Quant/Quant-Discretionary Trading Dilemma

Discussion in 'Risk Management' started by CPTrader, Dec 12, 2008.

  1. Quant/Quant-Discretionary Trading Dilemma

    I have a fully developed and back-tested multi-market multi-strategy quantitative trading program that I have near 100% confidence in.

    Additionally, I have a quantitative strategy with a substantial discretionary overlay, let say 70% quantitative, 30% discretionary. This “overlay” strategy is more dynamic than the 100% quantitative strategy, has a smaller minimum capital requirement, and is far more nimble in containing losses due to the discretionary element. However, due to the discretionary component of this strategy, I don't have that "near 100% confidence" in it as I do in the 100% quant strategy.

    If capital wasn’t an issue, I would trade the 100% quant program, but with such a high minimum capital requirement (high seven figures) I cannot trade it at this time unless I increase the leverage by 2-4 times and thus reduce its capital requirement. The risk of course in doing so is severe drawdowns.

    So my dilemma is:

    Do I trade a strategy (the 100% quant strategy) that I am virtually 100% sure of its potential at 2-4x higher leverage and risk severe drawdowns or do I trade the 70% quant-30% discretionary strategy; a more nimble, dynamic strategy with lower capital requirements but with lower confidence.

    Your thoughts and ideas, please. I think there can be many lessons here about risk management, portfolio design, quant vs. discretionary trading.

    Thank you.
     
  2. Define "severe drwadowns" for your strategy...or in other words - not your Q/Q dilemma but your R/R ( risk / reward ) dilemma.
     
  3. Pls explain why the 100% Q requires more capital. Don''t both startegies trade the same markets/instruments?
     
  4. Sluggo's post is insightful in some ways but doesn't address the dilemma.

    My philosophical belief is that truly robust systems work on multiple markets, multiple sectors, multiple regions and are truly a "suite of systems". It is very difficult if not impossible to achieve robustness and good risk/reward metrics with a small account size without accepting some risk-reward limitations. So perhaps some more extensive research could address this problem, perhaps, but I doubt it can be fully addressed...can you drive a car with three good tires and one flat tire....YES..but not for long. Research has led to "run-flat' tires...but even the great research that led to run-flat tires can't solve the three good tires/one flat tire dilemma

    I could easily have a one-market/one strategy system....but how robust is that; how "safe" is such an approach??