Extracted Thread (per thread starter's request) | Page 372 | 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)

  • #7,421
  • Jul 2, 2014 2:54am Jul 2, 2014 2:54am
Quoting GU-Night
{quote} I believe that it is not always the top or bottom of the range that has to be broken. But another price within the CZ, and that is for me the reason that price sometimes/often comes back into the CZ to retest that price. So from my perspective, has the SSBO trader a bias that the top or bottom has to be broken or a bufferzone for confirmation of the direction, there is as far as I know no single proof for that. I began to think when I saw often ‘stairsteps’ within the CZ when I looked for hours and hours how a CZ was formed, which was ‘smart...
I understand your thinking better now, it's an interesting idea
I entered two trades mid CZ this morning again, lets see how they turn out - however this was because I arrived late to catch the buffer zone break out, nice to save myself some DD So I'd be interested in seeing how I could systematically apply this idea, if possible...

Next time I review my charts I'll see if I find something along the lines you are saying - a breakout point within the CZ, if I find something interesting I'll get back to you. Thanks for sharing your idea.

Meanwhile busy morning for me - 4 CZs already, guess it could be pre NFP/Fed chatter indecision... better to collect them now and stay away this afternoon me thinks...
  • #7,422
  • Jul 9, 2014 5:42am Jul 9, 2014 5:42am
Finished for the day, before tonight's FOMC minutes and Draghi Speech

Attached Image (click to enlarge)
Click to Enlarge Name: Untitled.png Size: 23 KB


Attached Image
  • #7,423
  • Jul 10, 2014 4:33pm Jul 10, 2014 4:33pm
Quoting Dzt
Finished for the day, before tonight's FOMC minutes and Draghi Speech {image} {image}
Nice tight settings Dzt

Meanwhile, I have 2 pairs with new CZs just forming, and they're precisely within the last CZs - to the pip! GU and AU. Hoping its a nice sign for a good weekend coming
  • #7,424
  • Edited 5:44am Jul 29, 2014 4:40am | Edited 5:44am
Anyone else had their Br-out trade %s go to hell this morning? Had about 3 go from 100% in 3 trades to 100% on Trade 5 or 6* (EU and GU)! I guess people are sensitive this week, due to the huge dump of news coming later and its causing lots of unusual ranging after long periods of low Vol...

Time to see if I can tweak my settings to see if it could have been avoided...

I've started an experiment by basically having a few 'good to average' settings on same pairs, but back testing in different environments/circumstances. Assuming that different settings do better under different conditions, I have a few open and will just watch them for a long time. If the stats improve as the circumstances change, I will start to trade it. Assuming the others I have set up do badly under different circumstances, then I demote those and put them on my watch list. This way I have a cycle of different settings whcih go through rotation as their stats change.
Its quite a delicate balance, and not sure if it would be advisable practice from any long term SSBOers, but I know that the last few months weve had unusual Vol levels and I assume the settings working under these conditions dont work as well as they would have another time. For example I tried to replicate some of the good looking charts posted on here a year or more ago, to see if i could - knowing the conditions were different when this thread was alive and kicking - but I couldnt even get it close after spending hours on it. The settings I have now that are (asides today) working well the last 100/150 days would not have worked well a year or two ago.

Also to add another layer to this Ive back tested each setting to different days back (yes I know the placement of the CZs change etc, its not a science here!) and its interesting to see the stats from 150 days ago vs the last 30 days and every divide in between. I note them all down and some do go through a cycle of getting better and worse. I visually go back over these periods with manual boxes too to get an idea if the indi is giving miss-representative data. Hope I explained myself well enough here! Anyway, I only have very provisional data so far, any feedback from remaining readers welcome! In a few weeks I'll know more

*Edit: and it was only last week (I think) when I read this post (http://www.forexfactory.com/showthread.php?p=6173510#post6173510)while going through the thread again
  • #7,425
  • Jul 29, 2014 9:12am Jul 29, 2014 9:12am
Interesting
  • #7,426
  • Aug 1, 2014 4:04am Aug 1, 2014 4:04am
[quote=gjgjg;7636943]Anyone else had their Br-out trade %s go to hell this morning? Had about 3 go from 100% in 3 trades to 100% on Trade 5 or 6* (EU and GU)! I guess people are sensitive this week, due to the huge dump of news coming later and its causing lots of unusual ranging after long periods of low Vol... Time to see if I can tweak my settings to see if it could have been avoided...


gjgjg I’m assuming you are dialling in your settings by running your back testing for more than a 100 days, if not then simply ignore what I posted below.

The dilemma you facing when adjusting your settings to get the perfect BO %, every time BO %’s gets knocked down, is that you unknowingly fall into the trap of curve fitting – ( If your back testing stretches to far back).

The concept of curve fitting is as easy as pie, the realisation that you doing it, is a bit different, it took me a while to realise it.

I would venture to guess the following is going to happen, you are going to spend a couple of hours maybe a day or 2 (Depending on how far back your back testing stretches ) to tweak your BO %’s. In 2 or 3 weeks time maybe a month maybe 2, those BO %’s are going to get blown out of the water, again. Then you are going to go back tweaking BO %’s, till the next time those %’s gets blown apart, you will probably do this until you get frustrated, thinking this indi doesn't work, when you reach that point of frustration, that’s when you know you busy with curve fitting.

I’m not saying ignore the % and other stats, but if you “trade” the stats instead of trading the breakouts the results are less desirable. My advice, avoid stretching those back test too far back.

Back testing gives one a false sense of security, live forward testing gives one confidence, here’s the catch though ………………… it takes time.
  • #7,427
  • Aug 8, 2014 7:28am Aug 8, 2014 7:28am
Hi All,

I'm late to the party but I would like to say thank you to forexhard and all of the other people who have made a positive contribution to this thread. I feel that in taking this system and making it my own, I have found a new level of confidence and I'm finally starting to turn a profit after several years of frustration. I'm using the latest indicator from xux which is fantastic and saves me a heap of time - I'm not a coder by any stretch of the imagination but manged to add an expectancy score to the HUD as this is a lot more useful than net pips to me.

I've come up with some rules that I follow to suit me and would like to share as this thread has given me a lot.

 

  1. I only take 1 breakout per CZ, and my reward to risk ratio generally ranges from 2:1 to 10:1 depending on recent backtests. Resetbox set to 0.
  2. At the start of every month I backtest and tweak the full set of instruments and pick the top 6 performing instruments to trade based on expectancy - generally 1-2 CZ's per week produces the best results
  3. For every trade I set a stop loss and TP and then leave it alone except for weekends where trades will be manually closed and not held
  4. If there's a great imbalance in the recent average breakout long vs. short (by a ratio of 3:1) then I only trade in that direction
  5. I only trade higher volatility days (Tues-Thurs) and NFP (no US public holidays). So far I have found the Frankfurt through to mid-London sessions produce the best results but I need more data to confirm

I've attached my forward test data collected from the indicator which I started collecting from the end of June, 2014. The forward test contains about 16-17 instruments per month traded all week (In Column C, 1 represents Monday, 2 Tuesday etc.) It doesn't contain spread, commissions, slippage etc so it will be overstated and can only be used as a guide.

  1. Using a bunch of top performing different pairs or indexes my forward test data shows a return of ~52R over a 1.5 month period with an expectancy of 0.32 per trade.
  2. If the data is filtered to only trade the more volatile Tues-Thurs, then the return is ~82R over a 1.5 month period with an expectancy of 0.87 per trade. Note, to achieve these results the win-rate is only around 24% but winners are on average 6.6 times larger than losers.

It's a real shame this thread is in the bin - hit me up on PM as I'd like to find some like minded individuals to discuss this system to get the best out of it.

Attached File(s)
File Type: xlsx SSBO forward test data.xlsx   38 KB | 585 downloads
  • #7,428
  • Aug 11, 2014 7:08pm Aug 11, 2014 7:08pm
Quoting Dzt
Quoting gjgjg
Anyone else had their Br-out trade %s go to hell this morning? Had about 3 go from 100% in 3 trades to 100% on Trade 5 or 6* (EU and GU)! I guess people are sensitive this week, due to the huge dump of news coming later and its causing lots of unusual ranging after long periods of low Vol... Time to see if I can tweak my settings to see if it could have been avoided... gjgjg I’m assuming you are dialling in your settings by running your back testing for more than a 100 days, if not then simply ignore what I posted below. The
...
Roger that Dzt! Yes ive fallen into the tweak to death regularly - curve fitting trap before, something i knew would happen and i wanted to avoid initially when i decided to only hand draw boxes in the beginning.
i feel like i have mostly overcome this by now, and view the indi as a helpful tool for spotting the best cz sizes for any given pair in a given period, im expecting my settings to break down once in a while, if they do then see if some small tweak would have helped or not, but ultimately trying not to lose sight of the core of it all. not being a slave to the indi but the other way round;p

the testing above i described is a way of preparing for a time when my current settings fail consistantly and become 'untradeworthy'. this way ill have been watching the evolution of other settings unfold while only trading the best settings in the mean time. i defo wont do too detailed back testing and use the numbers as sacred, especially since changing the DB amount can change when a cz forms and then a ripple effect through time! ive seen one occasion where changing the db by 1 day actually changed the stats from 75/75/100 to 20/30/0/0/50/50 (approx)! all because of where the indi started to form czs! gives one pause for thought and good insight! my aim is just to get a feel for the 'cycle' of each setting and see if i can gain a deeper understanding and feeling for things on a longer term basis, and to keep my perspective as wide as possible so as not to be stuck suddenly one day because my favorite settings fell out of fashion.
hope that makes sense!
  • #7,429
  • Edited 7:37pm Aug 11, 2014 7:16pm | Edited 7:37pm
Quoting Spinal
... It's a real shame this thread is in the bin - hit me up on PM as I'd like to find some like minded individuals to discuss this system to get the best out of it. {file}
hi Spinal. I've decided to continue to post here as a kind of homage to the thread (yep, weird) despite the lack of activity. maybe some more like minded folk will join and things will get going again. re the bin status - it actually probably deters most bubblegum traders id imagine, so maybe its for the best! Id be happy to share notes anyway, right now im tracking this lovely one...

Attached Image (click to enlarge)
Click to Enlarge Name: 12.JPG Size: 56 KB


Regarding your notes:
2. At the start of every month I backtest and tweak the full set of instruments and pick the top 6 performing instruments to trade based on expectancy - generally 1-2 CZ's per week produces the best results.

In my experience (which is not very significant at this stage) the lower number of CZs per week in the stats generally do give the best % winners ...in the stats, but I find these are more prone to 'break down' more often. It happened to me already twice with the 1-2pw settings I had and my feeling is that because there are only 1-2pw, that makes the stats less 'reliable' going forward because less CZs have been 'tested' by the indi. when there are 2-3pw there is the extra level of certainty that the stats will 'survive' going forward as more CZs have been tested. If the CZs are 3-4 pw then overtrading and messy overlaps etc are a possibility. hope that makes sense and feel free to call BS on this if you disagree, Im still learning.
  • #7,430
  • Aug 28, 2014 4:34pm Aug 28, 2014 4:34pm
Quoting gjgjg
{quote}In my experience (which is not very significant at this stage) the lower number of CZs per week in the stats generally do give the best % winners ...in the stats, but I find these are more prone to 'break down' more often. It happened to me already twice with the 1-2pw settings I had and my feeling is that because there are only 1-2pw, that makes the stats less 'reliable' going forward because less CZs have been 'tested' by the indi. when there are 2-3pw there is the extra level of certainty that the stats will 'survive' going forward as...
Hi gj, I know what you mean. If your sample size is very small then it's hard to tell if your settings are robust over the short term or not with only a few CZ's per week. You can try setting the "DaysBack" parameter to a larger number, say 500. However, the issue with that is the parameters that work well in today's low volatility environment will produce less signals when the volatility was historically higher. It's not an exact science but from all the testing I've done, my conclusion is that quality is certainly more important than quantity from an expectancy per trade perspective.

I've attached a trade I'm currently in on the EURAUD, 20 pip stop with a 160 pip target. It's about halfway to the target and may run out of steam and reverse back - but the euro is showing a lot of weakness at the moment.
Attached Image (click to enlarge)
Click to Enlarge Name: Screenshot2.png Size: 99 KB
  • #7,431
  • Sep 4, 2014 3:51pm Sep 4, 2014 3:51pm
Nice pattern. If someone wants to play with the indicator on the right side, attached below.
Attached Image(s) (click to enlarge)
Click to Enlarge Name: EURUSDfo.png Size: 58 KB
Click to Enlarge Name: EURUSDfo1.png Size: 54 KB
Click to Enlarge Name: GBPusdfo.png Size: 52 KB
Click to Enlarge Name: GBPJPYfo.png Size: 57 KB
Attached File(s)
File Type: ex4 TPO-Range-v3.ex4   40 KB | 760 downloads
  • #7,432
  • Sep 4, 2014 4:38pm Sep 4, 2014 4:38pm
Quoting GU-Night
Nice pattern. If someone wants to play with the indicator on the right side, attached below. {image} {image} {image} {image} {file}
Hey, thank you. But could you explain a little how it works? I can't see clearly....



see you
TONI
  • #7,433
  • Sep 5, 2014 5:25am Sep 5, 2014 5:25am
Quoting forexdaist
{quote} Hey, thank you. But could you explain a little how it works? I can't see clearly.... see you TONI
The indicator shows at which prices the market was trading at most of the time in the given period. With the vertical lines on the chart you can change the period you want to see.

I wrote earlier in this thread that I had the feeling that not the upside, downside or bufferzone but a particular price in the CZ has to be broken. This indicator shows that particular price. (that is how I read the indicator). A trained eye can see this information without indicator, but I do not have a trained eye, I’m lazy, and the indicator shows more objectivity than my subjective mind ever can show.

Regards.
Attached Image (click to enlarge)
Click to Enlarge Name: EURUSDfo3.png Size: 35 KB
  • #7,434
  • Sep 5, 2014 8:59am Sep 5, 2014 8:59am
Quoting GU-Night
{quote} The indicator shows at which prices the market was trading at most of the time in the given period. With the vertical lines on the chart you can change the period you want to see. I wrote earlier in this thread that I had the feeling that not the upside, downside or bufferzone but a particular price in the CZ has to be broken. This indicator shows that particular price. (that is how I read the indicator). A trained eye can see this information without indicator, but I do not have a trained eye, I’m lazy, and the indicator shows more objectivity...
Thanks for sharing the indi, Ive been thinking about what you said a few weeks ago about that and Ive reviewed all the Czs that appear on my charts and taken extensive notes on how many times the price retraces into the Cz before TP and how far and if it reaches TP or only goes a few pips and on which trade and how far it goes past TP on how many times etc etc etc and the results are very nice. In these low vol days it can present many good % chances to enter in better positions than the edge of the Cz (or in addition to it) and this indi could be a nice compliment to it
  • #7,435
  • Sep 5, 2014 9:07am Sep 5, 2014 9:07am
Quoting Spinal
{quote} Hi gj, I know what you mean. If your sample size is very small then it's hard to tell if your settings are robust over the short term or not with only a few CZ's per week. You can try setting the "DaysBack" parameter to a larger number, say 500. However, the issue with that is the parameters that work well in today's low volatility environment will produce less signals when the volatility was historically higher. It's not an exact science but from all the testing I've done, my conclusion is that quality is certainly more important than quantity...
Ive tried a few longer DB settings on a few pairs, some do very well for long periods (if the data is available on them) but it does mean longer forward testing! However if you find a setting that has the same results for 250 days compared to one that results fall appart above 100 then theres obviously a better one to pursue!

Nice job selling Euro, and 160QTP is really nice if you can get it I guess you would have made it after the last few days!
Sadly I missed the huge down move yesterday I wasnt at my PC I usually have 1 lot without a TP and try to catch a wave or BE with it; yesterday i wouldnt have closed it until it started aiming up slightly after about 195p in the green lol what a day to skip trading
Attached Image (click to enlarge)
Click to Enlarge Name: Capture.JPG Size: 40 KB
  • #7,436
  • Sep 6, 2014 4:23pm Sep 6, 2014 4:23pm
Quoting gjgjg
{quote} Thanks for sharing the indi, Ive been thinking about what you said a few weeks ago about that and Ive reviewed all the Czs that appear on my charts and taken extensive notes on how many times the price retraces into the Cz before TP and how far and if it reaches TP or only goes a few pips and on which trade and how far it goes past TP on how many times etc etc etc and the results are very nice. In these low vol days it can present many good % chances to enter in better positions than the edge of the Cz (or in addition to it) and this indi...
Good work, gjgjg. I’m glad to hear that I’m not crazy, or didn’t you say that? My English isn’t what it ever was. Excusez moi. Es tut mir leid. Het spijt me. Lo siento mucho. I’m sorry.

I don’t know which style is the better one. I think it is personal preference.

Last week I set up the Rossbot stairsteps EA on a demo account, made by XUX99 on 12 (or so) pairs with all the same settings (only boxrange differs for high volatile pairs). Risk reward 1:1. No bufferzone. No close outside box. EMA 3. Newsfilter 10 minutes before and after high impact news releases. 2 trades max per CZ. No martingale.

If my math is correct (I drink some beer at the moment) it would have made more than 220 pips this week. Have to say that not all the trades were closed at RR 1:1 because I had set some pairs on manual. (Not all winners, losers too.)


My intention is to let it run 24/5 for a month of three, maybe longer, and see what happens. Until then I trade the retraces as shown in the posts above.

Cheers.
Attached Image (click to enlarge)
Click to Enlarge Name: Rossbot.png Size: 9 KB
  • #7,437
  • Sep 8, 2014 5:56am Sep 8, 2014 5:56am
Quoting GU-Night
{quote} Good work, gjgjg. I’m glad to hear that I’m not crazy, or didn’t you say that? My English isn’t what it ever was. Excusez moi. Es tut mir leid. Het spijt me. Lo siento mucho. I’m sorry. I don’t know which style is the better one. I think it is personal preference. Last week I set up the Rossbot stairsteps EA on a demo account, made by XUX99 on 12 (or so) pairs with all the same settings (only boxrange differs for high volatile pairs). Risk reward 1:1. No bufferzone. No close outside box. EMA 3. Newsfilter 10 minutes before and after high...
No worries, de nada, de rien, dim problem etc I don't recall calling anyone crazy recently, but if I did - sorry!

Good that youre getting results from that EA. I admit I tried it recently, but couldnt get it running straight away so I left it alone. Honestly I'd be very skeptical at this stage (that Im at) to use any mechanical robot to run things for me. Maybe after a few years of staring at SSBO Cz's it will be needed but for now I want to run things at my own hands to ensure I know what the best settings are etc.

Did you have any problems getting it running?
  • #7,438
  • Sep 8, 2014 6:08pm Sep 8, 2014 6:08pm
Quoting gjgjg
{quote} No worries, de nada, de rien, dim problem etc I don't recall calling anyone crazy recently, but if I did - sorry! Good that youre getting results from that EA. I admit I tried it recently, but couldnt get it running straight away so I left it alone. Honestly I'd be very skeptical at this stage (that Im at) to use any mechanical robot to run things for me. Maybe after a few years of staring at SSBO Cz's it will be needed but for now I want to run things at my own hands to ensure I know what the best settings are etc. Did you have...
I had no problems getting her (Rossbot) running. I don’t like robots, normally. Maybe because I have almost none experience with them. That is what I’m going to find out next months. It is more that I want to compare my manual retrace results, with her mechanical “in at breakout” SSBO results.

I don’t know at which stage you are right now, but I think that the TPO indicator can help to make better trading decisions too while trading SSBO as teached in this thread by Forexhard.

If I take a look at the CADJPY picture below (extreme example, no bufferzone), it broke out the CZ. But above that CZ is a resistance area. This can happen too above a bufferzone. The SSBO indicator gave a buy signal straight into resistance, for me it is a 50/50 trade. I shouldn’t take that one. Probability far to low for me. If it breaks the yellow dotted line, it meets another at 23 pips (counted from top CZ).

To reach Risk Reward 1:1 it has to break that first resistance area at 8 pips counted from top CZ. There are far better places for a Risk Reward 1:1 (or higher RR) trade.

If I look at the AUDUSD SSBO trade today (forget my drawings), that is one I may have taken in SSBO style as teached by Forexhard. If I put my SL above the upper yellow dotted line and my take profit at the lowest yellow dotted line, I have in my opinion a great change to make profit out of that 1:2 trade. I don’t care if I lose a trade like that because I know that it is a good trade, even if I lose that particular one. Next time better.

What I’m trying to say: I think it is not in the settings. I think it is on the chart. I think it is not in finding the best settings but in filtering CZ’s and trades. Because CZ’s appear in favorable and unfavorable situations, regardless what the settings are.

What was Forexhard trying to do in the thread below from september 2011 until April 2012, after creating this thread? Did Forexhard think that Stairsteps Breakout could be improved with a kind of heatmap S/R line? On page 108 of this thread he has the indicator from the thread below on his charts.

http://www.forexfactory.com/showthread.php?t=312952

I think he thought exactly that, and I think he was right. Last two pictures from today.
Attached Image(s) (click to enlarge)
Click to Enlarge Name: CADJPYM15.png Size: 63 KB
Click to Enlarge Name: AUDUSDfib.png Size: 58 KB
Click to Enlarge Name: EURJPYM15.png Size: 59 KB
Click to Enlarge Name: EURUSDfib.png Size: 52 KB
  • #7,439
  • Sep 10, 2014 1:56pm Sep 10, 2014 1:56pm
Quoting GU-Night
{quote} I had no problems getting her (Rossbot) running. I don’t like robots, normally. Maybe because I have almost none experience with them. That is what I’m going to find out next months. It is more that I want to compare my manual retrace results, with her mechanical “in at breakout” SSBO results. I don’t know at which stage you are right now, but I think that the TPO indicator can help to make better trading decisions too while trading SSBO as teached in this thread by Forexhard. If I take a look at the CADJPY picture below (extreme...

Interesting stuff GU, thanks for sharing the other thread by FXH too, it does reinforce what were discussing More layers to uncover!

This week has been very slow for my settings, only 1 or 2 Czs formed when normally Id have 4~5 by now... (similar last week too). If this level of volume continues I'll need to start looking into rotating between my other settings I have running. Its a bit early days yet, gives me time to play with the above ideas
  • #7,440
  • Sep 12, 2014 1:36am Sep 12, 2014 1:36am
Hi GU. Also a short thanks to the new ideas and the link to the other thread. :-)
3 traders viewing now