Assess my strategy, please

D

doublewobble

Dear All,

I have formulated my first strategy that looks anywhere near promising in backtests. Could you please give your opinion?

Software: Amibroker.

Frequency: hourly.

Data: SP500 stocks 1.04.2004-1.04.2021 (obtained from the FirstRateData, KMI omitted). 761 symbols traded (probably more because some are recycled).

Method: technical, no fundamentals included.

General rules:
* no leverage;
* no fixed annual interest for the money out of the market;
* long only;
* commission 0.005 per share;
* trade during market hours;
* trade at Open prices;
* 2% max loss stops executed immediately with "Exit intraday at stop" option (at the low price of the bar);
* position size not exceeding 1% volume of the -7th bar (hourly volume is heavily dependent on time), set as the following: SetPositionSize( round(Ref(Volume, -7)*0.01), spsShares );;
* position size not exceeding 20% of the Portfolio Equity;
* do not enter if not listed in SP500;
* exit when delisted from the SP500.

Profitability:
Initial capital: 100 000;
Ending capital: 3 469 965.70;
Exposure %: 30.71%;
Annual Return %: 22.84%;
Risk Adjusted Return %: 74.38%;
Transaction costs: 141 479.47.

Trades:
All trades: 5744;
Average bars held (Winners): 11.98;
Average bars held (Losers): 11.66;
Max. Consecutive (Winners): 23;
Max Consecutive (Losers): 14.

In that thread , somebody listed his set of rules (long & short):
  1. Take the Absolute Money into consideration, don't use the Leverage Money while you are calculating the Backtest Report. You should test atleast 5-6 years data, so that the system is going through every phase of the market.
  2. Commission (Slippage +Brokerage + Govt Taxes) = 10% on the turnover you have to count it. The more slippage while increase the potential of your strategy. If you kept 20% commission, after that if your system is making money then it is fantastic, still you need to consider the below fields in the backtest.
  3. CAR / MAX DD should be greater than 2.5 or more.
  4. CAR > 20-24%
  5. Pay off Ratio >2
  6. Max DD: 12-20% of the Absolute Capital.
  7. Winnings & Losers Ratio: 40-60% would be idle.
  8. Sharp Ratio > 1.5 is good.
  9. Go to Charts and now look at the Profit Table, QoQ is it Positive or not. If above all satisfies then we can shortlist the system and go to the Paper trading for 3months and if not found any errors then you can go LIVE.
  10. Screenshots attached. my 5th Rule has not satisfied, but we made it Live, as rest are satisfying my conditions.
  11. Forgot to mention, also please CODE CHECK & PROFILE, if it says GOOD, then fine, if it is referencing to the Futures Quotes, then the results which are showing in the back test can't be taken into the real time trading. Such systems are not worthy. Analysis > Formula Editor > Tools > Code Check and Profile.
  12. Avg. Profit/Loss% - should be greater than 0.25.
More...

My results (long only):

1. Yes.
2. I do not know how to include slippage into the calculations. However, the 1% vol rule applied to the position size should avoid most of the trouble, yes?
3. CAR / MAX DD 3.50. Better than his.
4. CAR 22.01%.
5. Pay Off Ratio 1.38. Not filled.
6. Max DD 6.53%. Better than his.
7. Winners 59.05%, Losers 40.95%. Better than his.
8. Sharp ratio 2.20. Better than his. Calculated with 5% Risk Free rate that has become nonsense in the recent years anyway.
9. I am not sure what he means by the QoQ (not included to the Amibroker standard report) but the strategy does not lose money any year and the average results are positive for all months. The profit table:
Year Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Yr%
2004 0.0% 0.0% 0.0% 1.4% 2.1% 5.6% 2.1% -2.9% 0.0% 1.7% 4.1% 0.5% 15.4%
2005 0.1% 0.0% 0.3% 0.0% 0.0% 0.2% 2.9% -0.1% 0.0% 0.0% -0.6% 1.6% 4.4%
2006 -0.2% 6.0% 2.7% 0.7% -1.4% 3.0% -0.1% 0.0% 3.8% -0.4% 2.1% 1.5% 18.8%
2007 0.0% -0.4% 0.0% 4.5% 3.6% -0.6% 2.4% 2.7% 0.3% 5.9% -0.4% 0.0% 19.2%
2008 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0%
2009 0.0% 0.5% 0.0% -1.0% 4.7% -2.9% 7.9% 2.0% 4.2% 3.0% 0.0% -0.6% 18.6%
2010 -0.7% 1.2% 8.3% 7.3% 0.0% 0.0% 0.0% -1.8% 1.6% 3.3% 1.1% 0.0% 21.8%
2011 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.0% 0.1% 2.3% 2.3%
2012 3.1% 5.5% 2.7% 0.5% -1.6% -0.8% 0.5% 0.6% -0.5% 0.2% -0.6% 0.0% 9.9%
2013 5.4% 3.5% 2.6% 12.7% 7.0% 0.0% 4.5% 2.6% 0.0% -0.9% 1.7% 2.9% 50.2%
2014 0.3% 2.7% 1.4% 2.8% -1.0% 3.6% 1.8% 0.4% -0.5% 4.6% 8.4% 0.4% 27.5%
2015 0.0% 3.3% 3.1% -2.3% -0.2% 1.2% -0.8% 0.0% -1.1% 3.2% 3.3% 2.7% 13.1%
2016 -0.6% 0.0% 3.5% -0.4% 1.4% 3.2% 1.8% 0.0% -0.1% 1.4% 7.6% 3.6% 23.2%
2017 1.9% 9.1% 1.6% -0.6% 1.6% 0.4% 6.1% 1.5% 6.2% 3.3% 3.4% 0.9% 41.3%
2018 6.0% -3.6% 1.5% 0.5% 1.5% 4.1% 6.1% 2.5% 0.1% -0.9% 1.1% 2.2% 22.9%
2019 4.7% 7.7% 3.8% 1.3% -0.6% 1.3% 6.9% 0.0% 5.9% 2.4% 4.1% 1.1% 45.7%
2020 0.6% -0.8% 0.0% 5.4% 14.2% 2.2% 4.0% 5.1% -2.7% 7.0% 17.4% 4.2% 70.9%
2021 5.8% -0.8% 3.4% N/A N/A N/A N/A N/A N/A N/A N/A N/A 8.5%
Avg 1.5% 1.9% 1.9% 1.9% 1.8% 1.2% 2.7% 0.7% 1.0% 2.0% 3.1% 1.4%
10. The 5th rule not filled as well.
11. Good.
12. 0.46%. Better than his.

The Monte Carlo table:
Percentile Final Equity Annual Return Max. Drawdown $ Max. Drawdown % Lowest Eq.
1% 2798247 21.32% -103593 -52.36% 49745
5% 2988408 21.78% -89118 -37.42% 71963
10% 3082482 22.00% -78860 -31.14% 78631
25% 3264393 22.41% -67363 -23.28% 88798
50% 3453929 22.81% -57843 -16.66% 96531
75% 3650861 23.20% -50478 -11.84% 99838
90% 3856601 23.60% -44613 -9.39% 100000
95% 3987112 23.83% -42360 -8.16% 100000
99% 4199208 24.21% -38126 -6.75% 100000

Thanks in advance.
 
That Quora thread you are linking to, contains much valuable information for any systematic trader, especially for new ones to be business.
 
From that thread

Joseph Wang, ex VP-Quant {sha-wing - had to include that}

..........

"6) Understand why your strategy really works. If you just look at the stock market over the last few years, it would be possible to get 10%/month return by just leveraging."

Can you answer that? And BTW did I miss the enter/exit rules somewhere?
 
I will not disclose the precise entry & exit signals.

No, all SP500 constituents are watched. Positions are sized not to exceed 20% of the Portfolio Equity and 1% of the volume of the same hour of the previous day. If multiple signals are received at the same time, the available resources are divided equally between them, taking into account the rules mentioned above.
 
doublewobble said:
I will not disclose the precise entry & exit signals.

No, all SP500 constituents are watched. Positions are sized not to exceed 20% of the Portfolio Equity and 1% of the volume of the same hour of the previous day. If multiple signals are received at the same time, the available resources are divided equally between them, taking into account the rules mentioned above.
More...
I can understand that you won´t disclose your strategy. But thank you for your reply.
 
I would not be to worried about not going further back than 2004 (i assume your intraday data has limitations, but i would try to get some more history to get you some hints about other market types) i would also look at average trd%.. i cant see that here, but your AB should output that in the metrics.. i would increase commission to further see how it affects. You say entry at next open... is it a intraday system using 60 min bars to crate daily bars and entry next open is next day open ? if not i would believe you would have slippage on entry and exit. Anyway, perhaps i missed something, but my thoughts at least.
 
You must log in or register to reply here.
Back
Top