I am playing around with a system variation/filter. Still trying to work this one out, but the idea is to process trades against the pivot. (The system itself does not include any pivot logic, it is not a trend system, looking to capture some of the range.) 1) process short only orders when in upper pivot ranges 2) process long orders only when in lower pivot ranges The chart below is today. There were some buy orders that got thrown out (and good ones, but we were banging away on the upper pivots all day.) The pivot calculation is my own concoction (maybe you can figure it out) the chart below is firmly between 1st and 2nd resistant points. 3) Cancel with option to reverse if price fluctuates heavily above or below the most extreme S/R (as in fomc last week) Happy with the system as it is, however it sends way too many orders for comfort level and I overmonitor. With the pivots I have a slower system. Now, mentally it is hellish to watch trade (countertrend is not my preferred method) as it is constantly betting that supp will hold or resis will fail. Always making one-way bets on an up day or down day. Any ideas or experiences on this? Is it foolish to add this kind of filter or should I explore other filters (volatility, trend intensity...). Still thinking about this one. <img src=http://www.elitetrader.com/vb/attachment.php?s=&postid=1620762>
06, I have added simple filters, the reduce risk kind as in modifying quantity up/down, a few time ranges for news, optional days, turn off after x points. A labyrinthine of filters/options. But I do not want to over fit so I have a perfect system...one that filters all orders out
Walk forward test continues with the S/R filter. Only one order, allowed Globex session, profit 3 points. (There was another short order near 2pm with a server error during the same time, I manual discarded). The last order was time filtered. Attached the S/R lines today, they calculated unusually tight. With the problem stated in previous post, now just thinking that I could allow more orders if the pivots have calculated tight that day and less if they are wide (!?) I need to think about this some more, when the pivots are tight we have a tight range day, I posted that on the ES Journal thread and that's what we got. Back to the thinktank. <img src=http://www.elitetrader.com/vb/attachment.php?s=&postid=1621881>
I will update today testing and chart this weekend. There were two orders. One I caught, held manually and eeked out 2 extra points at the day open. The second I cancelled and didn't take and it was the second best of the day. Most likely because I was blowing my head off over in the ES Journal. Getting in my own way my own fault...but it amuses me during the boring chop. I didn't enable the pivot filter. However, if I had, neither of these orders would have executed. So roughly 9 points compared to 0 points. (Guesstimating the points since I don't know what the profit engine would have randomized the profit points to, starts at 4) So all in all, I am not happy with this day, since the orders would not have executed unattended due to pivot. Filtering sell orders if too close to the bottom of the range needs tweaking. Right now I am looking at Mark Fisher's book on ACD (Logical Trader), and the section on pivots, not the pivot calcs (mine are unique) but the idea of a plus or minus day based on where the opening range is (I think, haven't finished it yet). This interests me because I had thought of something (maybe) similar (but not as bright lol), just a visual check of where the opening price was compared to where the pivots were calculating that day...I mean to say when they opened even slightly lower, a couple points or more, it was easier to sell against the lower range - of course none of this empirical right now. I will post a chart and additional thoughts this weekend. The goal is to let more of the good orders go through without upsetting the cart.