Forex Tester 3, better then MT4 back testing? | Page 2 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Is Trading Forex And Making Money Better Then Sex? 27 replies

MT4: Create objects in Tester in new build MT4 0 replies

Better Then MT4 - Any Suggestions? 0 replies

Testing a semi-automatic EA with a strategy tester 1 reply

2009.03.11 15:41:15Tester memory handler: tester stopped because not enough memory 19 replies

Attachments: Forex Tester 3, better then MT4 back testing?
Exit Attachments

Forex Tester 3, better then MT4 back testing?

  • #21
  • Dec 30, 2018 7:18am Dec 30, 2018 7:18am
Quoting broketrader
I think that people has a tendency to over-complicate things when at the beginning things are simple. The market has only 2 states, Trending or Ranging. Ranging is a flat market while when it trends, it can go up or down. So, from this finding, you have 3 possible scenarios: The market is flat. The market is in an up trend. The market is in a down trend. From there, if you can write an EA capable of performing the best it can in all those scenarios and you have the test data where all those scenarios are present, you will be well and you will also...
Nice points!

How about an EA works on ranging, designed based on

  1. BB, Envelop as tunnel
  2. Zigzag as high (resistance), low (support) point
  3. Trending signal for order direction
  4. When touch the high and low point, then open reverse order
  5. the multiple order open with dense pips(20~30pips), and maximum order numbers limited by daily range
  6. Basket or Equity % DD control

A scalping with ranging net idea. How do you guys think about this? Is it workable?

  • #22
  • Dec 30, 2018 7:20am Dec 30, 2018 7:20am
Quoting robots4me
{quote} Hi RickM -- thanks for your thoughtful response. Yes -- "over-optimization", "curve-fitting", "over-curve-fitting" -- does bring up some interesting and diverse perspectives. Surely -- there is such a thing as "over-curve-fitting". However, as far as I know, no one yet has come up with a way to quantify the level of curve-fitting -- from appropriately-fitted to over-fitted. What seems to be the norm around here is whenever a strategy doesn't perform well then someone starts waving their hands and says it's because it is overly curve-fitted....
I would defind over curve fitting as having too many external variables on a EA. For example in one of my EA's I am testing it uses four variables.
1.. SMA factor 2.. Gap between trades 3.. Size of secondary trades 4.. Take profit amount.
So this EA is only profitable if I optimize all these variables on a pair every three months to fit in with its current market type. It creates far too much work but the profits may make it worthwhile.
I then have a EA that is based on a inside breakout which works well on indexs. I would call this EA a under optimized EA because it uses only 1 variable which is
gap size of trailing stop. Its either profitable on a market or not and I have only retested it once.
Therefore myself and a few buddies call Two external variables as the ideal number to neither over or under calibration a EA.

The big arguement is does it matter if the EA works. I think not but the EA I mentioned will required constant 24/7 recalibration as it trades 46 charts in one account.
Trading thin liquidity at the boundary of the charts
  • #23
  • Dec 30, 2018 7:29am Dec 30, 2018 7:29am
Quoting cocozone9898
{quote} Nice points! How about an EA works on ranging, designed based on BB, Envelop as tunnel Zigzag as high (resistance), low (support) point Trending signal for order direction When touch the high and low point, then open reverse order the multiple order open with dense pips(20~30pips), and maximum order numbers limited by daily range Basket or Equity % DD control A scalping with ranging net idea. How do you guys think about this? Is it workable?
LOL, you've almost written the specs of the EA that I have already wrote and that I'm currently testing live
If you want to see the kind of backtest results, look here

Of course, this is a private work I've been on for years now so I cannot publish the details, but you have all the ideas in your post.
Simplicity is the Ultimate Sophistication.
  • #24
  • Dec 30, 2018 7:34am Dec 30, 2018 7:34am
Quoting broketrader
I think that people has a tendency to over-complicate things when at the beginning things are simple. The market has only 2 states, Trending or Ranging. Ranging is a flat market while when it trends, it can go up or down. So, from this finding, you have 3 possible scenarios: The market is flat. The market is in an up trend. The market is in a down trend. From there, if you can write an EA capable of performing the best it can in all those scenarios and you have the test data where all those scenarios are present, you will be well and you will also...
I design two models of EA's.
1.. One that works on trend breakouts.
2.. One that works on Reversal trade setups.

The trick is to match them together so one EA is always making nice profits while the other EA is hopefully only in moderate drawn down or has moderate losses.
I feel a balanced portfolio of around 7 would be ideal.
Trading thin liquidity at the boundary of the charts
  • #25
  • Dec 30, 2018 7:41am Dec 30, 2018 7:41am
Quoting RickM
{quote} I would defind over curve fitting as having too many external variables on a EA. For example in one of my EA's I am testing it uses four variables. 1.. SMA factor 2.. Gap between trades 3.. Size of secondary trades 4.. Take profit amount. So this EA is only profitable if I optimize all these variables on a pair every three months to fit in with its current market type. It creates far too much work but the profits may make it worthwhile. I then have a EA that is based on a inside breakout which works well on indexs. I would call this EA a...
I don't know which approach is the best but by contrast, I have a single EA that trades one single instrument (EURUSD) and with no external variable !

My approach when I witness a difficult market is to try to detect that with the most appropriate indicator and then to integrate that into the EA in order to help it to limit the damages during those periods.
Simplicity is the Ultimate Sophistication.
  • #26
  • Dec 30, 2018 7:42am Dec 30, 2018 7:42am
I found one tester is better ( Soft4FX ). It can use tick data ( not idea where is source ) but the best is not dependent of mt4 backtester. I will try for see if can run indicators as regular charts but have good look
Try don't lose pants never...
  • #27
  • Dec 30, 2018 7:44am Dec 30, 2018 7:44am
Quoting RickM
{quote} I design two models of EA's. 1.. One that works on trend breakouts. 2.. One that works on Reversal trade setups. The trick is to match them together so one EA is always making nice profits while the other EA is hopefully only in moderate drawn down or has moderate losses. I feel a balanced portfolio of around 7 would be ideal.
Good, why not merge the both in one single then ? That's what I did, of course you need an non USA account that allows edging, in the sense that you can have buy and sell orders simultaneously on the same instrument.
Simplicity is the Ultimate Sophistication.
  • #28
  • Dec 30, 2018 7:52am Dec 30, 2018 7:52am
Quoting cocozone9898
{quote} Nice points! How about an EA works on ranging, designed based on BB, Envelop as tunnel Zigzag as high (resistance), low (support) point Trending signal for order direction When touch the high and low point, then open reverse order the multiple order open with dense pips(20~30pips), and maximum order numbers limited by daily range Basket or Equity % DD control A scalping with ranging net idea. How do you guys think about this? Is it workable?
That's a Reversal design which will work if you keep the model simple. You probably need to design a good profit exit strategy first before working on the entry signals.
That means you need to have an objective why you are opening up so many trades, how are they going to profit the strategy. If you get the entry signal wrong, the extra trades can destroy a model.
Trading thin liquidity at the boundary of the charts
  • #29
  • Dec 30, 2018 7:57am Dec 30, 2018 7:57am
Quoting broketrader
{quote} Good, why not merge the both in one single then ? That's what I did, of course you need an non USA account that allows edging, in the sense that you can have buy and sell orders simultaneously on the same instrument.
I do that in some accounts already but one EA has to be dominant and the other EA's just kind of hedge during the Key EA's drawn down periods.
I still want 7 accounts in a folio to be able to capture a regular monthly return.
Trading thin liquidity at the boundary of the charts
  • #30
  • Edited 8:33am Dec 30, 2018 8:17am | Edited 8:33am
Quoting broketrader
{quote} LOL, you've almost written the specs of the EA that I have already wrote and that I'm currently testing live If you want to see the kind of backtest results, look here Of course, this is a private work I've been on for years now so I cannot publish the details, but you have all the ideas in your post.
It looks you have conquered a big mountain! So I believe my idea for the EA should be workable! But one of the keys, should be the treatment for the drawndown. How do you handle this tricky task?
  • #31
  • Dec 30, 2018 8:31am Dec 30, 2018 8:31am
Quoting RickM
{quote} That's a Reversal design which will work if you keep the model simple. You probably need to design a good profit exit strategy first before working on the entry signals. That means you need to have an objective why you are opening up so many trades, how are they going to profit the strategy. If you get the entry signal wrong, the extra trades can destroy a model.
I think TP is easy, but to exit the positions should be a tricky task. The object to opening up so many trades, I would like describe, like sawing a wood, a saw with so many teeth, sawing up and down....., so it can saw wood fast, and the chips fall from sawing, which are the profits. Sorry for my bad English, however I am trying my best!

Any good idea for the exit strategy?
  • #32
  • Edited 8:59am Dec 30, 2018 8:39am | Edited 8:59am
Quoting broketrader
{quote} LOL, you've almost written the specs of the EA that I have already wrote and that I'm currently testing live If you want to see the kind of backtest results, look here Of course, this is a private work I've been on for years now so I cannot publish the details, but you have all the ideas in your post.
Hi Brokertrader

I was looking at your trade explorer and to me its looks promising. You really need to judge it over a year before drawing any final conclusions. The equity curve is a bit up and down but I reckon I may have a clue on that one. EURUSD is neither a good pair for a Reversal model or a Trending model, its just neutral and I get little profits out of this pair on any model. If your'e trading a trending model, test it on a trending bias pair like EURAUD / GBPAUD / EURNZD / GBPUSD / GBPCAD.
I find most pairs have a Bias to show profits for either a reversal model or a trending model. I like any CHF pair for reversal models but EURUSD is hard to model on a EA.
It sucks.
Trading thin liquidity at the boundary of the charts
  • #33
  • Dec 30, 2018 8:44am Dec 30, 2018 8:44am
Quoting braintheboss
I found one tester is better ( Soft4FX ). It can use tick data ( not idea where is source ) but the best is not dependent of mt4 backtester. I will try for see if can run indicators as regular charts but have good look
Thank you for the information. Once if you have some results please share.

Cheers!
  • #34
  • Dec 30, 2018 9:06am Dec 30, 2018 9:06am
Quoting cocozone9898
{quote} I think TP is easy, but to exit the positions should be a tricky task. The object to opening up so many trades, I would like describe, like sawing a wood, a saw with so many teeth, sawing up and down....., so it can saw wood fast, and the chips fall from sawing, which are the profits. Sorry for my bad English, however I am trying my best! Any good idea for the exit strategy?
The pro's start with a spread sheet and manually plot a strategy and its exits candle by candle, they don't even code a thing tell they have defined the best entry & exit strategy's for their model. I suggest you open up microsofts Excel and start plotting your strategy but spend most of your time on how you can exit in profit.
Trading thin liquidity at the boundary of the charts
  • #35
  • Dec 30, 2018 9:30am Dec 30, 2018 9:30am
Quoting cocozone9898
{quote} It looks you have conquered a big mountain! So I believe my idea for the EA should be workable! But one of the keys, should be the treatment for the drawndown. How do you handle this tricky task?
It looks only, it will be true once I will be on top of it ;-)

So yes, it's workable but certainly not easy. And concerning the drawdown, well, you cannot avoid it or maybe if you can, you're better than me. Actually on my backtests, I have a drawdown up to 36% which many would consider as quite big or even too big but which in reality isn't if you know that on the medium term (months) you're profitable. If you know that, and are confident in it, then you go to the maximum profit you can get with an acceptable drawdown as long as the recovery period is not too long (weeks, months).

But I understand that this could also be a matter of character or account size. I want to fund my account up to 2k and make Gorilla profits from it! I don't know if I will reach that goal, at least I have a clear plan for that and risking 2% to have 10 or even 20% yearly on a 2K account is pointless to me as I don't need those profits to live.

I think the tricky task is to recognize the kind of market that brings you a drawdown and from there plan some strategy to mitigate what you can while at the same time you learn to accept the losses you can't avoid as the price to pay to do this business, that's what I'm doing right now.
Simplicity is the Ultimate Sophistication.
  • #36
  • Dec 30, 2018 9:43am Dec 30, 2018 9:43am
Quoting RickM
{quote} Hi Brokertrader I was looking at your trade explorer and to me its looks promising. You really need to judge it over a year before drawing any final conclusions. The equity curve is a bit up and down but I reckon I may have a clue on that one. EURUSD is neither a good pair for a Reversal model or a Trending model, its just neutral and I get little profits out of this pair on any model. If your'e trading a trending model, test it on a trending bias pair like EURAUD / GBPAUD / EURNZD / GBPUSD / GBPCAD. I find most pairs have a Bias to show...
I understand your point but believe me EURUSD and the others you cited are really different animals and I coded my EA originally for EURUSD and saw that trying to adapt for other pairs was painful or almost impossible without changing it's original ADN which I don't want as I think my original idea has really some good value in it.

As per the equity curve, yes it sucks right now but look at the monthly chart below,

Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot1.png Size: 85 KB


See the ATR(1) for this month, it's at a historical low, those who follow EURUSD know that this month is a flat month, some 200+ pip of range, so no trend no profit as I said before my EA is essentially a trend following machine, so even if I have losses in December (I hope no more than 36%) I'm not happy with that but OK and hoping for more volatile months in 2019.
Simplicity is the Ultimate Sophistication.
  • #37
  • Dec 30, 2018 10:32am Dec 30, 2018 10:32am
Quoting cocozone9898
{quote} Thank you for the information. Once if you have some results please share. Cheers!
I tested it and works well. Better that MT4 backtester solution. Only limitation is custom indicators can need small adjustment and multi currency tools not work out of box. But multicurrency tools works if you prepare it for backtester.

Inserted Video
Try don't lose pants never...
  • #38
  • Dec 30, 2018 6:22pm Dec 30, 2018 6:22pm
Quoting broketrader
{quote} I understand your point but believe me EURUSD and the others you cited are really different animals and I coded my EA originally for EURUSD and saw that trying to adapt for other pairs was painful or almost impossible without changing it's original ADN which I don't want as I think my original idea has really some good value in it. As per the equity curve, yes it sucks right now but look at the monthly chart below, {image} See the ATR(1) for this month, it's at a historical low, those who follow EURUSD know that this month is a flat month,...
Hi Broketrader

I do hope I don't come across as a rude but just offering some experiences I have learned along the way. I have just found EURUSD & USDJPY are very hard to code profitable EA's on. I am sure a few other traders may disagreed with me but that just my findings. I do find my most profitable EA's are reversal models and not trending models because its my belief the market is only in a trending market type for around 20% of the time. I would say Reversal models I build produce 70% of my EA profits so I form accounts that are dominated by one Reversal model and hedged by two Trending models.
Trading thin liquidity at the boundary of the charts
  • #39
  • Dec 31, 2018 12:34am Dec 31, 2018 12:34am
Hallo friends,
after a long lurking period, I am back again because discussion is getting interesting here. I have a question here, along the lines of what Rick was suggesting a few posts ago, which is innovative, in a sense, because it shifts the discussion about backtesting in a different perspective: spreadsheet planning. I mainly use R for such purposes, but my results are very mixed.
So, Rick, if you want/like to share. What are you studying with spreadsheets? Something on the line or Stridsman’s “trajectories”? MAE and MFE? What do you mean by “plotting a strategy” on a spreadsheet?

Thank you if you’re able to share and keep up with the nice work!
πάντα δοκιμάζετε τὸ καλὸν κατέχετε
  • #40
  • Dec 31, 2018 12:45am Dec 31, 2018 12:45am
Quoting braintheboss
{quote} I tested it and works well. Better that MT4 backtester solution. Only limitation is custom indicators can need small adjustment and multi currency tools not work out of box. But multicurrency tools works if you prepare it for backtester. https://youtu.be/XgQJ3VPKf9M
Hi braintheboss,

Thank you for the sharing. You mentioned limitation for custom indicators, how about EAs? Could be also testing?

Cheers, and Happy new year!
0 traders viewing now