THV MTF EA beta testing | Page 165 | 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

  • #3,281
  • Sep 23, 2010 6:09pm Sep 23, 2010 6:09pm
Good point Bert, since the capital growth is a function of the number of accurate entries, this lots size should be scaled to individual account value.

The major effort would be to confirm the entries as being valid, after that we can grow equity at our personnel setting. Lets face it, we will be checking on the EA every day that it is running...lol

With regards to tweaking for pairs, I believe the direction is in the trix values

Currently Fast Trix = 20, Slow Trix = 30, higher values can be used for the more volatile pairs (they act as smoothing) with a higher hot value
  • #3,282
  • Sep 23, 2010 6:49pm Sep 23, 2010 6:49pm
So what you're saying, Tread, is that we should keep the SL and TP levels as what is in Letsarb's set but that we should tweak the Trix settings, correct?

Thanks.

Brett
  • #3,283
  • Sep 23, 2010 7:46pm Sep 23, 2010 7:46pm
What I am saying is that the lot size should be discretionary to each trader, so for me I would use 1.5% of account as the stop loss. Since everyone will have different account and different risk appetites, the TP and SL should be to each trader own personnel choice.

The entry is 1lot ($10.00 per pip) stoploss is 54pips, so his risk is approximately 540.00 (depending on pair) which is 5.4% per trade. Of course this % will vary as the account increases or decreases

Trailing Stop I dislike in a scalping Ea..... a trailing stop is usually set to about 10pips(at a minimum) to have a acceptable chance of working, on a low volatile pair the spread is 2 to 5 pips....it might be wiser to exit the first trade at the full pips earned (not giving up the 10pips to the trailing stop) and reenter if the conditions are right.

Using the ATR to calculate the TP is a good idea. But on some pairs the volatility is very earratic. and the TP sometimes miss. I am a big fan of using the ATR, but I no longer am confident of it sustained profitability in the EA as a TP or SL method

The Trix of Fast to Slow has a ratio of 20/35 = 0.57

So i would use this multiplier in calculating the new trix Fast/Slow values. I while back I tried to automatically set the trix value based on the 4h ATR value, but got side tracked on another project.
  • #3,284
  • Sep 24, 2010 1:07am Sep 24, 2010 1:07am
Quoting Tread
Was not aware of that, all MTF entries are backtestable. I think multicurrency EA (strength meter /basket) type EA's are not backtestable, since it requires multiple feeds,

Once you have an uptodate History you should be fine.

The problem with MTF backtesting is that visually the indicator do not appear to match up with the actions taken by the EA. Since the EA calculation is based on tick, the EA will as an example OPEN a trade based on the indicator being Green at that point in time in that candle. However lets say the condition...
Thats great news. Thanks Tread, I'll begin backtesting and comparing with my recent forward testing to see if there are any differences.
  • #3,285
  • Sep 24, 2010 2:33am Sep 24, 2010 2:33am
Just to confirm for Ivoes and others one that all the others parameters are defaults parameters. I would like to add that I did not really refine my setting, so I am sure that it is possible to adapt SL, TP, TrStop ........
This is what I try to do in this moment.
And it would be necessary to do it for the others pairs for sure.
Otherwise, just have a look to this report, it is not so bad but does not mean anything because there was a very strong trend.............
To finish, to use 99%, 90% or less, is not so important for these setting, at least for this step of the work...so do not hesitate to perform backtest on others pair for example, it would be great.
Attached File(s)
File Type: pdf THV EA v1.131 No session.pdf   292 KB | 1,059 downloads
  • #3,286
  • Sep 24, 2010 3:02am Sep 24, 2010 3:02am
Tread you have some idea why it is my set stopped working?
  • #3,287
  • Sep 24, 2010 3:13am Sep 24, 2010 3:13am
Quoting letsarb
Just to confirm for Ivoes and others one that all the others parameters are defaults parameters. I would like to add that I did not really refine my setting, so I am sure that it is possible to adapt SL, TP, TrStop ........
This is what I try to do in this moment.
And it would be necessary to do it for the others pairs for sure.
Otherwise, just have a look to this report, it is not so bad but does not mean anything because there was a very strong trend.............
To finish, to use 99%, 90% or less, is not so important for these setting, at least...

but now put the same set on forward test and see the difference
  • #3,288
  • Sep 24, 2010 4:29am Sep 24, 2010 4:29am
Hi Letsarb. I have tested the exact same period as you on m1, but i`m only getting 30 trades comparing to your 1618. Do you have an idea why is it so. And please post set files, because it takes too much time to set the ea at your settings by reading them.
  • #3,289
  • Sep 24, 2010 5:59am Sep 24, 2010 5:59am
did. Advisor began to work, but very badly. very late entries
Attached File(s)
File Type: zip 5m.zip   1 KB | 485 downloads
File Type: zip 1m.zip   1 KB | 429 downloads
  • #3,290
  • Sep 24, 2010 6:34am Sep 24, 2010 6:34am
examples of very late entry
Attached Image(s) (click to enlarge)
Click to Enlarge Name: ea.jpg Size: 163 KB
Click to Enlarge Name: ea1.jpg Size: 167 KB
Click to Enlarge Name: EA2.jpg Size: 156 KB
  • #3,291
  • Sep 24, 2010 6:57am Sep 24, 2010 6:57am
Tread on the basic rules of Cobra is necessary to consider the color of the coral too. But this adviser does not account for. Why?
  • #3,292
  • Edited 8:47am Sep 24, 2010 7:09am | Edited 8:47am
hey letsarb,

how do you get a 99% model quality? i've found birt's tick-data solution. i downloaded eurusd from begin 2010 till now from dukascopy and merge the data into a csv file. i convert the data to fxt and hst for all timeframes and using the 226 loader script. but sometimes if i watch the visual backtest, the candles began to flicker and jumping. the ea opens position in zones they were never touched by the candles? backtesting two or three weeks takes a night long and it's not the badest pc i have. how long does your backtest need for a quarter? can you give me some hints please?

thanks and best regards, derox
  • #3,293
  • Sep 24, 2010 7:28am Sep 24, 2010 7:28am
Quoting ozma
Tread on the basic rules of Cobra is necessary to consider the color of the coral too. But this adviser does not account for. Why?
The rules make no mention of Coral Color

=================================================

The main V4 rules:


Entry: Above Coral only long, below Coral only short
Price crosses cloud - both Trix cross each other and have the same color.
Best and most secure trend confirmation is a cross of the fast Trix on the zero level.

 Exit:Faster Trix changes color and/or touch of a known good resistance/support level.
  • #3,294
  • Sep 24, 2010 8:13am Sep 24, 2010 8:13am
so I was wrong. sorry. Tread you seen my set? I think it should work, but the problem-set does not think so
  • #3,295
  • Sep 24, 2010 8:21am Sep 24, 2010 8:21am
Quoting ozma
Tread you have some idea why it is my set stopped working?
Market condition are changing, There will be time when the EA performs very well and other times when it bombs.

If the EA performed well last week, then backtest the EA again during that period, to see if the results are still consistent.

Also i see you are trading the NZD ... i would advise that you stay away from this currency for the purposes of scalping
  • #3,296
  • Sep 24, 2010 8:22am Sep 24, 2010 8:22am
Quoting ozma
so I was wrong. sorry. Tread you seen my set? I think it should work, but the problem-set does not think so

Running the test on the set, will let you know
  • #3,297
  • Sep 25, 2010 11:16am Sep 25, 2010 11:16am
Has anyone had any luck on Letsarb's set yet? I ran a forward test on friday on EURUSD M1 but nothing positive really.

@Tread,

Tread, my friend, are you still using the same set file you posted with the last version of the EA in the preset folder to give you entries?

Have a good weekend all.
  • #3,298
  • Edited 1:55pm Sep 25, 2010 1:12pm | Edited 1:55pm
hi tread,

what happenend to the divergence pattern?

i think we need a set file generator, that generates all different variations. and someone who test this million of combinations what this ea is able to do.
  • #3,299
  • Sep 25, 2010 2:32pm Sep 25, 2010 2:32pm
The Divergence when detected is 2 bars after it had happened. thus it was getting you late into the entry or keeping you out of the entry alltogether. It was removed along with the gartley pattern
  • #3,300
  • Sep 25, 2010 2:33pm Sep 25, 2010 2:33pm
Quoting pips r us
Has anyone had any luck on Letsarb's set yet? I ran a forward test on friday on EURUSD M1 but nothing positive really.

@Tread,

Tread, my friend, are you still using the same set file you posted with the last version of the EA in the preset folder to give you entries?

Have a good weekend all.
Yes, the only change I made was I turned on the Damani Volaltmeter and the ADX_3
0 traders viewing now