So how do you use DMI/ADX indicators on the 80% of the trading days when the market is non-trending? Is it useful for little intra-day scalps or do you just switch to oscillator indicators?
abishiai, Unfortunately, if you have TS7 (since you metnioned the @ES symbol I am assuming this is what you use) then you are stuck with whatever historical data they give you. I don't know how far back they keep ticks, but I remember that it wasn't enough for me to keep my subscription with them. I think their biggest mistake in trying to convert 2000i users to the online version was not allowing for data import. I've never understood how the brainiac that made that policy ever got a job at a backtesting software company. A while back I read somewhere that they were considering a change to this but I don't know if it ever happened.
Look for divergences. Note, for example, that on the 0945 NQ bar there were divergences in the ADX, DM+ and DM-. Therefore, a short.
Abishiai, What are the ADX/DMI settings you are using on the 233 tick chart. I have esignal and could try to back test it if you tell me what your parameters are. Savage
I use DMI everyday and am profitable far more than 80% of the days, so non-trending markets are not an issue. I don't switch indicators regardless of market conditions. To answer your question, yes it does work for "scalping". The majority of my trades I'm only going for .02 - .04. Choppy days just suck because I don't often get the good moves of .05 - .10 or more. Also, generally speaking the market does trend for at least the 1st hour of trading and that is the time I generally get the bigger moves of .05+. I expect mid-day to be choppy (or non-trending) and accept that I'll only get the .02 - .04.
At this point I'm just using the default settings as WarEagle wrote them. However, it might be better to use a Fib number, since the 233 tick chart is Fib based. Maybe 13 would work better. I'll look at and see if I like it. Cheer, kp
nyc, Are you using premarket data for your caculations? Gaps can really throw the ADX off for awhile in the morning making the signals untrustworthy during what is often the best time of the day. Also, are you entering on the cross as soon as it hapens or do you require the 1 min bar to close with the cross intact? Any time I have tested crossing type signals, this has been a problem, since in backtesting you never get the false signals that are no longer crossed at the close of the bar. Thanks again. abishiai, I just used the setting (15) that nyc uses in the code that I posted.
OK, here's my magnum opus. These signal pics are discretionary, as I haven't been able to figure out a set of rules that determine when to or when not to take a signal. Basically, I want to go with the trend, so I don't take the little counter-trend bars that may occur within the trend. However, this system seems to be very good at telling me when a trend is over and when to go the other way. The circles on the chart would be trades with a 2 pt stop/target. Now a couple observations: It would be great to back-test and find the optimal stop/target so the risk/reward ration would be better. One should try to figure out a 'sweet spot' for entering the trade. It seems like in most cases after the signal is given there is some sort of pullback and you would then be able to enter a few ticks to the better. Please comment/improve the idea here. (I can't say 'system' obviously because this is very discretionary.) However hold the criticism. I already know I'm naive and dumb. Cheer, kp
Just to reinforce my point, look what the market did while I was massaging that last picture and posting it! Cheer, kp
I filter out the pre-market data. I start my chart at 9:30 when the market opens. I do not trade pre and post. Generally yes, I enter as ADX crosses to the upside. I say generally because, I always play the first call of the day as it happens and do the same through about 10:30AM or so. Mid-day I may or may not play at all.