First some more studies about : -Daily indicators and tradedirection -Bull versus Bearmarket with ORBO System http://trader.gmxhome.de/breakout/breakouttest.htm and a new idea to define (real) Bull and Bearmarkets. In a Bullmarket there is a positive outcome with this rule: If monday closes negative, buy tuesdays open and exit wednesdays open. There is a positive moneyflow in this manner in Bullmarkets. To use some way to show the avgtrade of this strategy tells you if we are in a Bull or Bearmarket. If the avgtrade-indicator crosses under 0 then a Bearmarket arrived, since 1982 this happend only once at april 2000, the following crossover signaling the next Bullmarket did happen around Nov 2003. Seems to me the best indicator of Bull/Bear without Whipsaws i know. That seems a little bit out of the OR stuff, but as i mentioned, i believe the market behave different in Bull and Bearmarkets and that´s also relevant in a ORBO System. See the Ind. in the picture
Well I'll be the first to say it: The example trading system really took a bath today! A , fortunately, rare combination of a sharp reversal and and a large opening range combined to deadly effect. Are there any variations out there that managed not to get caught?
...aehmmm mine, my version i really trade, didn`t trade since 09/03/2006. Somewhat boring, but i like the shape of the equitycurve, and also for today i got a "red condition" to trade... As you can see in the pic in the left side (daily chart) i get every evening a green or red sign for the next day, the last 7 days i got a red one, NO trade. In detail i get a red or green (on 5 Min) after end of OR, this checks the conditions for OR and shows what side is allowed(if any...). Very boring stuff, but successful trading i heard is completely boring.... tomorrow is red condition, no trade Michael
If looking on some technical stuff, there is some massiv warning to go long, i track a 60min chart with a trendfollowing study of mine, but that´s of no interest here. The fractal dimension indicator seen on the chart is remarkably under the bottom yellow line, this is not seen a long time and means that this trend did run a very large distance and that some reaction is at hand with high probability of at least 1-2 days of a reaction back. I have to do some studies, but i expect this condition quite bad to trade breakouts in direction of the exhausted trend.
To the tradesituation on tuesday again, what we`ve seen was a exhausted/stretched situation on daily or 60-150min, under such conditions there ist a higher probability of exactly such kind of fake with sharp reversal, or any other erratic/reversal stuff. Fractal dimension on a medium timeframe seems to be useful, but you can have it more simple, just don`t trade under daily overbought/sold conditions, there is to much tension in the market. The results are better if not only exclude the tradedirection of the exhausted trend, don`t trade at all. By using a daily CCI(20) i found the very neutral numbers losing, means between -20 and 20 and the heaviest losers are in OS/OB numbers. A good rule would to take trades only if CCI(20) is not neutral and not overbought/sold. The pic shows the outcome if trading only under OS/OB conditions with a very basic system without filters. Michael
Unfortunately, combinations of large opening range and sharp reversal are not that rare. I remember having had three losses in a row with YM in december 2005: -40, -40 and -70 points. On each day a long was triggered, but then prices traveled down through the whole opening range and below - exactly like yesterday! As already mentioned before, a solution would be to abandon the opening range concept and use a kind of mid-day-range-breakout. Recently, I traded the ORB for YM from 9:30-11:45 EST without additional filters, but from now on I gonna trade a break from the 11:45-13:00 range with exit at 16:00 EST.
Hallo bounty, did you do some backtests with this strategy? Probably not... I did some studies with timerange breakouts and found no good enough results. Your Luchhour Breakout has a very low profitfactor, with the best stopsize of 0.3 * Lunchrange and 10,-$ commission/slippage you earn ~ 8,-$ per trade. Yesterday this thing worked fine, but the backtest over 18 months shows the truth. You have to search at least for some tricky filters.... See the results in the pic. A better idea would to extend the OR up to 1300, so you have now an afternoon breakoutsystem, this works better, but probably did not so good in 2002-3 in the higher Volatimes (just a guess). An unfiltered afternoonbreaker did ~ 18,-$ trade. Michael
@bounty Did some tests with this lunchbreaker, of course the ATR(5) Vola filter works the same good as for the example system (and may be there is some confusing und you already use it ) but the profits did happen only in some months and the rest is a mostly flat equitycurve. I found another condition, for going long the high of the Lunchrange must be some little portion of daily ATR over the openprice at least it has to be over the open at all, inverted for short. Well, additional rules: Yesterdays Range < DailyATR(5) * 1.4 Lunchhigh > Open + DailyATR(5) * 0.1 {long} Lunchlow < Open - DailyATR(5) * 0.1 {short} Stopsize = 0.5 * Lunchrange No mondays, no entry later 1530 Exit 1600 This gives ~ 48,- $ /trade, ~ 58% Win but no trade yesterday cause the lunchlow was not under the openprice.... Equitycurve in pic. you reach ~66,-$/trade with 66% win if open < close[1] for long open > close[1] for short but with reduced netprofit Michael
Hello Michael, indeed I didn't seriously backtest my idea, since I got only a few days of data for the current YM june contract. What kind of data do you use for longer term backtests? By the way, the profit/trade data can be misleading imo; because a high profit/trade combined with low trade frequency can lead to a rather low overall profit within a given time frame. Maybe you could post a comparison of net profits for both unfiltered lunchbreaker (11:45 - 13:00 EST) and afternoon breaker (09:30 - 13:00 EST) for the same time frame? Thanks for your efforts! I've attached a screenshot from the string of losses mentioned in my previous post.