I finished calculations for 15M, so far, the results improved for every pattern
butterfly with AB=CD comes to a amazing 92%, followed by crab with 86%, then gartley with 80% and finishing with bat at 72%.
The deepcrab is pretty good, about 90% but I just found 147 samples. The stats used data from December 12, 2011 to date.
I didnāt found any alternate bat yet, I donāt know if my code is wrong or not, so if you found an alternate bat, please tell me which pair, timeframe and date/time to look at it, and check if the program recognizes it and if not, fix the problem.
If you have got an EA that can scan these patterns with these resultsā¦ You should go ahead pack your bags and leave your computer running on your way to Hawaii.
When you say AB=CD, do you mean 100% of AD for CD leg, or can CD be 127.2, 168.% of AB? Also, what kind of convergence in pips are you allowing the system to trade on? Approximately how many trades on a given pair in one months time does the pattern make (for 15min, or 1hr tf)?
when I say AB=CD is that the pattern has a CD=xxxAB with xxx at least 100%, to 261.8%, both in time and price, and having a difference between the percent of price and time of less than 20%.
For example, if CD=127.2%AB for size and CD=138.2%AB for time, and the difference between 127.2 and 138.2 is less than 20%, so they are AB=CD for my program.
By the way, the results are for backtesting, live testing is different, and some time of live testing is necessary to get real information, until that, Iām going to consider the EA a dangerous thing for real money hehehhe.
On You first trade, I would say that it is not very symmetrical. Throughout your X-A swing, there are lows that almost hit your A level, so the XA is not really a swing that other traders would be looking for fib levels on. The key is to know what fib points the majority of traders have their eyes on. I believe that is why the convergence of both the large swing and AB=CD are so succesfull, because longer term traders are looking at the XA fib levels whereas shorterm traders are looking at CD. When you get the two groups of traders at one point, then that is when a diffident retrace-ment can occur.
Also note, somewhere in the thread, I believe TMB states, When BC is only a small portion of AB swing (.382), the pattern is less likely to be a Gartley, and more likely to be a pattern. Also when the BC is only a small portion of AB swing, then the AB=CD pattern will not be a 100% and is more likely to be 127.2 or 161.8.
Your EURJPY also looks like a bad setup because those dont look like very defiant swings. At least, that is my opinion. Someone with more experience could probably give you a better idea on that one.
I recently posted a EURJPY trade on meetpips. If you look at how I analyzed my trades on there, it may help you in finding trade-able patterns. Next week I plan to be posting a setup every other day and the results. PipMagnateās Channel | MeetPips.com
This pattern looks alright to me. The problems for this one may be the convergence selection. When a pattern converges at 1.0XA (double top), it maybe a warning that this convergence can be less tradable as the big guys often push the price further, so the next convergence is a better choice.
The patterns are evolving with the market, so even a pattern looks perfect, it would still be better to take other factors as further confirmation or warning, so you havve better chance to winning, such as candle patterns, price momentum and trendlines. If you go through the whole thread, you will find all the factors discussed in details.
If your system can pick up one or two patterns every day with this rate of success, it will be very profitable.
Since you have got exellent results with history data, you may focus more on live trading improvement. One way you could improve it is to use it for live trading first (demo), then trade it again afterwards with the same data as history test and analysis of the details of the successful and failure patterns so you can modify your code. If you put your results in the thread I am sure the community would be happy to help.
Hi guys,
I was monitoring the market this evening during the Asia session using the MT4 and MBT MT4 (special version of MT from my broker) I use the harmonic indicators to help me spot patterns, and then use eSignal to verify these patterns manually to make sure they r accurate.
Tonight around 8 pm my mt4 catches a bearish Gartley on USD/JPY as you can see in the charts I attached below. Then when I double check the pattern with MBT MT4, the pattern does not show up in the chart.(I use the same indicators in both MT4) Then I decide to verify the patter manually, and I find the pattern to be 5 pips off the fibb levels. I paid a closer attention to the charts and noticed that the tick info r slightly different in all three software which caused the inconsistency. So here are my questions:
I use free demo MT4, MBT MT4, and eSignal which all have different data feed providers. Is any of these more reliable than the others?
When price does not perfectly fall on the ideal fibb lines, like 5 pips away, is that tradable?
Can anyone check the indicators I use to see if they are reliable? I know I should train myself to spot the patterns myself, but I am still at the learning stage and I will have quite a hard time to spot the patterns right away.
Hi all! I first read Larry Pesaventoās books, then after searching in the internet for further information, I found this thread. TMB seems to be by far the most successfull Harmonic trader. Looking for the patterns, several times I found this (see picture): Price doesnāt always react at the best confluent fib levels (XA 61.8 and CD 138.2). Often price starts above (for buy) a good CD level (161.8 in this case), breaks through to the downside until a good XA level (here at 78.6) and then retraces above the CD level and goes our direction, EVEN THOUGH these levels have a certain distance (candle with arrow). I have found this as relatively reliable trades. Everyone else noticed it too?
Here is another example (sell): Price starts just below the XA 78.6 level, then breaks above to the CD 161.8 level, retraces back below the XA 78.6 andā¦ a big move. Although 78.6 and 161.8 arenāt really confluent. When this happens with good levels (by this I mean, that 61.8, 161.8, 78.6 are stronger levels than 38.2, 138.2 for example), often there will be a good move.
Ok, Iām having some difficulty understanding the proper location to take a trade based on symmetry. So on this trade the 2 hour GBP/NZD I found two divergences XA .618, CD 1.00 and the 2nd XA .786, CD 1.272. The first trade failed, while the 2nd trade was successful. My ? is why was the 2nd trade/convergence the right trade to take? Thanks again for all your help!
This is a Gartley as I see the ratios. The ārightā ratio for Gartley says that XA should be 0.786 and CD 1.272, so the first one was not a āperfectā Gartley.
Hi mulmil, the way to mast the pattern trade is to understand the basics and practice. This thread has provided excellent resource for understanding harmonic patterns and trade experience from TMB and others. I have read through all the pages and have learned a lot from it. For my understand, having confluent adds the odd that the price may reverse, but it does not mean that without confluence the price will continue. With confluent fib level, it does not guarantee you that you will get a reversal either. You need additional confirmations to have successful trade.
Bye the way, 0.707 is another fib level TMB used, which may be the case for your previous chart.
It actually depends on which definition you use . In this thread,it uses differently from some other traders. Apart from other rules, if D point is not beyond the X point, it will be considered a gartley (1.0 XA or less). IF it reached beyond the X point, such as 1.272XA, 1.618XA ā¦, it will be a butterfly. Butterfly is considered a failed gartley (the price did not reverse at or before reaching the X point).
The first two patterns are invalid gartleys as the CD did not reach 100%AB. The 3rd pattern is not valid as it was, but by moving the X to swings higher than B (the extrem high upleft), it may become a valid pattern.
If you are really interested in learning harmonic patterns, it would be better to use your eyes to find it. It maybe slow at the begining, when you get used to it, your will find it is an easy task. Really on the software may be not the way.
