Extracted Thread (per thread starter's request) | Page 174 | Forex Factory

Options

Go to
Permalinks
Print Thread

Similar Threads

1 trade per pair per year 6 replies

multiple systems per pair/multiple pairs per system? 3 replies

Extracted Post (per poster's request) 0 replies

Trading 24 hours per day, 5 days per week 14 replies

Attachments: Extracted Thread (per thread starter's request)
Exit Attachments

Extracted Thread (per thread starter's request)

  • #3,461
  • Sep 13, 2012 5:23pm Sep 13, 2012 5:23pm
Quoting dropkick-cat
The Final Edition Guide says for "Box Ranges: Maximum height in pips...". So decreasing this parameter should only decrease the number of CZs found, right?

Altering only this parameter from 20 to 19 to 18 gives 50, 23, 43 CZs as seen in the screenshots.

This is on EU M15. Am I missing something, or may that be a bug?

Thanks,
DKC

edit: for other values <18 or >20 it works as expected: less pips --> less CZs, more pips --> more CZs. It is just with 19 and some decimals around.
I can't say it's a bug as there other factors involved. Changing the maximum will eliminate the bigger CZs. That's correct. But then this affects the order simulation. It is possible that some of the other larger CZs were still in play (the trade set hadn't completed yet) when other smaller CZs occurred. Because of that, they wouldn't be in the list of CZs for the larger one. They only become visible when the larger one is ignored. Does that make sense?

Hopefully it's not a bug.
  • #3,462
  • Sep 13, 2012 5:29pm Sep 13, 2012 5:29pm
Quoting forexhard
Well if you know it, then why don't you stop? (<--Rhetorical question.)

There are dozens of possible tools at our disposal, some are less risky to use, and some are more, and I think it's wrong to ignore any of them; I think all of them should be given some attention some of the time, including so-called tabboo subjects like Martingaling. But there always needs to be a balance, and once the tipping of the scales becomes too obvious, it's time to reevaluate priorities and get things back on track. So work with me guys: tabboos are fun to...
Point made.....apologies.
  • #3,463
  • Sep 13, 2012 6:42pm Sep 13, 2012 6:42pm
Guys,

Can I just alert you to a fairly major bug I found in the indicator.

Let's say you are trading off the EURUSD 15 min. Assume the top of the CZ triggers a long (trade 1). Along comes some news and in ONE candle the price drops down to your short trigger (trade 2) and then shoots up to your trade 1 TP.

You would think (correctly) you have just lost 2 trades. Instead, the indicator shows only a TP off the first trade! This is because the indicator HISTORICALLY doesn't know which got hit first, the trade 1 TP or the trade 2 trigger and defaults to the TP! The widening of spread during news makes this problem even worse. You can verify this by checking the 5 min or 1 min charts to see how the candle developed on the 15 min.

Now the indicator probably shows correctly if you are watching it live, but be aware all those pips you would have won historically could be way out, especially with a narrow CZ and a few straddling candles.

Apologies for not posting a chart but I am on the road at the moment.
  • #3,464
  • Sep 13, 2012 7:28pm Sep 13, 2012 7:28pm
Quoting Fizzerr
Guys,

Can I just alert you to a fairly major bug I found in the indicator.

Let's say you are trading off the EURUSD 15 min. Assume the top of the CZ triggers a long (trade 1). Along comes some news and in ONE candle the price drops down to your short trigger (trade 2) and then shoots up to your trade 1 TP....
Yes, be careful, always, but this is not an idicator bug. You have discovered what is very likely the weakest part of an automated system, which is that every trade will be accounted for regardless of surrounding price action that we might have ignored (or would have happened too quickly to react to) when trading live. In this case, spike candles pierced clean through both the upper and lower boundaries of the CZ, triggering a sequence of Closes and Opens in quick succession on the same bar. Nothing is really "wrong" with this, since price action is price action is price action, and luckily these types of quickly zig-zagging sequences happen only rarely. For me, this only reinforces my committment to proper MM due to the fact that no matter how well prepared we are as traders, we really never can tell what price is going to do next, so we should always be prepared for the worst -- and this definitely qualifies.

Both pics show the same thing, I just zoomed in on the 2nd one. The one tall blue candle is the real culprit here:
Attached Images
To improve is to change. To perfect, is to change often.
  • #3,465
  • Sep 13, 2012 7:46pm Sep 13, 2012 7:46pm
Quoting subs
I can't say it's a bug as there other factors involved. Changing the maximum will eliminate the bigger CZs. That's correct. But then this affects the order simulation. It is possible that some of the other larger CZs were still in play....Hopefully it's not a bug.
Exactly right, Subs! Very cool that you recognize what is happening here. The CZ's are not static artifacts, they are extremely fluid and changeable, and changing any one of them has the potential of altering several CZ's in succession. This is due to the fact, as you were pointing out, that the next CZ in line only pops up when the trade set in front of it has completed. So very time you change a parameter, and that parameter alters a trade set, then every other CZ around it will be affected as well. This is why small changes to the inputs window can create very big changes to the diagnostic display, and can give back the type of results that Dropkick-cat found confusing. Hopefully it will become clearer to him as he plays around with it some more.
To improve is to change. To perfect, is to change often.
  • #3,466
  • Sep 13, 2012 7:59pm Sep 13, 2012 7:59pm
Quoting forexhard
Yes, be careful, always, but this is not an idicator bug. You have discovered what is very likely the weakest part of an automated system, which is that every trade will be accounted for regardless of surrounding price action that we might have ignored (or would have happened too quickly to react to) when trading live. In this case, spike candles pierced clean through both the upper and lower boundaries of the CZ, triggering a sequence of Closes and Opens in quick succession on the same bar. Nothing is really "wrong" with this, since price action...
Amazing, but the market always seems to come up with something that just kills an EA stone dead, yesterday was a case in point for me.

You only need this to happen twice in a month, even with conservative MM multiple trading off a single box, your account will be devastated.

I have now proved that with all the fine tuning of back testing, optimising of stops and profit levels, etc etc, the market will be different going forward, and there is no way of controlling that or dealing with it with an automated type of system.

It sure looked good for a while, but it is goodbye for me, which is sad because I really felt I had found a way of trading successfully here, but it was not to be. Pity.

Larry
  • #3,467
  • Sep 13, 2012 9:12pm Sep 13, 2012 9:12pm
Quoting forexhard
Exactly right, Subs! Very cool that you recognize what is happening here. The CZ's are not static artifacts, they are extremely fluid and changeable, and changing any one of them has the potential of altering several CZ's in succession. This is due to the fact, as you were pointing out, that the next CZ in line only pops up when the trade set in front of it has completed. So very time you change a parameter, and that parameter alters a trade set, then every other CZ around it will be affected as well. This is why small changes to the inputs...
Like I said, I did do my homework.
  • #3,468
  • Sep 13, 2012 9:20pm Sep 13, 2012 9:20pm
Quoting LarryK
Amazing, but the market always seems to come up with something that just kills an EA stone dead, yesterday was a case in point for me.

You only need this to happen twice in a month, even with conservative MM multiple trading off a single box, your account will be devastated.

I have now proved that with all the fine tuning of back testing, optimising of stops and profit levels, etc etc, the market will be different going forward, and there is no way of controlling that or dealing with it with an automated type of system.

It sure looked good...
Sorry Larry, but except for the "the market will be different going forward" bit, I simply can't agree with most of what you just wrote.

For starters, taking that same pic into consideration, that E/U trade set only had 1+3 trades, which is pretty typical all things considered, and it actually ended as a winning trade set. The only way this would "devastate" someone's account would be if they were using really bad MM, because it would mean reaching margin call midway through the set. So this is not the EA's fault, or even the system, it would be 100% the fault of the trader to over-extend so far that just 3 bad trades would hit margin.

Your final paragraph is a bit vague. Do you mean that you're no longer interested in the EA or the SSBO system itself? Since my personal experience with SSBO is now bordering on 6 years I sure hope it's not the 2nd part, because I don't see anything in the markets today that I haven't already survived a hundred times.
To improve is to change. To perfect, is to change often.
  • #3,469
  • Sep 14, 2012 2:10am Sep 14, 2012 2:10am
Quoting forexhard
Sorry Larry, but except for the "the market will be different going forward" bit, I simply can't agree with most of what you just wrote.

For starters, taking that same pic into consideration, that E/U trade set only had 1+3 trades, which is pretty typical all things considered, and it actually ended as a winning trade set. The only way this would "devastate" someone's account would be if they were using really bad MM, because it would mean reaching margin call midway through the set. So this is not the EA's fault, or even the system, it...
I think Larry has actually learned a lot but just doesn't realise it. You can back test and optimise as much as you want but it's pointless is what he has spent hours learning. The market is dynamic.

People here seem to be relying to much on the indicator (imo) and ignoring the bigger picture. Look at my 60 min chart we are clearly in an up trend on cable and close to testing this years high. If someone can get a margin call in this market they need a change of career.

I use your indicator on a live account and find it useful for entries on the 15m. I am selective with trades and time I trade and certainly would not short this market right now.

Point is the indicator works if you know how to work it and also have respect the market - like a woman there's no point arguing with it.
Attached Image (click to enlarge)
Click to Enlarge Name: fxh3.gif Size: 19 KB
  • #3,470
  • Sep 14, 2012 2:16am Sep 14, 2012 2:16am
Quoting forexhard
Sorry Larry, but except for the "the market will be different going forward" bit, I simply can't agree with most of what you just wrote.

For starters, taking that same pic into consideration, that E/U trade set only had 1+3 trades, which is pretty typical all things considered, and it actually ended as a winning trade set. The only way this would "devastate" someone's account would be if they were using really bad MM, because it would mean reaching margin call midway through the set. So this is not the EA's fault, or even the system, it...
I traded the GU not the EU, and believe me the price action was so quick that my EA could not react in time. If I had been at my screen I may have been able to reduce the damage, as it was the EA ended up taking an extra trade, just a dumb animal, what can you expect.

My account is intact, but would not be except for proper MM, just that there is no way back for an account when this sort of stuff happens twice a month.

Not to worry there are other ways of trading.

Cheers

Larry
  • #3,471
  • Sep 14, 2012 2:18am Sep 14, 2012 2:18am
Quoting subs
I can't say it's a bug as there other factors involved. Changing the maximum will eliminate the bigger CZs. That's correct. But then this affects the order simulation. It is possible that some of the other larger CZs were still in play (the trade set hadn't completed yet) when other smaller CZs occurred. Because of that, they wouldn't be in the list of CZs for the larger one. They only become visible when the larger one is ignored. Does that make sense?
Thank you, subs, it does make sense. What I find curious is that the trade set from July, 26, has not closed until now as seen in the screenshot. I have looked into the code, but it is way above my head. For the record (and Kissa), here are the settings:

Oanda
EUR USD, M15
AuthorizedSSBODynamicBreakoutBoxFinalEdition from post 1
BoxLength 23
BoxRange 19.0
BoxTimeFrame M15
BoxBuffer Pips 1.0
MinBarsClosedOutsideBO 0
MaxBreakoutTradesPerBox 4
InitialBreakoutDirectionOnly false
LongTrades true
ShortTrades true
QuickTakeProfit 20.0
StopLossOverride 0.0
EmergencyStopLoss 100.0
BoxMartingale true
AutoReduceBoxSize true
DaysBack 100

Setting MinBarsClosedOutsideBO to 1 results in the set being closed.
Attached Image (click to enlarge)
Click to Enlarge Name: ssbo_26_Jul.jpg Size: 83 KB
  • #3,472
  • Sep 14, 2012 2:57am Sep 14, 2012 2:57am
Interesting chart on cable at 8.50 ish today in London.

Nice break out overnight running straight through yesterdays high.

Consolidating now through Frankfurt/London openings and resistance at 94. I am going to wait for a box to form and look for a clear close above 200 before entering another long. Shorts too risky for me.

Its Friday and may see some profit taking dragging it back a bit. Who knows.
Attached Image (click to enlarge)
Click to Enlarge Name: fxh4.gif Size: 31 KB
  • #3,473
  • Sep 14, 2012 3:34am Sep 14, 2012 3:34am
Larry, and others, I read all the way down to post #3470 before making this post.

Quoting LarryK
Amazing, but the market always seems to come up with something that just kills an EA stone dead, yesterday was a case in point for me.

You only need this to happen twice in a month, even with conservative MM multiple trading off a single box, your account will be devastated.

I have now proved that with all the fine tuning of back testing, optimising of stops and profit levels, etc etc, the market will be different going forward, and there is no way of controlling that or dealing with it with an automated type of system.

It sure looked good...
You have encountered yet again (meaning several others also before you) the simple fact that EAs are dumb and usually not capable of dealing with impulse disturbances. EA developers in the past, including myself, concentrate on coding their rule-set without at first appreciating the whole army of caveats in place inside the human mind.

This simple fact has stopped me from coding EAs.

A real person would have done one of the following in yesterday's events: they would have closed out all their positions (automated or manual) and waited until after the event's transient effects, or they would wait for the impulse responses to play-out before resuming their plan; or maybe something else...

How does one code a capacity for reaction to news events into an EA? How does one even know in all cases whether events will cause upward or downward movement in an instrument? How does a machine interpret the news input? Impossible in my view.

Therefore a strategy is required which copes with events that cause major fluctuations in PA, and this needs to be built into the EA for SSBO. In order to do this one must have a means of detecting the onset of impulsive behaviour, such as rate-of-change of PA coupled with volume data, maybe it will be enough to detect sudden changes in the tick rate. Another possibility is to examine the calendar of events such as that posted by FF and have the EA 'un-hook' for a time period around the red-flagged events. Without some kind of strategy the spikes will always cause problems, including the unscheduled events such as Dexia going bust, tomorrows data being leaked - or whatever.

FH is right about watching position sizing. I do not believe one should abandon efforts just because the market will always contain unpredictable events. Dealing with these exceptional reactions is the challenge for any EA.

2-cents
Carbon-Dioxide: the gas of life! https://youtu.be/ttJBdr6eBuo
  • #3,474
  • Sep 14, 2012 3:56am Sep 14, 2012 3:56am
Something happened yesterday that sparked an idea for a change of setting to implement today....

I was away from my PC and had a pending order setup on the opposite side of my AU CZ in case price reversed. Remember, I'm not on an MT4 platform....

When my phone beeped to tell me that orders had been executed I was confused because I couldn't understand why I had such a quick order open and close.....I thought I'd accidentally setup up the wrong limit and hadn't placed it properly....

What actually happened was that price moved so quickly that whilst my pending order executed and hit TP within 40 seconds on my non MT4 platform (yay!) on the MT4 platform as FH said the price moved too quickly for the platform and indicator so in reality that big blue candle was the stopper and TP closer all in one!

Below a demonstration...

MT4

Attached Image (click to enlarge)
Click to Enlarge Name: mt4.jpg Size: 87 KB


ETX

Attached Image (click to enlarge)
Click to Enlarge Name: ETX.jpg Size: 94 KB


Now the problem I had with using the demo for CZ confirmation and my non MT4 platform was that spreads were higher on my live account (non MT4). This meant that stops were higher.

It got me thinking......what's the one setting I haven't played with so far....someone asked a question yesterday about stop sizes and my response was use what's best for you, but I just use CZ instead of custom stops...

Well today, having seen the chaos of yesterday, I thought why not try fiddling with the stop size. Wow was I surprised!

I've noticed my CZ's per week on the AU rising significantly....3.4 if you check my MT4 pic above....

So today I had a little play and look what changing 1 pip in box size and increasing the stop size does....

Attached Image (click to enlarge)
Click to Enlarge Name: new au.jpg Size: 90 KB


Incidentally, apologies if this has been covered but you see that latest CZ....well that didn't show up automatically. It was only when I made a change to the indicator this morning that it showed.

In fact as a matter of course I now go into my settings and change nothing on the indicator and come back out purely to refresh (if that's the word) the indicator on the chart and most times a new CZ shows up....shame I didn't do it yesterday else I would have been 30 pips better off today!

Anyone else had this issue?
  • #3,475
  • Edited 11:42am Sep 14, 2012 6:57am | Edited 11:42am
+ The EA can deal during the night and most of all is free of emotion
- But the EA must be robust to a lot of situations (market + platform)

+ Manual trading allows "trivial" erronous trades to be bypassed
- needs experience and it is not easy to master our emotions

BUT

If the market had strange spikes this week, this has nothing to deal with that question. It is purely a problem of how to handle a market behavior. Nobody can predict this and everybody MUST take ino account a percentage of losses due to strange market behaviour.

I am coding an EA and it works because it is simple. The simplest things are the most robust. Places an order and "the world can crash" after that, the trade will go as scheduled. And i limit the amount of trades per hour, filter the "0h00" candle, etc... then the EA is more and more robust. SO I CAN'T BLOW MY ACCOUNT WITH THAT.

This week 3 good trades, 1 bad and no problem with the spikes (trading on EURUSD M15, live).

The return of my experience for this week : choose a trading style that suits best for you (EA or manual), don't change, and enhance, enhance, enhance...

Don't use martingale at first and verify you are always in large profit, then you can add this kind of pip turbo booster if you like... but this has always killed a lot of accounts... Use proper MM to limit its effects (depth of 3 for example).
  • #3,476
  • Sep 14, 2012 7:34am Sep 14, 2012 7:34am
Quoting C345
Interesting chart on cable at 8.50 ish today in London.

Nice break out overnight running straight through yesterdays high.

Consolidating now through Frankfurt/London openings and resistance at 94. I am going to wait for a box to form and look for a clear close above 200 before entering another long. Shorts too risky for me.

Its Friday and may see some profit taking dragging it back a bit. Who knows.
Closed above the 200 and box formed easy 20 pips. Now the profit takers are in. Some of you guys are really making hard work of this for no reason. Tell me if I'm wrong FXH.
Attached Image (click to enlarge)
Click to Enlarge Name: fxh5.gif Size: 30 KB
  • #3,477
  • Sep 14, 2012 12:14pm Sep 14, 2012 12:14pm
As long as we're on the subject I'd like to point out that the EA Larry has been talking about, was one he had made for himself privately. So neither Xux nor I had anything to do with it.

I'd also like to remind everyone that I have repeatedly said to never to use an/our EA on a live account. I recommend that any software generated by this thread be used as a diagnostic tool only, and that it should never be linked to a live account.

That said, Xux sent me the latest overnight, and it looks fantastic.
To improve is to change. To perfect, is to change often.
  • #3,478
  • Sep 14, 2012 12:27pm Sep 14, 2012 12:27pm
Quoting C345
....Some of you guys are really making hard work of this for no reason. Tell me if I'm wrong FXH.
No, you're not wrong. It's human nature to second guess or doubt yourself and others. There's a lot to "unlearn" if you're going to follow SSBO as it's intended. It takes a while before it sinks in just how simple this system is, and until that revelation kicks in for real, there is the tendency to overreact or over-compensate. The really good ones are the guys that, to put it bluntly, just don't give a shit. They set their limits and pull the trigger and walk away and that's the end of it. I'm doing my absolute best to bring as many of you guys as I can to that exact point, but in the end it's not really up to me. I can say, say, say how simple it all is, but understanding it for real has to come from the inside.
To improve is to change. To perfect, is to change often.
  • #3,479
  • Sep 14, 2012 1:05pm Sep 14, 2012 1:05pm
Hey guys, This started happening today on my charts. The spikes last for about 2 seconds. It also has a tendency to happen when I change time frames. Has anyone else seen this?...

(I've already PM'd Xux about it, so don't bother.)
Attached Image
To improve is to change. To perfect, is to change often.
  • #3,480
  • Edited 1:40pm Sep 14, 2012 1:30pm | Edited 1:40pm
Quoting forexhard
Hey guys, This started happening today on my charts. The spikes last for about 2 seconds. It also has a tendency to happen when I change time frames. Has anyone else seen this?...

(I've already PM'd Xux about it, so don't bother.)

Nope.
Got 12 charts open and change TF quite often.


---

I am really happy about the EA coming along, my plan is to test it and if i am happy with the trade management it has, use it to enter trades after confirmation and / or manually if possible.

Tried about every trade manager EA there is, many have one good feature in them but to find all that i need combined into one has brought me no luck yet.


---

CHFJPY might be worth a look at Monday morning.
6 traders viewing now