Hello all,
I've been lurking at ET for several months but this is my first post here. By way of introduction, I'll disclose that I'm a newbie to futures trading - I have a little less than a year's experience. I have relatively little experience with any type of trading (about 5 years). Trading is not my profession. I have not been consistently profitable in my futures trading as of yet, but I have gotten to where I am treading water and not losing money.
Having said all that, I'd like to thank no_pm_please for posting his method and allowing us all to test it for ourselves. I have backtested 60 days worth of data (back to around the beginning of June). I realize this is not as extensive as OffshoreTrader's testing, but I did all testing manually in order to more closely simulate actual trading conditions (and also because I don't have a team of guys to do a more extensive mechanical backtest
). I will post my results here because they are significantly different from OffshoreTrader's. I donât know why my results are so different but I could speculate that itâs because a manual backtest allows for more discretion, or because market conditions over the past 60 days have been more conducive to NPPâs method as compared to market conditions during the previous 21 months or whatever.
A few caveats: I, like some others here, notice differences between my Chaikin Oscillator and NPPâs. I use eSignal and I found a Chaikin Oscillator file that claims to measure the difference between the EMAâs of a fast and slow Acc/Dist line (I set the fast at 3, and the slow at 10). I have found that while the values of my oscillator do not match NPPâs, the signals generated are nearly the same. I compared all the August trades with those of NPP and found 2 times where his oscillator showed a divergence, but on mine the divergence had not quite developed yet. Because I saw no divergences, I didn't take the trades. One of his trades was a loser and one a winner. The winner was larger than the loser by 0.5 or 0.75 points ES. Other than these 2 trades, other signals matched up well.
Also, while my Keltner channels are very close to NPPâs, Iâve noticed that they are not exactly the same. Sometimes this has caused my stop loss to be 0.25 points different from NPPâs. My Keltner numbers are close enough to NPPâs that the difference could be explained by using a different quote vendor. I am at a loss as to why my Chaikin values sometimes differ by a factor of 2 (even though the overall shape of the line looks very similar).
One more thing: I made two entries which could be said to fall outside of NPPâs methodology. Both were double tops/bottoms where the price did not quite make a new high/low, but came very close to doing so. The Chaikin, however, made a much lower high or higher low, and I read this as a clear divergence. Both double tops/bottoms occurred at what I considered significant resistance or support (IIRC, the R or S was provided by the previous dayâs high or low in both cases). Both trades were winners. Some may argue that these trades shouldnât be included in the test because no new high or low was made and, therefore, they do not fall within the bounds of NPPâs system. I claim that this is where the discretionary part of the system comes in and that these were valid trades. Maybe NPP can address this point. In any case, taking these trades better simulates the way I would use this system in realtime trading, so it makes perfect sense to me to include them in my personal backtesting results.
All trading signals were generated by the SP. I looked at using ES, but saw enough differences between it and SP that I ended up canceling my subscription to CBOT (I no longer look at bonds or notes regularly), and subscribed to the full CME data for only $5 extra per month. I did not look at ES to get the actual entry and exit prices; I assumed that these prices would match those of SP closely enough to not be a problem.
To reduce/avoid bias, I scrolled forward on the chart slowly once I saw ADX approaching 30. I made all entry and exit decisions while on the right edge of the screen, so I couldnât see the result of a decision before making it.
Because discretion is involved in this system, everyone will get different results from backtesting. As I said before, I offer my results to the thread because they are wildly different from OffshoreTrader's. Here they are:
Total number of trades: 35
Total number of winners: 22
Total number of losers: 13
W:L ratio = 1.7:1 (or 63% winning trades)
Total profits for winners = 106.75 points = $5231.90 per contract, including commissions
Total losses for losers = -33.25 points = -$1600.10 per contract, including commissions
Average profit = $237.81
Average loss = $123.08
Profit:loss ratio = 1.9:1
Net profits overall = $3631.80 per contract
Net profit per trade = $103.77
If anyone finds any errors in my calculations, please speak up!
Regards,
Jason
Edit: I should have mentioned that I use Interactive Brokers, so my commissions were estimated at $4.80 round trip.
