Nb 1.0 trend trade system | Page 21 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

Are you a trend trader or a counter-trend trader? 29 replies

What does it mean to trade with the trend - Define trend? 47 replies

Trading System - Trade with the Trend 5 replies

Multis-trend trade EA for scalp on short trend 27 replies

Would you trade this system? New system #2. 10 replies

Attachments: Nb 1.0 trend trade system
Exit Attachments

Nb 1.0 trend trade system

  • #401
  • Jan 7, 2011 6:18am Jan 7, 2011 6:18am
Dang havent seen this. Cad dropped like a rock and is low USDJPY is high and usd chf is in the middle. I have no read on tonights news gonna protect my 36 pip cad profit and see if it runs tonight.

I just saw an interesting scenario. The gbpchf and gbpjpy if you buy the gbpchf and price on the gbp goes up tonight you are in. if it goes down you are not. If you take a sell on the gbp/jpy and gbp goes down you are in but if it goes up you are not. Maybe a win win no matter what. I have never tried this before ................. hmmmmmmmmmmmmm.
Attached Image(s) (click to enlarge)
Click to Enlarge Name: gbp chf.gif Size: 22 KB
Click to Enlarge Name: gbp jpy.gif Size: 23 KB
  • #402
  • Jan 7, 2011 6:51am Jan 7, 2011 6:51am
Quoting SteveHopwood
I am going to have to program this into the auto-trader so I have to devise the code anyhow.

Rather than try to do a piecemeal development of several different ea's for what are really the same functions, here is what I propose:
[list=1][*]add an input to the auto-trader that will send it onto 'waiting' mode as if it were one of the buy/sell ea's we are using here. Users will use it in their manual trading and drag it onto the chart when there is a setup[*]add the visual stop loss that the criminals can see - a user setting of, say, 100 pips[*]add...
This is great.
I will wait.
Thanks.
A little beginner trader since 2006
  • #403
  • Edited 8:59am Jan 7, 2011 8:29am | Edited 8:59am
Took 3 trades:

gbp/chf - entered long manually a few pips late as I had just sat down to start the day ... 15 min prior to NFP release ... stopped out on a spike w/ 16p slippage,lol, ... luckily, the trade volume setting was 5X less that what my buy EA is set at, so cost wise -10pip [-50 p actual]... (note to self: best to stay on sidelines pre-report, something I already know well -> discipline)

nzd/chf - entered long @ 0.73191 w/ EA, manual TP @ 0.73511 on approach to upper TMA +32
Edit: Correction - apparently this trade was also manual entry. I sure thought when typing this up that two EA's had entered, but the charts say differently
Edit #2: hmm - can't quite figure this out. IBFX Expert log shows nzd/chf NB EA has sent a trade and suspended itself, and the trade volume confirms, but I don't think I manually removed the EA ... whatever

aud/chf - entered long @0.95899 w/ EA, manual TP @ 0.96342 after stall on spike up +44

I soon found out that you become busier than a one-legged waitress at IHOP restaurant w/ multiple entries firing off ... setting stops, taking pics, etc. At this stage, I'm pleased w/ the TP's, although feel much more is in the offering once I can settle in w/ all this.
Attached Image(s) (click to enlarge)
Click to Enlarge Name: gbp_chf.png Size: 28 KB
Click to Enlarge Name: nzd_chf.png Size: 28 KB
Click to Enlarge Name: aud_chf.png Size: 28 KB
  • #404
  • Jan 7, 2011 9:30am Jan 7, 2011 9:30am
My two major picks went well the usd/cad and the aud chf. One is at 60 and the other 40. I will take that 100 pips. I did decided to take the usd jpy but had a real late entry so not much there even though if it had entered on the line it also would have been a nice trade. The aud/nzd entered went back down and then came back to almost BE so I closed it out for the night. Dont feel like holding anything over the weekend. I tried the gbp/chf and gbp/jpy thing the gbp/chf went almost 100 and the gbp/jpy did a peeker at -55 so I closed them both for +35 total. So 135 pip night going to bed now.
  • #405
  • Jan 7, 2011 11:35am Jan 7, 2011 11:35am
Quoting nanningbob
Actually I see a point where we blend the contertrade with the trend. As the trend begins changing we can counter trade. I think during divergance we dont counter trade. That may do wonders for the countertrade thread. When we see convergance we begin to countertrade all the way to the trend divergance and then to the next convergance change to countertrading then. It maybe the one key that stops the big DD. I have to look at that on a 4H chart and see how to set it up but if not trading during divergance helps that system, you know we make a ton...
I see the same things and will not post my ideas on how to combine the 2 here to maintain the purity of this thread. If I hit on something I will PM it to you.

I did not trade the 60x90 and would have to go and read that thread, but I do know I like entering trades off the BB instead of crosses. The entries can be 20 pips or so better sometimes. We learned that on the 6.0/6.1 versions.
  • #406
  • Jan 7, 2011 4:56pm Jan 7, 2011 4:56pm
I ran tests over the same period on the EURJPY. I only ran these tests with the small stop loss. That means that I am using half the distance from the entry to the prior swing high/low as the measure for stop loss.

The results for this pair were even better than the last two pairs I measured.

Total trades: 172
Winners: 65
Winning Percentage: 38%

This is remarkably consistent. With EU, it was the same. With USDDKK it was 40%. The consistency over what are probably unrelated pairs gives me much confidence in the strength of the system. Now for the cool number:

Expectancy: 1.02R
This means that on this run, for every dollar risked, you could expect to see a profit of $1.02. That is just through the roof extraordinary. Anything above a .10 or .15 expectancy can make for a nice system, provided you keep costs down and use good money management.

Over this run, if you risked 1% of your account/trade, a 10K account would grow to approx. 47K. At 2% risk/trade, the growth would be more like 10K to 150K. I'm attaching an equity curve that shows both the 1% and 2% growth projections.

The caveats I have for this are twofold. To trade this system well, you have to be willing to take losses. 4-5 losses in a row are frequent. Sometimes there are as many as 8 losses in a row. I've seen that for every pair I've tested.

Also, it's not easy to say what will be a 1% or 2% risk when you put the EA on. If you really wanted these results, you would probably have to stay at the computer 24/7.

Finally, I've tried to find some patterns that would eliminate losers. I can't see one. Yes, there are series of trades that lose when the EAs are converging. But then these are often followed by low risk, explosive winners. Getting rid of the losers in this way, from what I've seen, would lower both the risk and the profitability of the system. That's a question of judgment and taste.

But, let's put this in perspective. A system with an expectancy anywhere from .75 to maybe over 1 simply isn't broken. Far from it. There's no need to try to fix this. And, in my opinion, its almost stupid to try. Test it yourself. If you think there is some room for improvement, do the tests and show the results.
Attached Image
  • #407
  • Jan 7, 2011 6:24pm Jan 7, 2011 6:24pm
Ok, I have adapted the auto-trader so that it can also act as a buy/sell 'waiting' ea as do the buy/sell ea's here. Rather than put it here where it will get lost in about 20 minutes, I have posted it in the auto-trading thread at
http://www.forexfactory.com/showthre...29#post4288129

Please go there and read post 1 for details. The post is a bit complicated, but so is the ea - the more flexible a bot has to be, the more complicated it is to use.

I have not been able to test the code to destruction, so live traders might be better off sticking with the ea's here until we know for sure that my bot works as intended. I am putting it up for public consumption before the markets open next week so that coders can see if they can find any holes in the logic.

  • #408
  • Jan 7, 2011 9:28pm Jan 7, 2011 9:28pm
Quoting duffypratt
I ran tests over the same period on the EURJPY....
Duffy I love your research thanks for all the work.
  • #409
  • Jan 8, 2011 3:55am Jan 8, 2011 3:55am
Duffy,

Please don't tell me you are using the TMA when visually scrolling the chart and manually calculating the trades...
Please tell me you are using TMAtrue...

I know i will get Bob angry at me by yet again putting up the TMA/TMAtrue fight, but everyone must be definitely convinced that using TMA rather than TMAtrue for visual backtesting on charts is PURE CRAP!

I fully agree with Bob that TMA and TMAtrue make no difference when used on the CURRENT BAR: they both give the same result, this is ABSOLUTELY TRUE.

However, this is not true anymore on the PAST BARS.
TMA gives an absolutely unbelievable visual result on charts.
To make it short, TMA uses FUTURE BARS to compute its current value.
Which obviously cannot be the case for the current bar...



SO PLEASE: DO NOT USE TMA FOR VISUAL BACKTEST !
USE ONLY TMAtrue !


You can use TMA or TMAtrue for real time trading, they will make no difference to measure crossings or bands breakout.

I hope this is clear enough for everyone...

Now, Duffy, if you have used TMAtrue for your backtests, then this is pure GOLD.
We should all reward you for this research, and Bob for this yet-another-fantastic-work !

Bod, as you ironically said earlier : just a 1$ fee per download (who can't afford that???) of your post-1 files would already make a bit of your living...

Sq.

quote=duffypratt;4297283]I ran tests over the same period on the EURJPY. I only ran these tests with the small stop loss. That means that I am using half the distance from the entry to the prior swing high/low as the measure for stop loss.

The results for this pair were even better than the last two pairs I measured.

...[/quote]
  • #410
  • Jan 8, 2011 4:18am Jan 8, 2011 4:18am
Quoting squalou
Duffy,

Please don't tell me you are using the TMA when visually scrolling the chart and manually calculating the trades...
Please tell me you are using TMAtrue...

I know i will get Bob angry at me by yet again putting up the TMA/TMAtrue fight, but everyone must be definitely convinced that using TMA rather than TMAtrue for visual backtesting on charts is PURE CRAP!

I fully agree with Bob that TMA and TMAtrue make no difference when used on the CURRENT BAR: they both give the same...
[/quote]

But the TMA is not the point of entrance, whether you use TMA True or TMA doesnt change the entrance point. Now it may affect the exit because the lines have moved if you are using the bottom line to TP but it would not effect the entrance. Either way the info about where one takes the entrance and SL hits would not change however the amount of TP might. The TP area, however, is discretionary because in back testing we usually take the more ideal exit but in real life we may not. So his info as far a percentage of profitable trades to SL would not change but the amount of profit could be idealistic. Anyways welcome, are you interested in developing the indicator that would give us the angle of distance from the TMA center to the 240MA. Something like a MACD would be nice. I would care if you used the TMA True or TMA.

Oh by the way when I developed the system I used TMA True. Just thought I would let you know.
  • #411
  • Jan 8, 2011 6:02am Jan 8, 2011 6:02am
Thanks bob for the explanations.
I am just very concerned about the results Duffy posted: the return is so high that it puts doubt on the type of TMA he used, especially because he scrolled the chart to find the entries and exits.
As far as i understand , TMA Bands are used to enter and/or TP, and TMA Bands directly depend on the type of TMA used.

As for the indicator development, i would be happy to help if i can.
One remark: i saw earlier in the thread the "TMA Angle" MACD-looking indicator that Kumars4x posted (very nice work!): it plots the diff between the 240MA and TMA(true), along with the convergence/divergence plotted with different colors.
Isn't that what you are looking for ?
Please post details of your request here, i'll check it out.

Sq
  • #412
  • Edited 8:52am Jan 8, 2011 8:42am | Edited 8:52am
Quoting squalou
Thanks bob for the explanations.
I am just very concerned about the results Duffy posted: the return is so high that it puts doubt on the type of TMA he used, especially because he scrolled the chart to find the entries and exits.
As far as i understand , TMA Bands are used to enter and/or TP, and TMA Bands directly depend on the type of TMA used.

As for the indicator development, i would be happy to help if i can.
One remark: i saw earlier in the thread the "TMA Angle" MACD-looking indicator that Kumars4x posted (very nice work!): it plots the...
Actually it is close to what I wanted, looks like it works good maybe a hair off but close enough. Dont know why I asked for it again. Sorry. Actually I will add to page 1 when I do my next update.
  • #413
  • Jan 8, 2011 12:06pm Jan 8, 2011 12:06pm
Quoting squalou
Duffy,

Please don't tell me you are using the TMA when visually scrolling the chart and manually calculating the trades...
Please tell me you are using TMAtrue...

I know i will get Bob angry at me by yet again putting up the TMA/TMAtrue fight, but everyone must be definitely convinced that using TMA rather than TMAtrue for visual backtesting on charts is PURE CRAP!

I fully agree with Bob that TMA and TMAtrue make no difference when used on the CURRENT BAR: they both give the...
[/quote]

I'm using TMA true for the tests. There may be some problems with my methodology, but the indicators aren't a problem. It's impossible to test for slippage. It's nearly impossible to filter out news. In areas where there may be judgment involved, its almost impossible to know whether I would make the same decision in real time. (On the last, I try my best to err on the side of unprofitability, so the results will be conservative at least in that way.)

I went through this before when Bob first posted version 6. I tested it using TMA centered at first. I was ready to get my Maserati, the results were so good. Then I tested it again using the strategy tester, and had a rude awakening.

There is one way in which the TMA Centered may differ, and be more useful, than TMA True in live trading. It gives a very smoothed look to price movement. If any decisions ultimately get made based on the shape of the TMA curve immediately before the trade is taken, then visual backtesting will become impossible (and I would then move to the strategy tester, which takes much more time).

Hope this helps. And if it makes you feel better, I've come across a pair that does not perform as well as the three I've already tested. I'll have results on that pair later today. It seems to do about half as well, at least through the first year of tests. And it is prone to losing streaks. More later.
  • #414
  • Edited 12:33pm Jan 8, 2011 12:16pm | Edited 12:33pm
I am not disputing the fact that your NB 1.0 Trend system is not profitable. It very well may be.

No one is disputing the fact that the TMA Centered and TMATrue are the same in real time (for the current bar). Everyone can clearly see this by looking at a chart with both indicators attached.

Quote
But the TMA is not the point of entrance, whether you use TMA True or TMA doesnt change the entrance point.

No one is disputing this either. We all know that the trade is entered by the 5x5 EA when the price crosses the 5x5 MA going in your trade direction.

Our issues are with the triggers used - the criteria that must be met before turning on the 5x5 EA. For example, if we are looking for a Long trade, we must have the triggers:

- In an uptrend - the center line of the TMA must be above the 240 MA.
- Price retracing to the bottom TMA line - anything close, touching, or crossing the green line is acceptable.

When looking at historical data, the TMA Centered indicator does not accurately reflect the actual TMA values during those bars. See charts below - the first chart reflects trades #1 and #2 from your manual using TMA Centered, the second chart is the exact same except it uses TMATrue.

Please explain how my analysis is incorrect?
Attached Image (click to enlarge)
Click to Enlarge Name: tma.gif Size: 78 KB
  • #415
  • Jan 8, 2011 3:04pm Jan 8, 2011 3:04pm
Quote
I'm using TMA true for the tests.

Good to hear that!

Slippage, news, etc, should not be a so big deal, they are probably negligeable compared to the profits.

Quote
There is one way in which the TMA Centered may differ, and be more useful, than TMA True in live trading. It gives a very smoothed look to price movement.

Agreed, TMA Centered gives a good feeling of overall past price movement for sure, although it "degrades" as you are nearing the current bar.

Quote
And if it makes you feel better, I've come across a pair that does not perform as well as the three I've already tested. I'll have results on that pair later today. It seems to do about half as well, at least through the first year of tests. And it is prone to losing streaks. More later.

Whôt ???? i thought this system was the Holy Grail ?!?

Good work Duffy for the manual backtest, this is very helpful and will certainly serve as a reference to estimate the EA's performance.
If you saved the list of trades on one of the pairs you tested -- unless you just noted the trades results in your spreadsheet, which i could easily understand
If you are going to test another pair manually teh same way, may i suggest you to draw a vertical line on top of the opening and the closing candles for each trade ?
That is very quick to do with MT4;
Then save the resulting chart as a template and post this template, so we can all see those trades.
-- not an obligation at all of course, this is just an idea to have a reference to compare with the EA trades run thru the backtester on the same pair.
It could also serve for discussions on possible areas of improvements perhaps.

Thanks Duffy for this hard work.
Sq
  • #416
  • Jan 8, 2011 7:01pm Jan 8, 2011 7:01pm
Quoting zznbrm
I am not disputing the fact that your NB 1.0 Trend system is not profitable. It very well may be.

No one is disputing the fact that the TMA Centered and TMATrue are the same in real time (for the current bar). Everyone can clearly see this by looking at a chart with both indicators attached.



No one is disputing this either. We all know that the trade is entered by the 5x5 EA when the price crosses the 5x5 MA going in your trade direction.

Our issues are with the triggers used - the criteria that must be met before turning on the 5x5 EA....
When I was testing the 5MA series, it did not take long before I realized there needed to be another filter. One that improved the odds by showing that price had retraced to a point where you could turn on the EA. The second thing I needed was something that would define the trend. In reality any tunnel type system would probably work. Bollinger bands were too erratic and other systems like Vegas, Cog, etc. basically all had the same problem. I needed something that gave a range. That is what TMA does based on ATR. There are other tunnel systems based on other indicators but the basic idea is the same. We want to enter on the retrace and second we want to avoid as much as possible the peeker trade. Those are the three problems.
1. What is a trend?
2. What is a retracement?
3. Avoiding peeker trades?

Now due to the fact that prices fluctuate and there is no real science to it you go with what may work best. The problem is this: Is TMA true any more accurate than TMA regular? First off you must set them to the same settings to compare. TMA true is set at 3.0 and TMA is set to 2.0 but I corrected that in my chart template. Even though there are differences I dont consider them significant because in real time they both say the same thing. So ............... here we go again. What do we do? If we take TMA off the screen then I would have to teach when price breaks a clear distance from the 5MA you want to turn it on in the direction of the trend. The TMA just gives a visual of that process. I mean if the TMA bothers you that much take the outside lines off the screen and just use: when TMA center is above the 240MA only buy and when it is below only Sell. Well the first thing you will need is: When do I reenter? Well, some kind of tunnel system would work here. Which one should I use? They all will move. You could use a MA with 100 pips above and below but then you would have to develop different distances for every pair. Why do all that work when there are tunnel systems already there to use. TMA uses ATR so that makes some sense.
Four rules of trading:
1. Follow the Trend. OK define it.
2. Buy the dips, sell the rallies. OK define it
3. Where do you close out your profit? Ok define it
4. Where and when does the trend change? Ok define it.

That is what this does. If you think it will fit every single trade scenario you are smoking something. However, and this is the question, Does this give you an edge in trading? AHHHHHHHH that is the million dollar question. Starting Monday I am putting up 500 bucks to find out. If I start racking up profits over a long term of time all the questioners will have to go run and hide like they did when I showed them you could counter trade with high leverage and the RWB indicator. Same problem was presented then, the MACD line moves The MACD line moves, .... well 1 year and over 10,000 pips later no one is arguing with me anymore on that issue. It took govt. regs to blow that system out of the water.

Starting Monday I will attempt to prove that moving TMA lines are not a real problem. I wish my self luck.
  • #417
  • Jan 8, 2011 7:05pm Jan 8, 2011 7:05pm
Quoting zznbrm
I am not disputing the fact that your NB 1.0 Trend system is not profitable. It very well may be.

No one is disputing the fact that the TMA Centered and TMATrue are the same in real time (for the current bar). Everyone can clearly see this by looking at a chart with both indicators attached.



No one is disputing this either. We all know that the trade is entered by the 5x5 EA when the price crosses the 5x5 MA going in your trade direction.

Our issues are with the triggers used - the criteria that must be met before turning on the 5x5 EA....
The quickest and shortest answer to your question is this.
1. They both show the same thing in real time so it is not an issue.
2. We saw we had plenty of possible trades in the first week and we had more winners than losers and most of our winners were bigger than our losers.
3. So ...................... we look to the future not the past and I am not so sure that TMA true is so true. That is an assumption that may not be true.
  • #418
  • Jan 8, 2011 7:35pm Jan 8, 2011 7:35pm
Quoting squalou
Good to hear that!

Slippage, news, etc, should not be a so big deal, they are probably negligeable compared to the profits.



Agreed, TMA Centered gives a good feeling of overall past price movement for sure, although it "degrades" as you are nearing the current bar.



Whôt ???? i thought this system was the Holy Grail ?!?

Good work Duffy for the manual backtest, this is very helpful and will certainly serve as a reference to estimate the EA's performance.
If you saved the list of trades on one of the pairs you tested...
It would take too long for me to write down dates and times of the trades. I used to do that for my own back tests. This is just quick and dirty. I might play with the vertical lines some. I think that's a good idea, if someone was really interested in seeing how the tests work. I hadn't thought about saving it as a template. That's a good idea.
  • #419
  • Jan 8, 2011 7:48pm Jan 8, 2011 7:48pm
I ran two years of tests on this pair, using the same system. Tests run with Bob's 1.0 Trend With TMA True template. Stop loss set at 1/2 the previous swing low/high. Take profit at opposite TMA line. No break even.

Total Trades: 222
Winners: 77
Winning Percentage: 35%

Expectancy: .56

This pair seems much more erratic than the other three I have tested. It doesn't swing. Instead it seems to jerk around wildly. I passed on many winning trades because the 5 hadn't moved close enough to the band yet. Then, when they did come together, the trend would already have turned on a dime, and there would be no retrace. Even so, the results are pretty good, just not as great as on the other pairs.

I didn't test it, but there are many consecutive losing trades when price the TMA is coming strongly back to the 240. And this pair didn't seem to offer alot of big trades at the end of those runs as compensation for the losses. Or at least not as regularly.

And there were many losing streaks: 1 nine, 2 eights, 3 sixes, and 3 fives. Each of those gives a pretty hefty drawdown. The nine in a row should be pretty obvious on the attached equity curve. Again, the curve shows growth for both1% and 2% equity risked/ trade.
Attached Image (click to enlarge)
Click to Enlarge Name: Untitled.jpg Size: 54 KB
  • #420
  • Edited 10:47pm Jan 8, 2011 10:36pm | Edited 10:47pm
Quoting duffypratt
I ran two years of tests on this pair, using the same system....
I think you will find this true with most any pair with the CHF. It is the one that gave me the most fits this last year counter trend trading. If it wasnt the gbp/chf then it was the usd/chf or eur/chf or chf/jpy. The CHF was the haven currency from the Euro and GBP and every time news happened people would hide their money there until things got better and then they would sell out causing long trends as they jumped in for long periods and then jumped out for long periods as they went to other investments. So I guess your analysis doesnt surprise me. So it had a tendency to jump around unpredictably.
0 traders viewing now