Feedback on Backtest Results
Hey Fellow Monkeys,
I developed a new strategy recently and hired a software engineer to put it into code and I just got some preliminary results I was hoping you all could help me interpret. Please bear in mind this strategy is designed to trade commodity futures as one of the biggest edges to this strategy is that futures trade 24 hours, but I only had a TD account at the time and they only allow one-month historical data on equities and nothing on futures. However, I am in the process of opening an Interactive Brokers account that’ll allow me to backtest futures for the entire year of 2020. Anyway, here are the preliminary results of this strategy on AAPL for August 9th-September 9th. Though it wasn’t designed to trade equities, I feel the results are encouraging. They came up as follows:
Chart Timeframe: 20 day 1 hour
Trades per day avg: 2.77 long, 2.88 short
I’m having an issue with entries and unnecessary reentries being triggered and I’m trying to work my way through that, but regardless the commission costs are currently not overwhelming at all, especially if I ever get to the point where I’m trading decent size and am in the position to negotiate rates.
Winning Trades: 58.33% long, 49.33% short
Once again this metric is a bit misleading because of the unnecessary reentry and immediate exit issue, but regardless seems sustainable.
Avg Profit Per Winning Trade: 0.547% long, 0.723% short
All of these metrics are conveyed as a proportion of the overall portfolio, so for a $100k account, an average long winner would be $547
Avg Loss Per Losing Trade: 0.207% long, 0.210% short
This seems that the profit factor would then be 2.64 for longs and 3.44 for shorts
Max Profit: 6.55% for longs, 1.67% for shorts
Clearly longs have performed better for this particular month, but with AAPL reaching all-time highs and practically going parabolic, this is not necessarily surprising
Max Loss: 0.832% for longs, 0.696% for shorts
This would make max profit factor 7.87 for longs and 2.40 for shorts
Average Candles Per Trade: 2 h:15m:49s for longs and 1h:08m:12 for shorts
This is not an HFT strategy, so with access to unfiltered CQG data and an excellent internet connection, which I do not currently have, the fills on these positions are extremely attainable
Avg Candles Per Winning Trade: 3h:16m:12s for longs and 1h:30m:17s for shorts
Avg Candles Per Losing Trade: 51m:16s for longs and 46m:41s for shorts
This follows one of the foundational adages of trading, “Let your winners run and cut your losers short”, and with my winners lasting 2-3 times as long as my losers, it seems this strategy adheres to this concept
And most importantly....
Overall P/L for Aug 9th-Sep 9th: +33.48%
P/L for longs: +27.91%
P/L for shorts: +5.58%
To a critic one might say the profitability on longs makes sense as AAPL is reaching all-time highs and seems to continuously rage on, but what I feel illustrates the robustness of this strategy is that even with a sub 50% win rate and AAPL going parabolic this past month, this strategy still returned over 5.5% going short. As mentioned, this isn’t even the asset class this strategy was designed for; about 60% of the trade opportunities for this strategy in the futures market trigger from 7pm-5am, so it’s a much better fit for that. This backtest does show implications of cross-asset functionality however. This strategy also functions quite well on the 1YD, 180d4h, and 3YW charts, but I don’t have the time or resources to run a formal backtest on those. I will post updated results once the IB account is open and a longer-term backtest is run on NG and ES. For now though, what do you all think? Thanks in advance.