There's a higher probability for price to continue rather than reverse.
nick203
⋅
@Alexander_Nikitin, considering the bat pattern is bearish what makes you say this ?
Alexander_Nikitin
⋅
@nick203, Market research makes me say this. I have tested hundreds of thousands bats and they have a higher probability of continuation than reversal.
But what makes you say the bats have a higher probability of reversal? You have read it in a book? You watched it in a video?
nick203
⋅
@Alexander_Nikitin, interesting! So far in my backtesting a bearish bat will have a higher probability of a bearish move and a bullish will have a higher probability of a bullish move. Maybe our rules/targets are slightly different though.
Alexander_Nikitin
⋅
@nick203, Of course target and stops may vary. If you set a 100 pips stop vs 10 pips target then you will have a higher probability to hit your target. But the expectancy of your strategy will be negative.
As for bats there are some rules that are neither yours nor mine. Bats creator, Scott Carney, says that the pattern exploits the retest of the X-point. He calls a retest the area of 0.886-1.13XA. He also applies this area rule to identification of double tops/bottoms and to sharks.
This means that the stop should go beyond 1.13 (preferably to 1.272) in order not to ruin the structure. If you set your target as standard 0.382AD then your stoploss will be greater than your takeprofit. If executed with limit orders bats have 48% probability of success averaging on 20 pairs. Of course, you can increase the accuracy if you use some reversal formations inside the PRZ. But ask your self a question: "If the profitability of the pattern depends on my execution, maybe it is not the pattern that generates money??" As you know Scott Carney is using HSI, RSI and amplified RSI indicators together with candlestick reversal formations to confirm the entries.
I am inclined to think that your backtest was not objective. When I first backtested harmonic patterns and advanced formations I also had very nice results. But when I switched to mechanical pattern recognition I found that in my backtest I was missing 2/3 of total amount of patterns. Subconsiously, human brain tends to skip something that doesn't work in order not to waste time - that's why majority of losing patterns weren't noticed.
I did my research analyzing the data period of 1999-2018. I was using wpr, cci, rsi, stochastic, prex, atr, pips and percent swing identification algorithms. As the average swing in all markets is approximately 3 atrs I set the period of algorithms to draw average swings close to this value. E.g. wpr had period 14, overbought -20 and oversold -80 and rsi had a period 5, overbought 70 and oversold 30. And so on... Different algorithms help to analyze the market from different angles. Luckily, MPS finds tens of thousands of patterns within a matter of seconds. I was only left to type down the results into excel spreadsheets. If you want to learn more about my market research you are welcome to visit my website: mps-trade.com/ Click "education" tab and then choose "market research". You will be asked to register but the information on my website is totally free.
nick203
⋅
@Alexander_Nikitin, I apreciate your answers Alexander. My stop loss is at a little bit different of a level, a bit higher. Also the expectancy is positive, meaning target/stop loss ratio is over 1:1. Also if the trade is moving against me I have a precise way of adjusting target. Also maybe our XA waves have different characteristics that improve the patterns reliability. I'm not at all trying to discredit your research in any way but right now the strategy is working for me. I will have a look at your website since I believe other opinions/information is important.