System Details: Pairs Trading QID QLD Scalper Hypothetical Monthly Results (limitations of these statistics) Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 2007 +7.6% (5.1%) +9.5% +3.6% (8.9%) +11.8% +11.3% +17.4% +2.1% +4.9% 2008 (12.9%) +19.2% +12.7% (2.4%) +5.5% (8.9%) +2.3% +3.0% (17.5%) - (6.1%) - 2009 (4.9%) +12.4% (14.8%) +39.4% (6.8%) (6%) (18.5%) +9.5% +11.9% (11.3%) -11% for October 2009? how did that happen? you seem to have entered long QLD yesterday. QLD is down ~6%, are you using 2x leverage?
Yes, I got smoked today. The exact sizing is 192% of equity. My opinion was that today was pure panic selling and shorts looking for an opportunity to pile on piled on today. It really isn't any surprise that the data is bad. It wouldn't be irrational to sell off after poor economic news, but it would be to think that we'll never see DOW 14000 or S&P 1500 again.
I doubt it would, and, mechanically, most days we are oversold, then revert back to fair value the next day and I sell the day after the entry, (vice versa for overbought), so I'm pretty sure it would not as some entries are good with little MAE and some have a lot more. This is the eighth worst trade in my backtest.
Well, it's disappointing for me to say that compliance finally came through with an answer for me, and they denied my request to trade this system for any client, unfortunately. Well, I guess my edge will just have to last a bit longer. I've gone back to see if there were some risk adjusted improvements I could make to my model PTQQS and I believe I was succesful in that regard: Long + Short Starting Capital $100,000.00 Ending Capital $637,853.61 Net Profit $537,853.61 Net Profit % 537.85% Annualized Gain % 77.71% Much higher annualized gain at 100% of equity, it was at its previous highest 57% before these model fitness tests for fair value calcuations were added. Exposure 24.06% Cash Interest $0.00 Margin Loan Interest $0.00 Total Commission ($1,240.00) DividendsPaid $0.00 Number of Trades 78 Avg Profit/Loss $6,895.56 Avg Profit/Loss % 2.56% Avg Bars Held 2.62 Winning Trades 58 Winning % 74.36% Higher win percentage contributed to a profit factor from 1.6 as the old to the current 2.47 Gross Profit $902,221.06 Avg Profit $15,555.54 Avg Profit % 4.74% Avg Bars Held 2.79 Max Consecutive 9 Losing Trades 20 Losing % 25.64% Gross Loss ($364,367.44) Avg Loss ($18,218.37) Avg Loss % -3.76% Avg Bars Held 2.1 Max Consecutive 4 Max Drawdown ($125,159.69) Max Drawdown Date 8/17/2009 Max Drawdown % -17.45% Drawdown percent at 100% of equity is consistent with the previous max dd date. Max Drawdown % Date 8/17/2009 APD .7519 Wealth-Lab Score 266.6894 RAR 323.063 Profit Factor 2.4761 Recovery Factor 4.2973 Payoff Ratio 1.2596 Sharpe Ratio 2.0187 And this ratio had been as low as 1.6 in my backtest previously, now to over 2, and that's really a holy grail in the sense that over long periods of time this will trend lower. Ulcer Index 5.0002 WL Error Term 6.213 WL Reward Ratio 12.5084 Luck Coefficient 3.4036 Pessimistic Rate of Return 2.5933 Equity Drop Ratio 0.0438 Mostly all I added was a closing price based stop, and an improved calculation of fair values, including an rsquared test for model fitness that drastically changed certain trades. I see from my APD I was letting losses run a bit too much, but in my mind when drawdown from dividends is counted, it affects average profit and loss characteristics, as well as total drawdown, inappropriately. Can't get too much better with the leveraged results at 189% of equity, I found the 192% of equity had daily exposure levels exceeding 200%, meaning these are periods where there is less that 50% net equity in the account, which leads to a margin call ultimately. Long + Short Starting Capital $100,000.00 Ending Capital $2,697,715.44 Net Profit $2,597,715.44 Net Profit % 2597.72% Annualized Gain % 178.02% Exposure 43.22% Number of Trades 78 Avg Profit/Loss $33,304.04 Avg Bars Held 2.62 Winning Trades 58 Winning % 74.36% Gross Profit $5,216,469.59 Largest Winning Trades $544,077.44 Avg Profit $89,939.13 Avg Bars Held 2.79 Max Consecutive 9 Losing Trades 20 Losing % 25.64% Gross Loss ($2,618,754.16) Largest Losing Trade ($344,254.19) Avg Loss ($130,937.71) Avg Bars Held 2.1 Max Consecutive 4 Max Drawdown ($1,051,429.00) Max Drawdown Date 8/17/2009 Max Drawdown % -30.85% Max Drawdown % Date 8/17/2009 APD 0.5428 APAD 1.4658 Wealth-Lab Score 284.7828 RAR 411.8399 MAR 5.7702 Profit Factor 1.992 Recovery Factor 2.4707 Sharpe Ratio 1.9469 Sortino Ratio 7.7662 Ulcer Index 9.1261 WL Error Term 10.573 WL Reward Ratio 16.8369 Luck Coefficient 6.0494 Pessimistic Rate of Return 1.4142 Equity Drop Ratio 0.0256 K-Ratio 0.4997 Seykota Lake Ratio 0.1128 Expectancy 0.6418 Expectancy Score 14.9995 Max Losers Held 1 Max Winners Held 1 Some of the changes to the most recent trades were the recent 8% loss would have stopped out at 48.7, a 4.5% down move from the previous close would have triggered this exit, and the point at which I was selling around 47.34 would have been a buying point, so what I must do is "synch up" with the current version of the model by buying back into QLD.
Long + Short Starting Capital $100,000.00 Ending Capital $3,011,975.44 Net Profit $2,911,975.44 Net Profit % 2911.98% Annualized Gain % 186.92% Exposure 45.93% Cash Interest $0.00 Margin Loan Interest $0.00 Total Commission ($1,272.00) DividendsPaid $0.00 Number of Trades 80 Avg Profit/Loss $36,399.69 Avg Profit/Loss % 2.58% Avg Bars Held 2.61 Winning Trades 61 Winning % 76.25% Gross Profit $5,937,278.26 Avg Profit $97,332.43 Avg Profit % 4.63% Avg Bars Held 2.74 Max Consecutive 10 Losing Trades 19 Losing % 23.75% Gross Loss ($3,025,302.82) Avg Loss ($159,226.46) Avg Loss % -3.99% Avg Bars Held 2.21 Max Consecutive 3 Max Drawdown ($1,053,370.50) Max Drawdown Date 8/17/2009 Max Drawdown % -29.60% Max Drawdown % Date 8/17/2009 Wealth-Lab Score 286.472 RAR 406.9367 Profit Factor 1.9625 Recovery Factor 2.7644 Payoff Ratio 1.1605 Sharpe Ratio 1.9802 Ulcer Index 8.7781 WL Error Term 10.676 WL Reward Ratio 17.5081 Luck Coefficient 3.4868 Pessimistic Rate of Return 2.6426 Equity Drop Ratio 0.0208 So after a few more iterations and adding the fitness test, or rquared as I call it, I've managed to filter out the core script with a surprisingly robust indicator that is useless by itself, but wonderful with the sort of statistical arbitrage I use in Pairs Trading QID QLD Scalper. It gives me a level of confidence in certain conditions that act as a test for validity when I'm placing my trades. I was re-entering QLD today around 48.21, not through any slippage, but that I was unsure if I should be synching up at a higher price than we closed at friday. Ultimately, you could say greed got the better of me, but I'm up after hours, have some stops in place, and a target of 54.87 that could be hit because we are in fair value range but the fitness, r-squared indicator wasn't sufficiently large enough to exit the trade on the open tomorrow. For Cash Cow from BWolinsky we had a trade in SDS at 10:30 today EST and quickly exited at 10:45 for 0.47% loss. It seems there is a rounding issue with C2 in that they will round up your shares on decimalized buy orders and round down your shares on decimalized sell orders, leading a stray share to be closed once I had observed this problem. I corrected it with a "floor" call in WL5, and I don't think I'll have that problem again.
Pretty nice day today. Makes up for the bullshit panic selling that had me stopped out of my last trade. I had re-entered QLD after buying at 51.35 and selling at 47.32, to buy in at 48.21 and most likely to sell around 50.1 tomorrow. I'm pretty excited about the changes to the model. It does what the proverbial holy grail would have you do, which is to let winners run and cut losses quickly. I think the most percentage wise I've ever lost is 8% on QLD, or about 4% on the market, which for some is tough to stomach but it's all relative to the market. No trades to report for Cash Cow from BWolinsky. For QLD I have the following fair value predictions: QLD Projection: 52.7708331181453 QLD Close: 50.1399993896484 QLD Projection: 47.9643256447532 QLD Close: 50.1399993896484 QLD Projection: 48.8560368855794 QLD Close: 50.1399993896484 It would appear we are mostly overvalued at this point.
So its sounds like you made changes to your system hopefully for the better. Should you not just close your current system on C2, and start with the new system?
Has your model changed after your stop out? Buy at 51.35 sell at 47.32 Buy at 48.21 plan to sell at 50.10 This doesnât seem to be letting winners run or cutting losses quickly.
The stop out for the first trade was at 48.70, and the buy in would have been 47.32. The percentage of winning trades in excess of 10% was increased by 20%.