One of my strategies has been treading water since the middle of 2009. I find the following scenarios plausible and list them in decresing order of my estimate of their plausibility: 1. a regime shift in the market makes this strategy unprofitable for now, it may recover at some undetermined point in the future. 2. a structural change in the market makes this strategy obsolete, it will never make any more money. 3. my broker (who happens to run an internal quant HF fund) reverse engineered this strategy and found a way to effectively front run my entries and take all the juice out. 4. other random people stumbled upon the inefficency this strategy was exploiting and found a faster or more efficient way to capture it, leaving no crumbs on the table for me. From a practical standpoint, what I did was to reduce the size of this strategy and keep monitoring it and deployed the resources towards other strategies that keep on working. I am particularly worried about scenario 3 above, and was contemplating obfuscating my remaining strategies. Do you guys think this paranoia is justified?
More info needed (1) market traded (2) timeframe of strategy (3) for how long it worked (4) how many parameters it has (5) who is your broker (6) was that automated?
ive been questioning whether it is illegal to front run a technique/method which would have a good chance of pushign the market for a bit, for example if HF all used a secret technique to enter on stocks, and you found it out would that be illegal?
(1) market traded very liquid NYSE and NASDAQ listed equities (think SP500 constituents) (2) timeframe of strategy day trading (no overnight) (3) for how long it worked the spreadsheet is a backtest permited by my database I went live in the middle of 2008 and the actual equity curve is better than the spreadsheet (I have very conservative transaction costs in the backtest). So it worked for about a year before stopping. No way for me to go earlier with the backtest as I do not have the data. (4) how many parameters it has I use estimates for volatilities and beta for each stock just to manage risk (to determine position size and to hedge out the market exposure). Other than that there is one "threshold" parameter which I set to the most "common sense" value of 0. (5) who is your broker I would rather keep this info confidential. (6) was that automated? yes
I have some strategies (also daytrading liquid stocks, mostly reversion-to-mean type strategies) that were fine in 2006-2008 and died when the volatility died around April, May, June '09. Lack of volatility? Lack of liquidity? Lack of retail participation? Effect from QE/stimulus? I don't know.
Most optimized (many ways to optimize - not just parameter optimization) strategies work for about a year and then turn catastrophic Have you calculated your profit factor? win rate? avg. win to avg. loss? Sharpe ratio? All from real trades only?
From the actual trades for 2009 (made money first 6 months, with the last 6 months flat) I get a daily PF of 1.4 and an annualized Sharpe of 2.0. Do not have the other stats. There is always the chance I was just lucky, and the strategy is simply a datamining artifact but the sheer number of trades, over 2000 would argue against it.
The reason the realized PF and Sharpe are lower than their counterparts in the backtest is due to some unfortunate leverage decisions I made along the way, as well as various production screwups.
Most people conclude that all strategies have lives. A closer look reveals just a distribution of the traders of these strategies. Good strateies are oriented to market principles. You are not in that boat nor are hedge funds and their strategies. Look at your question set to be able to conclude where you are operating relative to market principles.