The Beast stirs. He's got a slope.

[Bunker]

Hello Fellow Traders,

Thanks Steve for your hard AND RAPID work.

Just like slowkey, Ver b also gave me a lot of small profit as far as CT goes. The trigger seems good and it is probably all about finding a good MM strategy. Pls, have a look on my thread. Today would have been a good day to use the strategy I posted for CT trades. I think using the 5.1 cross (back) for TP would have worked great today (instead of the 0.42).

http://www.stevehopwoodforex.com/phpBB3 ... 458#p19458


Cheers!

Bunker

[slowkey]

Hello Fellow Traders,

Thanks Steve for your hard AND RAPID work.

Just like slowkey, Ver b also gave me a lot of small profit as far as CT goes.

Bunker

Switching the ltf to the higher 240 minute time period will get you profitable with the Bunker immediately, no counter trend and no white box turn on. Just running Bunker all by himself. He just made 7 trades all but one profitable in the tester over a quick 6 month test. So we just need to find the right settings I think.

Maybe 240 and some different settings so I am going to run some optimization just to see if I can get some numbers to forward test from. None of the optimization on the 60 minute period seemed to give me any numbers at all profitable to forward test with.

Edit just noticed that on the setting on the last version of Bunker the ltf buy level are negative and the sell levels are positive. I do not know if I did this somehow or they were that way. Disregard the above until I can figure out what is going on.

[scartree]

fxozgirl
Thanks for your thoughts on this matter. I have set Use Max to 'true' and UseChaserL to 'false' as suggested by Macman in an earlier post. Being a coding klutz, I can not offer any suggestions as to reasons for the differences on our charts. It would seem that our inputs are the same. I have dropped Bunker and returned to SlopeyBeast for now using Steve's revised v1c. The only change from OOTB settings is the lot size; I have changed that to .02. The sixth lines are drawn by the EA and I will assume that they are correct for now.
Karl

[slowkey]

I have been staring at this setting page to long. They look fine the way they are I do think and work at 240.

To many settings. Goofed again.

Oh rats ... I am going to get some food and a shower and come back to this when my sanity returns.

I have been at this to long.

[SteveHopwood]

I am going to attempt to answer this one, as it concerns IMHO a critical part of "The Beast".

As some you will know, TB started life as macman's test EA. It was originally designed by me to be very fast & lightweight, nothing like the monster it became -

The key to the EA was to be how many bars were used in the look back routine, to decide how the EA should react to the current market situation. As you all know, we can't predict the future, so the best we can do is base the actions of our EA on a relevant past period.

My original concept was to have a dynamic system to adjust the look back period for every pair.
For instance we might find a currency pair behaves similarly most years in the summer quarter, so we could set that currency pair to use that look back period. If a currency pair is 'flavour of the month' (safe haven) for instance, we might want a substantially shorter look back period etc.

When Steve asked if it was OK to release the original into the forum (then on FF), I agreed, hoping that the idea of a dynamic input routine would be taken up. However it was not to be and instead ever increasing bits were bolted onto an unfinished concept.

OK, enough of the history lesson, let's look at the code.

Firstly the reference to Max is slightly wrong, so don't bother to try and look it back, as Max supplied a routine for adjusting ReEntryLinePips (part of Recovery), nothing to do with the look bar bars used.

The statements like this one below use my default number for bar count & BTW the bar count for AUDNZD should probably be 120 not 20 - my bad ...
if (StringSubstr(Symbol(), 0, 6) == "AUDNZD") { BarCount = 20; }

Just out of interest, here is a complete set of optimised results from early 2011:
if (StringSubstr(Symbol(), 0, 6) == "AUDCAD") { BarCount = 142; }
if (StringSubstr(Symbol(), 0, 6) == "AUDCHF") { BarCount = 196; }
if (StringSubstr(Symbol(), 0, 6) == "AUDJPY") { BarCount = 98; }
if (StringSubstr(Symbol(), 0, 6) == "AUDNZD") { BarCount = 76; }
if (StringSubstr(Symbol(), 0, 6) == "AUDSGD") { BarCount = 170; } //not optimised
if (StringSubstr(Symbol(), 0, 6) == "AUDUSD") { BarCount = 104; }
if (StringSubstr(Symbol(), 0, 6) == "CADCHF") { BarCount = 176; }
if (StringSubstr(Symbol(), 0, 6) == "CADJPY") { BarCount = 72; }
if (StringSubstr(Symbol(), 0, 6) == "CHFJPY") { BarCount = 107; }
if (StringSubstr(Symbol(), 0, 6) == "CHFSGD") { BarCount = 170; } //not optimised
if (StringSubstr(Symbol(), 0, 6) == "EURAUD") { BarCount = 161; }
if (StringSubstr(Symbol(), 0, 6) == "EURCAD") { BarCount = 217; }
if (StringSubstr(Symbol(), 0, 6) == "EURCHF") { BarCount = 292; }
if (StringSubstr(Symbol(), 0, 6) == "EURGBP") { BarCount = 224; }
if (StringSubstr(Symbol(), 0, 6) == "EURJPY") { BarCount = 221; }
if (StringSubstr(Symbol(), 0, 6) == "EURNZD") { BarCount = 131; }
if (StringSubstr(Symbol(), 0, 6) == "EURSGD") { BarCount = 170; } //not optimised
if (StringSubstr(Symbol(), 0, 6) == "EURUSD") { BarCount = 209; }
if (StringSubstr(Symbol(), 0, 6) == "GBPAUD") { BarCount = 34; }
if (StringSubstr(Symbol(), 0, 6) == "GBPCAD") { BarCount = 214; }
if (StringSubstr(Symbol(), 0, 6) == "GBPCHF") { BarCount = 192; }
if (StringSubstr(Symbol(), 0, 6) == "GBPJPY") { BarCount = 50; }
if (StringSubstr(Symbol(), 0, 6) == "GBPUSD") { BarCount = 150; }
if (StringSubstr(Symbol(), 0, 6) == "NZDCAD") { BarCount = 276; }
if (StringSubstr(Symbol(), 0, 6) == "NZDCHF") { BarCount = 300; }
if (StringSubstr(Symbol(), 0, 6) == "NZDJPY") { BarCount = 298; }
if (StringSubstr(Symbol(), 0, 6) == "NZDUSD") { BarCount = 214; }
if (StringSubstr(Symbol(), 0, 6) == "SGDJPY") { BarCount = 170; } //not optimised
if (StringSubstr(Symbol(), 0, 6) == "USDCAD") { BarCount = 117; }
if (StringSubstr(Symbol(), 0, 6) == "USDCHF") { BarCount = 269; }
if (StringSubstr(Symbol(), 0, 6) == "USDHKD") { BarCount = 170; } //not optimised
if (StringSubstr(Symbol(), 0, 6) == "USDJPY") { BarCount = 233; }
if (StringSubstr(Symbol(), 0, 6) == "USDSGD") { BarCount = 170; } //not optimised

So, if you are not running your own optimisation routine and updating these numbers as the market changes, and as far as I know, nobody is - then the "UseMax" input is a simple default.

BTW the information in the CAT section on my website at fx-works.com comes from my own version of TB, which was being optimised quarterly.

Later Chaser added a routine ("UseChaserL") which uses CCI to add some dynamic content into the look back bars being used. However as it simply uses the last 720 (hourly) bars, irrespective of what the market is doing as it's basis, it too is a best guess.

When you see posts by traders saying they have been profitable using x,y & z currency pairs, they are effectively saying that these currency pairs best fitted the look back bars being used - either CCI 30 days (720 hours) or my default values.

Before even more routines are added, it might be an idea to take a step back and see if we can improve TB in this area. This is a new forum and there are some brilliant minds around here, so I look forward to seeing some creative thinking. It is also possible that the system as it is now is good enough - I don't know.

You now have a bit more information to add to your knowledge of The Beast ... so go forth and conquer

macman, you know what? I am coming to the (early and possibly premature) conclusion that it does not matter how the extremes are calculated. Early experiments with SB are showing this: counter-trend sell trades are being taken when D1 Slope is showing buy-and-hold. Counter trend buy trades are being taken when D1 Slope is showing sell-and-hold. Much is explained by the D1 Slope.

Given that counter-trend trading has ultimately resulted in disaster for me and that I ditched The Beast a year ago because of this, this comes as no surprise.

This is not to say that SB has no future, but that all this confirms my long-decided opinion that fighting the trend is a waste of time. My only problem has been finding a method of recognising a trend that works for me. I have tried a few thingies; D1 Slope does it for me. Interesting that my previous fav D1 Rsi(20) does the same, so long as it is >=60 and <=50, conclusions that I was coming to a few months ago.

[SteveHopwood]

I have been staring at this setting page to long. They look fine the way they are I do think and work at 240.

To many settings. Goofed again.

Oh rats ... I am going to get some food and a shower and come back to this when my sanity returns.

You think this forum leads to sanity? You need therapy. Harry has some to spare.

[slowkey]

I have been staring at this setting page to long. They look fine the way they are I do think and work at 240.

To many settings. Goofed again.

Oh rats ... I am going to get some food and a shower and come back to this when my sanity returns.

You think this forum leads to sanity? You need therapy. Harry has some to spare.

I definitely need to pull my self away from the computer long enough to imbibe a nice brew and a meal so I will be sparing Harry his senses. The Beast will have his way with me no mater what

[spyderman]

I'm getting a continuous loop of the following in my experts tab:

2012.07.17 06:50:18 Slopey Beast EURAUDlc,H1: unknown ticket 1867462 for OrderClose function


Keeps cycling two previously closed ticket numbers. Anyone else getting this? :

[slowkey]

I'm getting a continuous loop of the following in my experts tab:

2012.07.17 06:50:18 Slopey Beast EURAUDlc,H1: unknown ticket 1867462 for OrderClose function


Keeps cycling two previously closed ticket numbers. Anyone else getting this? :

Well I am seeing a repeating error in the journal in the tester very similar. When I put Bunker on a ltf of 240 he becomes profitable but I get these two errors. Which leads to a loss of over 653 dollars because this one ticket does not close. Otherwise Bunker is profitable on 240 time period. Small little profits and a few bigger ones. Nothing I looked at on the 60 minute period looked good. Slope breaks down below 240 it seems which is what I expected. The 60 minute is 1/24 of a Day candle, just to big a ratio.

Edit actually it happens with more than one ticket but there is usually one that opens and does not close for the period of the test.


2012.07.17 09:00:32 2012.07.12 11:34 Slopey Beast EURUSD,H1: unknown ticket 58 for OrderClose function

2012.07.17 09:00:21 2012.01.10 11:59 Slopey Beast EURUSD,H1: OrderClose error 4108

[slowkey]

Oh Steve,

I think I will be optimizing the settings of the White Box and trying to see if I can get a little more productivity out of it. There is gold in that there box. If it can forward test half as good as it looks here.

We may have ourselves a little bot inside the big bot to separate out. I got no long trades but the trades were of a long duration and the Pair has been headed south most of the year. I am going to setup a demo and just test the WB on about 25 Pairs.