Going Home - a multi-pair auto-trader for Dreamliner's strategy, by Steve Hopwood | Page 11 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

building multi-pair multi-timeframe scanner 48 replies

multi-pair percentage scale charts 28 replies

Price info indicator needed - one pair's price x another pair 5 replies

Multi-pair trading or Few-Pair Trading. [poll] 15 replies

Need help modifying Multi Pair Stochastic indicator 0 replies

Attachments: Going Home - a multi-pair auto-trader for Dreamliner's strategy, by Steve Hopwood
Exit Attachments

Going Home - a multi-pair auto-trader for Dreamliner's strategy, by Steve Hopwood

  • #201
  • Feb 6, 2011 11:56pm Feb 6, 2011 11:56pm
Steve,

Different platforms have different pairs available, so having a fixed list can cause some problems.

Here is a function that I derived from code developed by 7bit. It uses the History folder file - symbols.raw - to get the exact list for individual platforms.

Use it if you want or just disregard.

int GetAvailPairs( string& AvalPairs[] ) // Derived from code from 7bit
{
int i, SymRawHandle, PairCnt;
string AddChar = StringSubstr(Symbol(),6,4 );
SymRawHandle = FileOpenHistory("symbols.raw", FILE_BIN | FILE_READ);
PairCnt = FileSize(SymRawHandle) / 1936;
ArrayResize(AvalPairs, PairCnt);
for (i = 0; i < PairCnt; i++)
{
AvalPairs[i] = StringTrimRight(StringTrimLeft(FileReadString(SymRawHandle, 12)));
AvalPairs[i] = StringConcatenate(AvalPairs[i], AddChar );
FileSeek(SymRawHandle, 1924, SEEK_CUR);
}
FileClose(SymRawHandle);
}

It is not ready yet, but I am developing a multi-pair capable tester that I can integrate with any EA to manage multipair strategies such as this. No limit on pairs. I have it working on single pair and debugging the multi-pair version. It gives an output much like StrategyTester with equity curve also. I am anxious to integrate Going Home with it to test different settings.

Regards,

KingHigh
  • #202
  • Feb 7, 2011 12:23am Feb 7, 2011 12:23am
kinghigh for mt4??
  • #203
  • Edited 4:56am Feb 7, 2011 4:21am | Edited 4:56am
Quoting KingHigh
Steve,

Different platforms have different pairs available, so having a fixed list can cause some problems.

Here is a function that I derived from code developed by 7bit. It uses the History folder file - symbols.raw - to get the exact list for individual platforms.

Use it if you want or just disregard.

int GetAvailPairs( string& AvalPairs[] ) // Derived from code from 7bit
{...
Absolutely fantastic contribution. Thanks.

I have been wondering how to automate the pair selection and will indeed incorporate this. I shall need to add an exclude pairs input so that users do not have to trade pairs they do not like.

As to your tester, it would be wonderful to have a tester not developed by a dimwit. I bet I am not the only one drooling at the prospect.

Cheers



Edit:
As soon as I added the auto-detect function, I spotted the problem with this. The function detects umpteen symbols I would not even dream of trading. It would take the user longer to enter all of these into an ExcludePairs list than it would to check that each pair in the PairsToTrade list exists.

Thanks anyway, but including this feature would not be helpful here.
  • #204
  • Feb 7, 2011 4:28am Feb 7, 2011 4:28am
Quoting KingHigh
Steve,

Different platforms have different pairs available, so having a fixed list can cause some problems.

Here is a function that I derived from code developed by 7bit. It uses the History folder file - symbols.raw - to get the exact list for individual platforms.

Use it if you want or just disregard.

int GetAvailPairs( string& AvalPairs[] ) // Derived from code from 7bit
{
int i, SymRawHandle, PairCnt;
string AddChar = StringSubstr(Symbol(),6,4 );
SymRawHandle = FileOpenHistory("symbols.raw", FILE_BIN | FILE_READ);
PairCnt = FileSize(SymRawHandle)...
Wow a multipair tester
Excellent
  • #205
  • Feb 7, 2011 1:20pm Feb 7, 2011 1:20pm
Quoting SteveHopwood

Edit:
As soon as I added the auto-detect function, I spotted the problem with this. The function detects umpteen symbols I would not even dream of trading. It would take the user longer to enter all of these into an ExcludePairs list than it would to check that each pair in the PairsToTrade list exists.

Thanks anyway, but including this feature would not be helpful here.
Steve,

Have already added a BlockPairs list to my version, but maybe it would be better to block by currencies such as MXN, PLN, DKK, CZK. Justta though.

KingHigh
  • #206
  • Feb 7, 2011 3:23pm Feb 7, 2011 3:23pm
The pairs were EURCAD(sell)/EURJPY(buy) the EA took the trade at 18:45 utc. I have it on a 5M chart. The difference never gets above 20. not all day, but it took the trade anyway.

I used the params ootb except for
UseScottHEdgeFilter = False
SendPairTrade = True

and I have it set for 80 open, 20 close.

GH took other trades today that looked appropriate. Don't know what else I can tell you about the trade.

Shazbut

T
  • #207
  • Feb 7, 2011 3:30pm Feb 7, 2011 3:30pm
Quoting t1minator
The pairs were EURCAD(sell)/EURJPY(buy) the EA took the trade at 18:45 utc. I have it on a 5M chart. The difference never gets above 20. not all day, but it took the trade anyway.

I used the params ootb except for
UseScottHEdgeFilter = False
SendPairTrade = True

and I have it set for 80 open, 20 close.

GH took other trades today that looked appropriate. Don't know what else I can tell you about the trade.

Shazbut

T
My limited understanding is that Stoch is subject to spikes and that these spikes do not always show up on a chart - somebody discussed this briefly earlier in the thread (or was it DL's thread?). If the diff briefly went above 80 as a result of such a spike, then GV should take the trade.

  • #208
  • Feb 7, 2011 3:33pm Feb 7, 2011 3:33pm
Quoting SteveHopwood
My limited understanding is that Stoch is subject to spikes and that these spikes do not always show up on a chart - somebody discussed this briefly earlier in the thread (or was it DL's thread?). If the diff briefly went above 80 as a result of such a spike, then GV should take the trade.

Yes, I was monitoring my table of correlations this morning and this is exactly what happened. It is a double edge sword, because sometimes you get a successful trade that would not have been picked up.
  • #209
  • Edited 4:05pm Feb 7, 2011 3:49pm | Edited 4:05pm
Quoting smjones
Yes, I was monitoring my table of correlations this morning and this is exactly what happened. It is a double edge sword, because sometimes you get a successful trade that would not have been picked up.
Would it be helpful to put in some hysteresis, not averaging that would just slow down the responsiveness, just so x number are above the threshold before the trade goes?

T
Something like this


if(MathAbs(stoch1-stoch2) > OpenTradeStoch && lastStochDiff > OpenTradeStoch)
{
................
}
LastStochDiff = MathAbs(stoch1-stoch2);

perhaps would filter out the instantaneousness blips.
  • #210
  • Feb 7, 2011 4:05pm Feb 7, 2011 4:05pm
Quoting t1minator
Would it be helpful to put in some hysteresis, not averaging that would just slow down the responsiveness, just so x number are above the threshold before the trade goes?

T
If steve put a 1 in the shift of the stochastic, it would only take a trade on a closed and registed difference. He could put in a input like

extern bool StochOnClose = true;

and then in the:

iStochastic(currencyname,timeframe, StochOnClose);

because a true = int 1 and a false = int 0

That would give the user a choice.
  • #211
  • Feb 7, 2011 4:29pm Feb 7, 2011 4:29pm
Quoting smjones
If steve put a 1 in the shift of the stochastic, it would only take a trade on a closed and registed difference. He could put in a input like

extern bool StochOnClose = true;

and then in the:

iStochastic(currencyname,timeframe, StochOnClose);

because a true = int 1 and a false = int 0

That would give the user a choice.
Am I correct in thinking that this means there is only any point in testing for a trade trigger at the start of a new candle?

  • #212
  • Feb 7, 2011 4:38pm Feb 7, 2011 4:38pm
Quoting SteveHopwood
Am I correct in thinking that this means there is only any point in testing for a trade trigger at the start of a new candle?

I am not necessarily in agreement with getting rid of the way you have it setup now. If you were to make the shift a 1 as in the above example, then the EA would only make a trade if the closed difference were above the required amount listed in the inputs. I suppose the reason I suggested this is because it would give people a choice either way.

T1minator had a slightly different approach, by requiring a certain amount to spike above, instead of what I suggested. I really do not know what would be the best idea.
  • #213
  • Feb 7, 2011 4:47pm Feb 7, 2011 4:47pm
Quoting smjones
I am not necessarily in agreement with getting rid of the way you have it setup now. If you were to make the shift a 1 as in the above example, then the EA would only make a trade if the closed difference were above the required amount listed in the inputs. I suppose the reason I suggested this is because it would give people a choice either way.

T1minator had a slightly different approach, by requiring a certain amount to spike above, instead of what I suggested. I really do not know what would be the best idea.
Thanks Scott. Aside from GV, this adds another useful thingy to my coding armoury for the future.

  • #214
  • Feb 7, 2011 4:55pm Feb 7, 2011 4:55pm
Quoting t1minator
Would it be helpful to put in some hysteresis, not averaging that would just slow down the responsiveness, just so x number are above the threshold before the trade goes?

T
Something like this


if(MathAbs(stoch1-stoch2) > OpenTradeStoch && lastStochDiff > OpenTradeStoch)
{
................
}
LastStochDiff = MathAbs(stoch1-stoch2);

perhaps would filter out the instantaneousness blips.
Thank you. I have come to value your contributions highly within a very short time.

I have copied this into the comments section at the start of the code so I do not lose it. I am not going to do anything with it for now; I think we can wait to see if these spikes prove to be a problem rather than an opportunity or a neutral.

Cheers.

  • #215
  • Feb 7, 2011 8:45pm Feb 7, 2011 8:45pm
Quoting SteveHopwood
Thank you. I have come to value your contributions highly within a very short time.

I have copied this into the comments section at the start of the code so I do not lose it. I am not going to do anything with it for now; I think we can wait to see if these spikes prove to be a problem rather than an opportunity or a neutral.

Cheers.

Thank you, I will contribute where I can. I am testing that change now, I let you know how it goes.
  • #216
  • Feb 8, 2011 5:22am Feb 8, 2011 5:22am
Hi Steve

I think we have a few problems wiht the exit strategies of
pips and pound

I will try to capture on another test system
The stoch exit seems ok , (although this morning one dropped below 20 stoch and did not close )
I think we all knew that it would not be easy with multiple pairs (and arrays)

Anyway no rush as I said I will try to capture

5m , 95 and 20 is doing quite well
Seems to filter out the losers a bit
I have one testing at 90 and one testing at 95 and also I have been logging to file using the file write command the amount of time we get of each , I changed SMJ correlation table alert routine

Id rather have like 3 good trades a day than 6 good and 5 bad

Like I say i reckon we could grab the bad one if we had the exit strategies working or even some sort of trailing stop
has anyone been testing , similar results ??
  • #217
  • Feb 8, 2011 6:17am Feb 8, 2011 6:17am
Hi steve,

I am reporting the performance of going home EA :

Starting date : 1 st feb TF : 5 MIN
Pips earned : 549 untill today.

My pairs to trade : According to tradeble pairs indi i choose every day biginning. Entry and exit : 90 & 10.

Observations :

1) Profit by pips do not seem to work. Exit by stoc works.
2 ) I still get error message in most of platforms except one platform.
Intraday only.
  • #218
  • Feb 8, 2011 6:35am Feb 8, 2011 6:35am
Quoting navk
Hi steve,

I am reporting the performance of going home EA :

Starting date : 1 st feb TF : 5 MIN
Pips earned : 549 untill today.

My pairs to trade : According to tradeble pairs indi i choose every day biginning. Entry and exit : 90 & 10.

Observations :

1) Profit by pips do not seem to work. Exit by stoc works.
2 ) I still get error message in most of platforms except one platform.
Excellent pips , about 100 a day yes, I hope it stays at that average
  • #219
  • Feb 8, 2011 8:39am Feb 8, 2011 8:39am
Quoting navk
1) Profit by pips do not seem to work. Exit by stoc works.
Thanks for that. As soon as I looked at the code I spotted some mis-codings.

I have not put up a fix yet as I am looking at a couple of other things first.

  • #220
  • Feb 8, 2011 8:51am Feb 8, 2011 8:51am
The fix for the pips profit thingy is in post 1.

I needed to check the tp/sl calculation because I had an invalid stop this morning. It isn't the code; I had forgotten to update GV with CriminalIsECN enabled by default. Check this setting if you are getting these.

0 traders viewing now