T101 basket trading system - math analysis - page 16

New comment
 

I've tried BasketBull v8 over last week (or slightly more), yet I've noticed that somehow it creates problems. It was running on Alpari uk demo accounts, and, for some reason 3 accounts were disabled one by one. I mean, I opened one account ran BB on it, it got disabled for unknown reason after 2-3 days. So, I opened new one. Same story. And 3rd one. Same story. Of course I've been using other Alpari uk demo accounts on the same vps for long time and they are perfectly ok. So, I suspect that for some reason BB is doing some not nice operations.. wont try it again, I think.

As for the profits, it managed to open like 3-4 baskets over that time, they seemed to be profitable. I got at least 1 losing basket, but others were quite ok - though have no full data, as accounts got disabled and I do not know at which level the baskets would close..

 

at my end, both H4 and D1 settings are disasterous. The EA impresses me with its sharp entry timing and retarded holding losing positions for too long or always losing floating profit. seemingly V7 is more promising than V8 to me...

Nopik:
As for the profits, it managed to open like 3-4 baskets over that time, they seemed to be profitable. I got at least 1 losing basket, but others were quite ok - though have no full data, as accounts got disabled and I do not know at which level the baskets would close..
 

Ok, I've restarted BasketBull test, so far it oscillates around zero, so more time is needed.

In meantime, I've also revisited this system and started to think how to trade it. Previously, I quite quickly went off the original idea of jumping pair slots, anchor pairs etc., and it did not ended too well (though it was not disaster ). Anyway, basket built by BasketBull EA is simpler and more balanced, so when I looked at it, I started to think again about basic principles behind this system.

It seems, that I overlooked one, very important basic thing. Namely - if you open full basket, e.g. 14 longs or shorts (or just 8, as in case of BasketBull), when this basket will go in profit and when in loss? Previously I had very analytical approach - measured exposure, got info that the profit is related to eur/jpy. And it was all right. But looking at those details, I forgot about something important.

So now I revisited it and started to think more about those slots and jumping pairs. If you have 7 buys at bottom and 7 sells at top, and you open long basket (14 longs), when it will go in profit? Can it go into loss? If yes, when?

The answer is simple. You measure cumulative profit of 7 buys and also cumulative profit of 7 sells. Since in moment of opening long trade all 7 buys were in loss, your 14-long basket will get in profit when cumulative profit of 7 buys will go up. Simple as that, yet very important. You want your basket to be as balanced as possible, to oscillate. Without oscillations, you risk that 7+7 account will just polarize and never go back. Here is where BasketBull choice of 8 pairs helps, as the exposure is almost nonexistent, and polarization will occur only if usd and jpy will get weak together, while eur/gbp/nzd/aud all go strong.

Anyway, apparently, if you draw cumulative profit of buys and sells on one chart, (they should sum to zero, or almost zero), you can trade those lines directly, by opening 14 pairs in the same direction. Actually, original poster, Trader101, was using this approach, but I did not understood its simplicity until now.

So, lets take a look at this picture:

We are interested solely in the bottom Profit Calc indicator, and its green and red lines. If green is at bottom, our IA have 7 buys at bottom as well. So, we go long with all 14 pairs. 7 sells are in profit, but they are likely to go in loss, so our long on those pairs will get into profit in the same moment.

Basically, this created virtual instrument. Instead of trading eur/jpy directly, we trade green line - we can buy it, sell it, and profit 1:1 from its movements.

Why this is important? Because this line movements is much smoother, much more predictable, and you can split it into 14 sub-lines to get even more insight what is going inside. You just have more information, something you dont have with eur/jpy. And, you have 'zero' level, so you can really put odds on your side.

I started to look closely at the shape of those lines. It seems, that they are trending very nicely, with higher highs and higher lows. So, if I 'buy' green line (by opening 14 pairs long), I put SL below last low of that line. When the line gets into new high, i move SL to more recent low. Something, that most manual traders are doing, I believe, just with our virtual instrument, not real pair. What are our profits? On the right hand we see that the line reach values like -1279 to +1089. So, our profit and potential loss is basically double of this (because we bought green line and in the same time we sold red one - so we catch each move with twice bigger volume). So, if you would

looking at the line, starting with the exact bottom around 26 May 20:00, you would get uptrend confirmation, for example, around 27 May 5:00, as by that time there were 2 higher lows and price is above highest high. So, entering long at that moment and moving SL successively up and up, would finally yield about 2000 pips (here we have 1 pip = 1 usd) by 28 May. And still with long position with tight SL, allowing to get even more profits.

This week I tried to do that. So far it seems, that it is harder than it looks like I've misinterpreted some movements and lost like 20% of the testing account, in 3-4 basket trades. Yet, I'm now sure that with more practice, my accuracy will grow. By that time I'll also see if my theory meet practice

 

One more thing. This 14-pair basket is very strongly correlated with eur/jpy, what is also visible at the previous picture. Though, profit calc lines seem to be smoother and more trending than eur/jpy (and they have 0 level!). Is it more profitable to trade green line, or eur/jpy? I dont know, though I believe in the former.

Anyway, later on I will also try 8-pair basket as introduced in BasketBull, since it is much more balanced and I expect that it will oscillate much better.

 

Hi Nopik,

Is it possible you could attach a copy of profit_calc indicator , I cant get the one I have to work.

And thank you for a great thread.

 
Nopik:
I've got mine from original thread: Simple Trading Method with trader101 - Page 155 @ Forex Factory, I think it was this post: Forex Factory - View Single Post - Simple Trading Method with trader101

Thanks for that, I am still having problems getting it to load, ut I will keep at it.

 

Basketbull 8.3

Do you know where can i download basketbull 8.3 . I cant find the link in the russian forum.

Boki

 

Pair Suffix

My broker using this 'm' suffix for the pairs. I've change the expert parameter to include this suffix but still not work. one thing appear on my chart is 'Run' (it set up black blank), but for like one trading day, no trade placed. any advice?

echelon4x:
Hi Guys,

I was conscious of having "hijacked" Nopik's thread to a degree.

I think that the mods. can safely split the thread into two from the point of my first post - the latter half of the thread being transferred to "BasketBull - T101 System EA" or something similar.

Are you aware that Version 8 of basketbull has been released with a few tweaks and this time the indicators run from the Daily timeframe ? I'm now running this version in myfxbook stats., however in the changeover, v7 entered some trades which I manually closed at a loss before installing v8.

I've attached the English version.
 

Hello Andy,

We would love the opportunity to make your indicator part of our system. I'll appreciate an update on the latest fixes.

Thank you,

Cal

hopfi2k:
@thor

Yes, new version is ready, but still doing some final testing.

- Logging o trend changes to files is done

- showing the time passed since last trend change is done

- small bug fixes adjusted

I'll release the new version soon.

Cheers,

Andy

P.S. It would be possible to show daily date, but I don't think it will be a real benefit. If you need different time frames shown at a glance, I highly recommend using Orest T101 indicator...
1...9101112131415161718
New comment
Reason: