THV MTF EA beta testing | Page 78 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

THV system, final edition 26,001 replies

MTF Stochastics into MTF Stoch Histogram 8 replies

THV ind 8 replies

My Journal - Based on THV V3.0 by Cobraforex 30 replies

THV RSI_V2 with Alarms --- need help! 6 replies

Attachments: THV MTF EA beta testing
Exit Attachments

THV MTF EA beta testing

  • #1,541
  • Jul 30, 2009 2:51pm Jul 30, 2009 2:51pm
Quoting kaizasoze
How about making your own?
OK

I can do that.

Good Luck Everyone

Gene
  • #1,542
  • Jul 30, 2009 6:17pm Jul 30, 2009 6:17pm
@Kaz

Hi Kaz - this is from post 1517 - your answer to me:

"2. What are your settings for trix and cloud histos..they should match with what you have on EA. Seems like you are just using the default template provided by Xux with version 8."

So what do you mean ? Have I to switch the indi parameter "autotimeframe" to "true" ? Or what else ?
  • #1,543
  • Jul 30, 2009 7:16pm Jul 30, 2009 7:16pm
Quoting kaizasoze
Can I suggest a few things:

1. FastTrixDirection = 310.
2. OppositeSignalExit = True.
3. CloseifProfit = False.
4. Enable Martingale, Lot size =1.5
5. OB/OS Levels of Higher TF ie 30Min
6. Use H4 instead of H1.
7. Use only majors/ with spread under 4.

Hope it should help us reduce some draw down.

Cheers

Kaz
Since this is mostly a trend following concept, I recommend to closely examine the entry times of the losing trades. They may happen during Asia in which case the EA needs to be traded during London and US sessions only. This simple change may avoid some of the losing trades and bring the profitability up. Just a suggestion anyway.
If it backtests good run it live!
  • #1,544
  • Jul 30, 2009 8:23pm Jul 30, 2009 8:23pm
No - analyze the 15 M Kaz setfile: trading time is 0700 - 2000 - asian session is excluded !
  • #1,545
  • Jul 30, 2009 8:40pm Jul 30, 2009 8:40pm
Quoting Mandl2007
@Kaz

Hi Kaz - this is from post 1517 - your answer to me:

"2. What are your settings for trix and cloud histos..they should match with what you have on EA. Seems like you are just using the default template provided by Xux with version 8."

So what do you mean ? Have I to switch the indi parameter "autotimeframe" to "true" ? Or what else ?
Bro..I seriously recommend going through the EA manual. What I mean is whatever MTF and Cloud DOT values you are using in EA don't seem to match with the chart. DOT's in EA should be equal to DOT's in chart.

Kaz
  • #1,546
  • Jul 30, 2009 8:41pm Jul 30, 2009 8:41pm
Quoting Mandl2007
No - analyze the 15 M Kaz setfile: trading time is 0700 - 2000 - asian session is excluded !
You can trade during Asian session on YEN and AUD pairs.
  • #1,547
  • Jul 30, 2009 8:43pm Jul 30, 2009 8:43pm
Quoting fxspeedster
Since this is mostly a trend following concept, I recommend to closely examine the entry times of the losing trades. They may happen during Asia in which case the EA needs to be traded during London and US sessions only. This simple change may avoid some of the losing trades and bring the profitability up. Just a suggestion anyway.
Why trade against trend when you can trade with it!
Also CT trades can be picked up on lower TF like 5min and 1min up to you how you want to play.
  • #1,548
  • Jul 31, 2009 1:35am Jul 31, 2009 1:35am
Hi All,

the Trix Histro bar has 4 lines of dots, with dot 4 being the first line shown followed by dot 3 etc. (The Cloud Histro bar is the opposite). In screenshots posted people show the shortest time frame on top, with longer time frames below - this is easier to take in at a glance and is therefore quite logical - BUT - are you taking this into account in the set files? For example, MtfTrixDot1 should apply to the longest timeframe - not the shortest. In the case where all mtftrixdot settings are the same then this won't matter, but in my case I'm using closed candle settings for open/close on the shortest timeframe, and open candle settings for the higher time frames - so it does make a difference.

If I'm simply pointing out the obvious then sorry to waste your time.

Tim
  • #1,549
  • Jul 31, 2009 2:31am Jul 31, 2009 2:31am
You guys gotta love this one..more patterns at work! This time on 15 min..sweet!! Didn't manage the trade very well..but trade management is secondary at this point for me. Just trying to optimize entries and exits.
Attached Image(s) (click to enlarge)
Click to Enlarge Name: M15GBPUSD29610307Open.gif Size: 36 KB
Click to Enlarge Name: M15GBPUSD29610307Close.gif Size: 36 KB
Click to Enlarge Name: M15GBPUSD29612937Close.gif Size: 35 KB
  • #1,550
  • Jul 31, 2009 8:03am Jul 31, 2009 8:03am
Something Interesting!
Didn't hit the SL, actually I had both set up..manual=15 and 2*ATR(20) which incidentally was also 15 and Emergency one was 30(Visible SL). Now is it EA which was not able to close or its the brokers who shut down all requests during news time? Beyond my thinking..Trade was picked up before news and Emergency SL was hit during news where EA had disabled trading.

Also hidden SL EA shows was 89.364..don't know how it calculated that?

Any thoughts are welcome.

Cheers
Kaz
Attached Image(s) (click to enlarge)
Click to Enlarge Name: M5USDJPY29623535Open.gif Size: 32 KB
Click to Enlarge Name: M5USDJPY29623535Close.gif Size: 31 KB
  • #1,551
  • Jul 31, 2009 8:57am Jul 31, 2009 8:57am
no explanation from me - but yesterday I watched 2-3 times the same ! Didn't made a screenshot.
  • #1,552
  • Jul 31, 2009 9:00am Jul 31, 2009 9:00am
Quoting NosferatuMan
Here is a backtest with EA 0.8 and the set from kaizasoze (closeprofit=false) for eurusd M15 from may till today.
Here is the backtest for the same period with a set file with some of the changes Kaizasoze suggested.

2 more trades (171 total) and 85% positive trades against 77% of the last test.

Here are the two graphics.

EDIT: percentages were not correct
Attached Image(s) (click to enlarge)
Click to Enlarge Name: TesterGraph.gif Size: 12 KB
Click to Enlarge Name: StrategyTester2.gif Size: 10 KB
  • #1,553
  • Jul 31, 2009 9:04am Jul 31, 2009 9:04am
Quoting NosferatuMan
Here is the backtest for the same period with a set file with some of the changes Kaizasoze suggested.

2 more trades (171 total) and 90% positive trades against 80% of the last test.

Here are the two graphics.
What were the changes you implemented?
  • #1,554
  • Jul 31, 2009 9:09am Jul 31, 2009 9:09am
Quoting NosferatuMan
Here is the backtest for the same period with a set file with some of the changes Kaizasoze suggested.

2 more trades (171 total) and 85% positive trades against 77% of the last test.

Here are the two graphics.

EDIT: percentages were not correct
Hi,
Very nice job! Can you post the report itself so that we can look at the trade by trade and try to figure out how we can avoid some of the losers? Also what is the MaxDD on the backtest?

Regards
If it backtests good run it live!
  • #1,555
  • Jul 31, 2009 9:23am Jul 31, 2009 9:23am
Quoting kaizasoze
What were the changes you implemented?
Take a look at the set file. I'm really not a specialist, so feel free to correct me and suggest some more changes
Attached File(s)
File Type: doc kaizasoze_changes.set.doc   12 KB | 341 downloads
  • #1,556
  • Jul 31, 2009 9:32am Jul 31, 2009 9:32am
Quoting fxspeedster
Hi,
Very nice job! Can you post the report itself so that we can look at the trade by trade and try to figure out how we can avoid some of the losers? Also what is the MaxDD on the backtest?

Regards
Here are the reports. Please comment.
Attached File(s)
File Type: doc reports.rar.doc   18 KB | 265 downloads
  • #1,557
  • Edited 11:05am Jul 31, 2009 10:52am | Edited 11:05am
Quoting NosferatuMan
Here are the reports. Please comment.
Try this:
1. Use all closed candles for entry ie. use 110 instead of 100. and FreshSignal= False. Use FastSlowTrixCross = 310.

2. All patterns except divergence turned to 3. Bars back =7.

3. Why monday open is false? U want to trade into gaps? MondayOpen=true;and put start hours at london open.

4. PSARTrailingStop=false. Use Classic TS.

5. CloseIfProfit=true? Use False.

6. DailyClose=false? Use true CloseHour 2000

7. FridayClose=false? Use true Close at London Close.

8. This is funny..Martingale=true; LotMultiplier=1.5; MaxProgression=0; Do you thing EA will use martingale? use MaxProgression = 2.

9. You are still trading in direction of hourly trend..use this for trix histos--
M15-Fast.
M30-Fast.
M60-Fast.
M240-Fast. then try M1440-Fast.

use this for cloud histos--
shortperiod =30
longperiod =60

10. Also in TP use first exit at 25(manual) and second at ATR based use period 20 and multiplier 2. ADX multiplier 3.5 ADX period 20.

11. Use 2% risk per trade. ie leverage should be 2% not 4%..its too high.

12. Use fixed SL =25.

I have a question..what are mismatched chart errors?

Hope it should push the curve further. It would be great if guys back testing this set could post the results on Sat. I'll spend some more time on evaluating EA.

Cheers
Kaz
  • #1,558
  • Jul 31, 2009 11:03am Jul 31, 2009 11:03am
Quoting kaizasoze

I have a question..what are mismatched chart errors?


Cheers
Kaz
These are my first weeks with EA so I don't know much about it. Maybe someone else can answer yuor questions.

I will try to change the set file according your post (maybe I will need your help). I will post the results later.

Thanks once again.
  • #1,559
  • Jul 31, 2009 12:05pm Jul 31, 2009 12:05pm
Did you use ADXMultiplier? It may explain this wide stop loss.

Quoting kaizasoze
Something Interesting!
Didn't hit the SL, actually I had both set up..manual=15 and 2*ATR(20) which incidentally was also 15 and Emergency one was 30(Visible SL). Now is it EA which was not able to close or its the brokers who shut down all requests during news time? Beyond my thinking..Trade was picked up before news and Emergency SL was hit during news where EA had disabled trading.

Also hidden SL EA shows was 89.364..don't know how it calculated that?

Any thoughts are welcome.

Cheers
Kaz
  • #1,560
  • Jul 31, 2009 12:22pm Jul 31, 2009 12:22pm
Quoting xux99
Did you use ADXMultiplier? It may explain this wide stop loss.
I got it XuX. ADX multiplier was 3.5. Didn't realize it at all. Thanks!
One more question..Can we backtest patterns as well? Also, can we use OB/OS levels other then those on current timeframe during backtest?

Forward testing is taking a lot of time and there are not many people who are doing it. I am considering using back-testing from now on as well on my laptop. My only concern is lack of news-time response. Oh well, lets c how it goes..

Have a great weekend all

Cheers
Kaz
0 traders viewing now