A couple of notes about this morning: I missed a retracement trade at 10:07 am that probably would have been good for about 10 points. Also, there was technically another retracement entry at 10:17 am. However, this entry occurred right at support at 9333 that was established at 9:38 am. In addition, it was the second retracement in the series, and also occurred right after a long signal at the low of a range. This raises a couple of points regarding the method.
On retracement entries, I will only take the first retracement entry in a series. This is because by the time a second retracement occurs, the trend has already gotten a little long in the tooth.
The other modification I am going to make is for range trades. I will only take range/reversal trades if price has met or exceeded prior support or resistance on the 5-minute chart. For example, if the prior low on the 5-minute chart is 9330, then price must hit 9330 or lower before I will take a long entry signal for the trade. The reason for this is because there are often a lot of resting orders around these levels which act like a magnet, so there is a good probability of another push up or down to these levels if you try to anticipate the reversal and take a signal prior to them being taken out.
Regarding changes to the method. Iâve found that in the past when trying to develop and improve a system, itâs easy to keep piling on new filters and indicators to try to correct for every shortfall and false entry signal that the system gives (over-optimizing). Iâm trying to avoid using all kinds of derivative indicators and such, since that approach has never seemed to work for me. What it ends up doing is filtering out as many good signals as bad, and the system ends up having just as bad of an expectancy (or worse), than the one I started with. Therefore, Iâm going to try to avoid adding too many new indicators (although I reserve the right to add one or two). For example, Iâm considering something simple like a 3 period RSI on the 3 or 5 minute charts to help identify ob/os, but in general, I want to keep it simple, if for no other reason than I have a difficult time watching so many things at once. Rather than use additional indicators as filters, I prefer to use price criteria (such as testing or exceeding support/resistance in conjunction with the entry signal).
As I mentioned in my earlier post, I have some issues, reservations, etc. that I am working on and thinking about.
First of all, I realized that no backtesting has been done on this method, so I went back about a month and looked at some charts. A cursory visual exam revealed that from a mechanical trading point of view, this method sucks. There are a lot of false signals, whipsaws, etc. However, from the beginning (in Baggerlordâs post), it was presented as a discretionary system and the issue of mechanical vs. discretionary was discussed in another thread:
Emini divergence journal
Therefore, Iâm going to make slight modifications as I go along based on what Iâm seeing, and then go back and visually review historical data to see how these changes would have affected the performance, as well as forward testing them with paper trading.
Also, I never really considered the question, âWhat is the theory behind this method?â
Basically, the HMA and LMA serve as both a momentum, and as an overbought/oversold indicator. In a retracement trade, having two bars completely above/below the moving average is an indication of high momentum, and then we wait for a retracement to re-enter in the direction of the trend. In a range trade, a bar that is completely outside of the appropriate MA is indicative of an overbought/oversold condition (when occurring at prior support/resistance), theoretically giving you a slight edge on the entry. One thing that doesnât make sense to me about the parameters is why the system uses a 10 sMA of the highs, but an 8 sMA of the lows for the signal moving averages. I suspect that what Mr. Bernstein did when he created the method (or copied it or whatever he did), was use odd-ball numbers like 34, 10 and 8, to make it look like there was some magical quality to the averages he chose to use. Perhaps if you believed there was a bias one way or the other you could use different averages, but I donât see that the system would be affected one way or the other if I used a value of 8, 9 or 10 for both of the averages, so I may make that change in the near future.
Another issue that Iâm aware of is that the Mini-Dow futures is still very thin, relative to the ES and NQ, so I donât have a realistic idea of just how bad my fills will really be. I tend to give myself fairly bad fills anyway, but it could be even worse in real life.
Finally, because of the small point value, commissions are very high. At $5/rt, commissions equal 1 point/contract traded. That can really eat in to profits or add to losses. However, for purposes of this exercise, I want to use the YM because the small point size will make it easier for me to ignore my emotions if/when I eventually try to trade the method for real. If I can at least pay for commissions and maybe make a bit more besides (over a period of months), then Iâll feel that this exercise was a success. So, with all that said, here are the updated rules:
Updated Rules as of 8/18/03
Chart Setup:
1 minute chart
34 eMA
10 sMA of highs (HMA)
8 sMA of lows (LMA)
Also, use a 3 or 5 minute chart to identify key intraday support/resistance levels.
Pullbacks:
Long: Wait for two consecutive bars completely above the HMA. Wait for a lower high and go long on the breakout of the high of the prior retracement bar.
Short: Wait for two consecutive bars completely below the LMA. Wait for a higher low and go short on the breakout of the low of the prior retracement bar.
Range Trades:
Long: Wait for price to meet or exceed (go below) prior support level on the 5-minute chart. Wait for a bar completely below the LMA. Go long at the break of the high of this bar.
Short: Wait for price to meet or exceed (go above) prior resistance level on the 5-minute chart. Wait for a bar completely above the HMA. Go short at the break of the low of this bar.
Stops: For range trades, place a stop above/below the pivot that is formed prior to the entry. Similarly, for pullbacks, place a stop above/below the entry bar. Trail stop behind subsequent retracements until there is 10 points profit in the trade, then begin trailing stop 3 bars behind the current 1 min. bar.
Profit targets: Take profits at key support/resistance levels, or tighten stops to behind previous 1 min. bar when approaching these levels. If trading multiple contracts, I can take profits on a portion at key sup/res, and leave some extra room for the remaining contract to run (since price can be expected to correct at these levels).
For range trades, the range should be at least 15-20 points to leave enough room to make some money on the trade.
Volume: Volume should be at least 150/minute. I'll measure this by using a 5 period sMA of the 1 minute volume.
