Caillou: another always-in auto-trader by SteveHopwood | Page 13 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Another Day, Another Chair, Another Day in the Office 203 replies

SSS EA v1.0 - Auto Trader Bot 623 replies

Candle Stick Pattern Auto Trader 12 replies

Auto trader for Currenex? 2 replies

SteveHopwood EAs 4 replies

Attachments: Caillou: another always-in auto-trader by SteveHopwood
Exit Attachments

Caillou: another always-in auto-trader by SteveHopwood

  • #241
  • Edited 8:39am May 22, 2011 7:16am | Edited 8:39am
Quoting magft
Ok, i've had a go at implementing Edd's idea and current backtesting seems to have reduced DD from 61% to 12%.

I have also tried to correct the HiLo Stoploss for Greg but it still gives the ordersend error 130 so if someone with more experience in this could have a look that would be great.

So Updated dated version based on gspe's CT caillou is v1.08 and Edd's version is v1.08 Edd. I've added the option to allow the user to select only buy, sell or both. Also a CounterTrend option to compare original caillou against gspe caillou methods is added...
Mike,

Thank you. You are a star, you deliver as you promised.


Sincerely,

Edd
  • #242
  • May 22, 2011 12:03pm May 22, 2011 12:03pm
Quoting Hardpike
Sorry, none of them backtest for me....:-(
i've spotted a slight typo that might be causing issues with the trades, i've just tested both and they do trade both ways.

The Edd version seems to work very well on the H1 EU chart, DD down to 9.6% and 260 on 1000 account over 4 month period and 76% profitable trades. This using current default settings.

Mike
Attached File(s)
File Type: mq4 Caillou v1.08.mq4   21 KB | 282 downloads
File Type: mq4 Caillou v1.08 Edd.mq4   22 KB | 291 downloads
  • #243
  • May 22, 2011 12:12pm May 22, 2011 12:12pm
Quoting Marcuss
some tests
Marcuss,

I have tried to copy your backtest (1.07 with the same input as yours) and to my big suprise my account blown up (contrary to your fantastic result). I have also run the same exercise on my second account - with the same result.

It looks like our systems differs starting from April 20, after the two buy orders... See atttached statements - original and two mines...

I understand there might be the difference in the quotes but not that significant...

Have you changed the 10.07 source code?

Thanks,

BTW: Thanks Steve, Caillou, Magft, Hardpike and others for very interesting thread...

Mariusz
Attached File(s)
File Type: zip MM backtest.zip   66 KB | 218 downloads
  • #244
  • May 22, 2011 12:14pm May 22, 2011 12:14pm
there is some problem with the version 1.8 edd, this is form debugging the EX4 log

1;39;;;Function "AreBuysAllInProfit" is not referenced and will be removed from exp-file
1;39;;;Function "AreSellsAllInProfit" is not referenced and will be removed from exp-file
  • #245
  • May 22, 2011 12:31pm May 22, 2011 12:31pm
it is possible to add maxspread funkction as it is on the beast. It will be good when there are some news, where the spread is bigger - to avoid trading
  • #246
  • May 22, 2011 2:51pm May 22, 2011 2:51pm
Quoting hankeys
there is some problem with the version 1.8 edd, this is form debugging the EX4 log

1;39;;;Function "AreBuysAllInProfit" is not referenced and will be removed from exp-file
1;39;;;Function "AreSellsAllInProfit" is not referenced and will be removed from exp-file
Hi,
this are not errors but warnings.
They means that the function defined in the source is not used so the complier will remove it from the ex4 file.

gspe
  • #247
  • May 22, 2011 3:09pm May 22, 2011 3:09pm
Quoting gspe
Hi,
this are not errors but warnings.
They means that the function defined in the source is not used so the complier will remove it from the ex4 file.

gspe
Thanks for replying gspe but i think i do actually need to be calling those functions i was messing about last night trying different things and forgot to change back.

So, please find the modified v1.08 again for the third time.

I have also added the max spread check and restructured part of the code grouping some of the filters like max spread min margin check and number of trades. This is present in v1.09 and in BT it does make trades and the fix above in 1.08 increases profit from $260 to $305 and DD to 13%.

So feel free to try different combinations of settings and share your findings and thoughts on the way forward. I'm sure we are back on track to finding a winner.

Mike
Attached File(s)
File Type: mq4 Caillou v1.08.mq4   21 KB | 287 downloads
File Type: mq4 Caillou v1.08 Edd.mq4   22 KB | 267 downloads
File Type: mq4 Caillou v1.09 Edd.mq4   22 KB | 363 downloads
File Type: mq4 Caillou v1.09.mq4   21 KB | 382 downloads
  • #248
  • May 22, 2011 3:44pm May 22, 2011 3:44pm
Thanks magft!
  • #249
  • May 22, 2011 9:46pm May 22, 2011 9:46pm
Quoting gregart
Eu and GU M1
caillou v1.01 (inverted)
you'll find bellow the settings

[size=2]DisplayGapSize=30
Lot=0.10
WithTrendLotMultiplier=1
ProfitTarget=5
SingleTradeProfitTarget=7
MagicNumber=110
TradeComment=
CriminalIsECN=0
tt=----Trading hours----
Trade_Hours=Set Morning & Evening Hours
Trade_Hoursi=Use 24 hour, local time clock
Trade_Hours_M=Morning Hours 0-12
start_hourm=6
end_hourm=11
Trade_Hours_E=Evening Hours 12-24
start_houre=13
end_houre=18
fsh=Friday stop trading hour
FridayStopHour=14
fta=For the Aussies
rsiin=----Rsi inputs----
UseRsi=0
RsiTf=1440
RsiPeriod=20
rsap=Applied...
Hi Gregart, great results. I will be testing these settings this week. I have discovered the "Trading Hours" inputs are actually based on my PC's clock - so in order to match your settings - can you tell me what time zone you are in ? Or better yet what hours in GMT are you using?

Thanks
  • #250
  • May 23, 2011 12:06am May 23, 2011 12:06am
Hi,

I use Activtrades as broker for this test, it's the hour of broker important not the place tou are.


Quoting DojiMojo
Hi Gregart, great results. I will be testing these settings this week. I have discovered the "Trading Hours" inputs are actually based on my PC's clock - so in order to match your settings - can you tell me what time zone you are in ? Or better yet what hours in GMT are you using?

Thanks
  • #251
  • May 23, 2011 3:15am May 23, 2011 3:15am
My demo went to $ -600, trading 0.01 lots on EU GU and EG.

So now I use Steve's MPTM with shirt protection (maximum loss for the open trades) on at $ -200.
(set everything off, except trade by magic number, manage basket trades, basket takeprofit at $3 and shirt protection)

Maybe this is an idea. Trade the EA for one week. The profit of that week is the shirt protection for the next week. Kind of locking your profits for the large drawdown that can/will happen
  • #252
  • May 23, 2011 3:55am May 23, 2011 3:55am
Whatever version or timeframe you use, they all fail.

My 1.03 on H4 now has -2500 Fp...
No one wants to be defeated~
  • #253
  • May 23, 2011 9:13am May 23, 2011 9:13am
Quoting Hardpike
by the way I readed many times that backtesting on weekends cant be performed because high spreads.
there is a easy trick for this.

shut down MT4 before get close to friday close.
if you swictch on again with no internet connection will keep same spreads that the moment you shout it down.

all weekend with low spreads

oh yeah
Hi Hardpike, that's a nice easy solution for weekend backtesting, but in case you did not shutdown prior to market closing, here is another way.
The MT4i guys wrote a script/EA to allow you to manually change the spread to whatever you want during backtesting - pretty cool!
Here is the website with info and download link:

http://www.mt4i.com/spreadcontroller.aspx

Happy Testing.
  • #254
  • May 23, 2011 9:31am May 23, 2011 9:31am
Quoting gregart
Hi,

I use Activtrades as broker for this test, it's the hour of broker important not the place tou are.
Hi Gregart, Activtrades must be using GMT time...

I have tested on several brokers and the "trade hours" settings are actually based on my PC time not the brokers time.

I want to trade only from London Open to NY Close.
Using HotForex, London open is 09:00 broker time and London open on my PC time is 0:00. I had to setup the "Trade Hours" time to start at 0:00 to make it start at London open - so the EA is really using PC Local time for the settings.

Can anyone else confirm this?

thanks
  • #255
  • May 23, 2011 9:44am May 23, 2011 9:44am
Here are my results for today since the open.
Using V1.04 24hr trading on EU, GU, EG with 0.1 lots.

Net PL so far $619.42

Attached Image (click to enlarge)
Click to Enlarge Name: 5-23-2011 7-33-29 AM.png Size: 25 KB
  • #256
  • May 23, 2011 12:04pm May 23, 2011 12:04pm
Already tried the bot 3 times (2 times the original one and 1 time v1.04). While it sure looks great in the beginning, it always fails at some point. I stopped all of the forward tests after 3/4 days, because the current floatings compared to the profits were so negative, I never want that to happen live. Most of the times there were like a couple of hunderd pips profit for a couple of days but the floating pips were minus a couple of thousand.

It just goes against one very very important lesson i learned: never add to a losing trade.

Don't know how the latest versions behave by the way, but i just like to warn anybody trying to take this thing live already: dont! (yet)

As for the programmers, thanks a lot so far for all your hard work!
  • #257
  • May 23, 2011 12:45pm May 23, 2011 12:45pm
Quoting DojiMojo
Here are my results for today since the open.
Using V1.04 24hr trading on EU, GU, EG with 0.1 lots.

Net PL so far $619.42

Attachment
Dfault settings? What timeframe?
  • #258
  • May 23, 2011 1:00pm May 23, 2011 1:00pm
My thought about the EA after a week + of testing is that when things are good the EA works fine and we got profits but when things are getting bad then it's a dissaster.

I do believe that the EA could be used not as a robot but as a tool to take a trade, since we all following different trade systems then I am thinking what about using the EA to take a trade, we should use our system and when it comes to the moment of the trade we should let the EA to handle the trade. This is not a set and forget thing but since we are expecting some sort of movement then version 1 or 1.01 should work fine.

Some goodies from the later versions should be added to version 1 like the max trades (it would be nice if we could choose max trades for each direction for example 2 buy and 2 sell) and the money managment.

Another option will be to drop the candles and the EA will take trades based to pip count and this because when a large candle appear the EA take a trade at the end of it and the price most of the time return and sit inside the canal that is created by the two orders whilst when a new trade will be taken every 3 pips our TP will be hit sooner.

Hope it make sence.
  • #259
  • May 23, 2011 6:08pm May 23, 2011 6:08pm
hi, first of all, I know I should not take this thing live yet. However, I do not like backtesting at all (too ideal situations such as nice and quick trade closures, incomplete tickdata, etc) and even demo trading I'm sceptical about.

So, I took a 500:1 micro-account, put 100$ on it, trade 0.01 lots (meaning basically that I can take thousands of orders), put version 0.7 on about 10 pairs (M1), all TP 0.25, and no RSI and stuff, and just see what happens after a few days. And here are my findings:

- the account was increasing, went from 100 to about 112 since thursday evening to monday evening
- drawdown was fine last thursday and friday but monday was a bit more difficult, went to -50$, having almost 400 trades on one pair and 1600 in total on the 10 pairs. But by the end of the day, drawdown was reduced to about -10 and sets started to close.
- and now the problem that I think is not revealed in backtesting. It might be just a problem with my trading setup or whatever, but here it is. For some reason, once a set with lots of trades hits its TP and starts closing, it closes slooooooooooooow. And with slow I mean reaaaaaaally slow (meaning that it can take up to an hour to close 400 trades). Consequently, as price keeps moving, I saw my 12$ profit being reduced to 5$ profit, when closing this '0.25$ profit trade'.
- worst case scenario: broker doesn't like the huge number of trades and is throttling me
- best case scenario: the EA set closure logic is suboptimal (especially considering that there are 10 pairs and thus 10 EAs running) and this is causing the slowdown.

I do not think the broker is throttling because
- I do not see any such messages in the terminal
- in the time between starting to close a set and ending that, the comment is never updated, indicating to me that ClosePosition is called only once and that the entire slow operation takes place in this single function.

Anybody else have similar experiences? I will perform a test with only one pair and see if there is a difference. If so, the solution is clear, only use the EA on one pair per account :-).
  • #260
  • May 23, 2011 9:38pm May 23, 2011 9:38pm
Quoting Marcuss
Dfault settings? What timeframe?
I was using M1 TimeFrame. 0.1 Lots, 24 hours, TP=10. I don't remember what I had the Single trade PT set at. I changed platforms and lost my setting.

I am now trading live: M1, 0.01 Lots, PT=2, STPT=2, 24 hours (but only when I am watching it).

Been doing very well. I have been doing some manual trades along the way to try to boost the performance. Like over the midday slow down, I had lots of open orders with many in profit. I felt like the market was going to turn, so I closed most of the orders manually....

Good luck
0 traders viewing now