...fading afternoon breakouts/trends is a fools game, also look at Linda Raschke about this stuff, or test it..... with optimized parameters, 20,-$ commissions/slippage, 1 trade/day the best timewindow to trade was between 1015 and 1200, the very worst was any after 1300. But with your original exit this was not profitabel either....
the exit after CCIAV cross on 5 min is crap, you get whipsawed most of the time, it`s useable if you want to save some nice profits you already got. Changing the rule to use this exit only if High of bar > entryprice + 4 points Short inverted rule. This was profitabel in theory, you see the elipse on chart where you would have been whipsawed without this add. rule. 4 points are to simple, it`s just an example, you have to adjust to volatility, and you need more, may be 4-8 different exitrules, depending on marketaction, including a timestop for example, if not profitabel after X bars then exit, etc. trading from 1015- 1200, optimized parameter, one trade only, the above rule about cciexit, 20,-$ comm/Slippage did ~3500,-$ but only an avgtrade of ~ 11,- $ , not realistic to trade till now, you need at least ~ 35,-$/trade to survive the game, and did i vergot the 11 consecutive losers?
Before going further with the system, we should know why this thing did perform well in march und very ugly in other months. On 15min, i colored the days, you can see very wide and fast, often complete dayreversals without any remarkable retracements. Of course you could make money with reversalplays. That`s a somewhat unusual, extrem behavior, not very often in this frequency in one month, and it`s very unlikely to continue... The most obscure is the LARGE width in points of this Reversals, making good advance over the other losing days. In more normal environment we see reversals to, but not in this big size, so the ratio of Win/Loss is much more worse in a more "normal" market.
In other months, or more usual months, you have reversal to, but not big enough to compensate the loss in trenddays. Here we see the Nov. last year. First you can see a lot more trendiness, more trenddays, and the reversalsdays have much less points then we had in march. Reversalplay was a loser here...
The most logical things don`t work... A stop 1 Tick over/under HOD/LOD is used by most amateurs and you often get faked out by the big boys by stopfishing... By adding some more Ticks , this thing gets more profitabel, a Stop ~ 0.75-1.0 point from HOD/LOD makes a lot of difference. Best testresult 0.75 points from HOD/LOD 6447,- $ with avgtrade 20.67 $, but the last 3 consecutive months were losers nonetheless.
...a 30 Min CCI over it`s average by just, lets say 1 or 2, or even just 0.4, means nothing, it`s more or less congruent with the average, that does not define any Oversold/bought condition. By using a trigger threshold, means the CCI has to be some numbers over/under it`s average it`s more meaningful, this reduces #trades and makes a better avgtrade results, 20 was a nice number, means 30min must be by 20 over it`s average to define overbought or 20 under it`s average to define oversold. Makes additional ~ 3 $/trade and little bit more netprofit.
A short Replay: The original approach with up to 5 entries/day, originalrules, trading every time of day, AND 20,-$ Comm/Slip. did something around MINUS -17000,-$ the described version did PLUS 6800,- $ If you are not testing strategies you are going the suicidal way....100%
Obviously not a man of your word. Your rightoues indignation and attitude are a total drag. Thanks for keeping us all safe from ourselves. And in reality it is the way to develop a system. Start with an idea test some stuff trade some stuff change some stuff. Make 'em pretty, Chris
Well, Golly Gee MechTrader And how much money do you make? How freakin typical. Methinks Mechtrader is harpin because he tried to complicate and over optimize his own trading and has lost his ass. If this simple method really does make money he will feel like a fool for not thinking of it himself.
mechtrader, can you post the easylanguage code as a text file? I'd love to be able to run some backtests but haven't been able to get the same trades as the ones I take live. I've got lots of ideas and if you've got the code I don't mind testing them all and posting the results. Afterall we're both after the same thing...a great, stable, trading method.