Here it is - I've included the P&L for each trade (commissions rounded up to $5/trade) so you can see where I did good and where I f#&*ed up. db, No worries, it's all constructive debate. Now, in your earlier post, are you talking about Emini NQ? 'Cause it sounds to me like the big contract, with resistance at 1001.5 etc. And what's a 2B? hii, I have no problem with your suggestions, I know you mean well. And your comments are really helpful, thanks. There are two reasons I didn't start trading SSFs and I'm blowing my money on Eminis instead: I don't feel that I completely understand how to trade SSFs yet, and secondly, because I have this mechanical method that backtests like crazy every day, and I know I can stick to it. I have to prove it to myself. If I get burned again, I WILL size down, most likely to SSFs. Good trading tomorrow guys.
Yes, the reference was to the emini. As for a 2B, it's a means of trapping people on the wrong side of the trade. Below is a copy of the post I made at the time: For future reference, today provided a good example of a 2B. First, there's a TL break just before lunch. This tells people that momentum is slowing and that there may be a trend reversal in the offing. However, around 1245, a new high is made. This puts everybody who entered shorts or exited their longs either on the wrong side of the market or out of the market altogether. So they either cover their shorts if short or enter if out. Ten minutes later, however, price drops below the previous high, This puts everybody on the wrong side of the market again, upping their "fear response". There is a feeble effort during the next 10 minutes to move the price up again, but this fails quickly and price drops decidedly below the 1048 high. This can be an excellent place to enter as it places the trader so close to resistance. However, if one wanted to wait for evidence of real weakness, price hit 990.5 three times (on the 3m chart) between 1330 and 1345. Dropping below that level would not be viewed with glee. The long expansion bars which began at 1400 are not typical, or at least they haven't been for quite a while. Perhaps there was a sell program. Who knows? But a sell stop below 990.5 would have been legitimate regardless of the reason for the ramp down due to the 2B setup (note also the rally attempt after 1345; the failure of this just feeds the fear of those who on the wrong side of the market). Note also that we came within a point or two of target, on both ends. --Db
I did two trades today, followed the rules but got stopped out with a nasty slippage. P&L=-$130 I stopped trading for the day, as I was by no means composed and objective anymore. Watched SSFs for a while and stalked what could've been a nice short in QLGC - from 12:34 to 13:10, down about 0.50 - hii, you must've seen this too. One thing that backtesting with StockWatchPro doesn't account for is the stops - if my initial stop is at 1.5 NQ pts, and I get stopped out 2-3 times (not to mention slippage) I'm already down over $100. So I gotta try a backtesting program that allows for stops too - any suggestions? Meanwhile will keep watching and maybe even try SSFs.
You've mentioned backtesting a few times but you haven't posted the results. Since you keep plugging away at your particular system, I assume you were satisfied with the results of your backtesting. If so, how was the backtesting done? What were those results? --Db
db, I backtested an EMA crossover ( 20 and 5). Entry - open of the next 1 min bar after the crossover (long if crossing up, short if crossing down) Exit - open of the next 1min bar after crossover in opposite direction. Backtesting results - profits of $50-200/day per NQ contract traded. If the daily trend is up you get higher profits for long positions and lower for short positions, and viceversa. This turned out to be unrealistic, because of the large number of whipsaws if the stop is set at 1.5 NQ points away from the entry. Also because of slippage/hesitation, the fills at the entry are never at the opening price of the next 1 min bar after crossover, but rather at its close or even at the high/low, which results in smaller profits /larger losses. Changing the backtesting parameters to entry on the close of the next 1 min bar reduces the predicted profits to about $50-100/day - still very good. Another problem with the entry signal is that it's lagging price action. It generally occurs after 4-5 bars up/down, and the correction that usually follows is likely to stop me out. I've also tested variations of the SMA crossover, such as: next bar after crossover, or the slope of the 20 EMA must be in the direction of the trade. It doesn't make much of a difference - it reduces the number of whipsaws but it also misses on some of the big moves. Now I've downloaded Wealth Lab (what a great piece of software) and I'm trying to program the strategy including the stops.
[ So, if I understand you correctly, the results you got from backtesting were not realistic because the entries you got in real trading were different enough from what you guessed you'd get - in backtesting - that you got stopped out more often than anticipated? The problems you've encountered with MA XOs are exactly what I encountered, which is why I abandoned it. But you might be able to make it work where I couldn't. For one thing, consider increasing your stop. Even 2pts is so narrow that you can be taken out by a hiccup. If you're going to be taken out, you want to be taken out for a good reason, e.g., a reversal. You might also want to consider increasing the bar length so that the pullbacks aren't so damaging, e.g., using a 3m or 5m chart. Or you could take your signal from the 5m chart and enter using a 2m chart. Ever thought about abandoning the MA XOs entirely and trading based on support and resistance? --Db
That's right. I've programmed the same strategy into Wealth Lab, but adding the stop losses and it's still profitable (about $120/week per NQ contract). The funny thing is that WITHOUT stop losses the strategy is 3x more profitable and the maximum drawdown is 2x less. You get a few bigger losers but more big winners too. So you're actually better off not taking the stops if you can afford the max drawdown (-$470 over 8 weeks tested). Taking signals from the 5m chart is also profitable, with profits of about $150/week with stop losses, and 50% more without. The advantage is significantly fewer trades. The max drawdowns are comparable at $350 over 8 weeks tested. I will play a little more until I find something that I like then I will simulate it for a while to see how it performs. I did think about levels trading but that is too discretionary for me yet, I need a simple strategy so I can work on my discipline.
Haven't been quite following strategy. Plus the small account size - a few losses in a row can be fairly demoralizing.
Couldn't trade today as the cable provider had maintenance scheduled for early Mon morning. It was supposed to be done by 5 am, but it's not up even now. They better fix till tomorrow.