Reversi: a dynamically calculated x-candle reversal EA by Steve Hopwood

Quoting the-game

Disliked

Nice observation... maybe you can incorporate the statistics for how often IB's form after x same coloured candles to further confirm/deny this trade trigger?

Ignored

Quoting f451

Disliked

I may have a go at coding up and presenting the statistics for this pattern. If I do I'll try to expand the framework behind Reversi to include candle range as definable aspect of the pattern, and then allow us to select which patterns to trade.

Currently Reversi works like this
determine statistics for reversal after consecutive candles in same direction
user specifies number of consecutive candles (X, = 6 OOTB) after which they want Reversi to trade an anticipated reversal

Ideally it sould look like this
determine statistics for reversal...

Ignored

Quoting f451

Disliked

I may have a go at coding up and presenting the statistics for this pattern. If I do I'll try to expand the framework behind Reversi to include candle range as definable aspect of the pattern, and then allow us to select which patterns to trade.

Currently Reversi works like this
determine statistics for reversal after consecutive candles in same direction
user specifies number of consecutive candles (X, = 6 OOTB) after which they want Reversi to trade an anticipated reversal

Ideally it sould look like this
determine statistics for reversal or...

Ignored

Quoting magft

Disliked

I really like the sound of this, i did start to look in to this last month but not had enough time recently.

For refence i ran reversi last month til the demo ccount expired and made 39%. I was using M15 charts on several pairs, i can provide set file at mo as on my other pc that is not switched on at mo as i recently moved house so nowhere to plug in yet! I chose M15 as it seemed to give a good balance between frequency of trades and no of pips gained. I used Squalou's indy http://www.forexfactory.com/showthre...75#post4629175...

Ignored

Quoting SteveHopwood

Disliked

I missed this somehow when you first posted it.

This sounds like a brilliant idea. Have you made any progress?

Ignored

Quoting f451

Disliked

I have made *some* progress. I was going to do a little backtesting to iron out any bugs (others those I've already fixed) and select some likely parameter values for thresholds, but what the heck... Let me introduce TrendStat.

TrendStat is the first building block on the ideas outlined in my earlier post. Currently it only addresses the probability of continuations or reversals after a sequence of n up or down candles - so it ignores volume, candle range and other potential factors which we might like to consider later.

How does TrendStat...

Ignored

Quoting magft

Disliked

I really like the sound of this, i did start to look in to this last month but not had enough time recently.

For refence i ran reversi last month til the demo ccount expired and made 39%. I was using M15 charts on several pairs, i can provide set file at mo as on my other pc . . .

Ignored

Quoting magft

Disliked

I really like the sound of this, i have been looking into candlestick patterns for marker reversal patterns. This has taken me on to Harmonic trading looking at patterns like butterfly, gartley, bat etc and there is a lot of info over on TSD if your interested and a MT5 EA that works very well.

I will stick this on a demo and see how it goes but I like the sound of the methodology so lets see how much we can squeeze out of it.

Thanks for all your hard work i know how long these things take and which i had more time at the moment.

Mike

Ignored

Quoting acostafulano

Disliked

Hi Magft:

Were you using the defaults, only on M15, or did you make any other changes to the settings? I've tried M15 before but my results weren't nearly as good (also, you were using 0.01 lots, so 39% is VERY impressive)

Ignored

Quoting f451

Disliked

Thanks Mike.

Word of warning though - there may still be bugs with the trade execution logic - I've fairly hammered the sequence and probability calculations, so I believe they are right, but I haven't backtested and visually examined the trades TrendStat takes to confirm they are as per intent. Been a bit busy with another trading project.

The harmonic EA sounds very interesting - can you post or send me a link?

cheers
f451

Ignored

Quoting f451

Disliked

I have made *some* progress. I was going to do a little backtesting to iron out any bugs (others those I've already fixed) and select some likely parameter values for thresholds, but what the heck... Let me introduce TrendStat.

TrendStat is the first building block on the ideas outlined in my earlier post. Currently it only addresses the probability of continuations or reversals after a sequence of n up or down candles - so it ignores volume, candle range and other potential factors which we might like to consider later.

How does TrendStat...

Ignored

   //Long 
   //if (TradeDirection == up && !DoesTradeExist(OP_BUY) )
   //if (TradeLong && TradeDirection == up && !DoesTradeExist(OP_BUY) )//f451 original code
   if (TradeLong && TradeDirection == up && !DoesTradeExist(OP_BUY) && !DoesTradeExist(OP_BUYSTOP) )//SH mod
   {
      //Balanced pair trade filter. 
      if (UseZeljko && !BalancedPair(OP_BUY) ) return;
      if (TakeProfit > 0) take = NormalizeDouble(Ask + (TakeProfit * Point), Digits);
      if (StopLoss > 0) stop = NormalizeDouble(Ask - (StopLoss * Point), Digits);
      SendTrade = true;
   }
   
   //Short
   //if (TradeDirection == down && !DoesTradeExist(OP_SELL) )
   //if (TradeShort && TradeDirection == down && !DoesTradeExist(OP_SELL) )//f451 original code
   if (TradeShort && TradeDirection == down && !DoesTradeExist(OP_SELL)  && !DoesTradeExist(OP_SELLSTOP) )//SH mod
   { 
      //Balanced pair trade filter. 
      if (UseZeljko && !BalancedPair(OP_SELL)) return;
      if (TakeProfit > 0) take = NormalizeDouble(Bid - (TakeProfit * Point), Digits);
      if (StopLoss > 0) stop = NormalizeDouble(Bid + (StopLoss * Point), Digits);
      SendTrade = true;      
   }

Quoting magft

Disliked

I have just checked my results running trendstat ootb on a $50k account with 5 lot size (GoMarkets wouldnt let me use a $1k with 0.1 lotsize!), it puts on trades until not enough funds to add more so typically 5 trades at a time. But account now at $96K, had 2 EJ trades open just before it shot up 300pips and made a nice $40k profit! I noticed tht NU had a few possible trades but no funds so if we can change the logic to limit to 1 trade per pair at a time then this looks very promising!!!

The logic in the LookingForTradeOpportunities needs resolving,...

Ignored