sufficiently impressive that when you put on your resume that you won, a future employer should be impressed.
I have something of a following these days, so for the right competition I believe I could attract a good number of entries, and if could act as one of the judges.
A good competition problem has the following characteristics:
1: You can't do well merely by googling a standard algorithm and turning pseudo code into C
Conversely, it should not require high levels of knowledge in a highly specialist domain.
2: It must be something like the work quant devs do in real life.
This partly conflicts is (1), I didn't say it would be easy...
3: It must have some objective measure of success, like speed or accuracy
I like the idea of X% for objective quality and Y% for being clever and innovative
4: Absolutely all criteria for success must be publicly available, including the formula for (3)
5: The problem should be small enough in volume that one smart person can do it, but we allow teams. I allow teams because there will be teams anyway, this just keeps it honest.
6: The problem should be sufficiently impressive that when you put on your resume that you won, a future employer should be impressed.
