Slopey Peaky Bob

[SteveHopwood]

live trading is the only way to test both a strategy and our appetite for risk
[/quote]

I always thought exactly this.


[/quote]
I trade 0.01 lots per dollop of $4,000....
[/quote]

Yes, I know. I planned to use 4000 for larger balances. It would be tough on 5k...
What could be a reasonable safe minimum? (considering that I would obviously not be happy to be margin called but it would certainly not be the end of the world. The only real pain would be the unsuccess of the system).
I know it is the mother of all questions but hey, better than asking lame stuff no? [/quote]


I am running a number of demos testing different settings for the basket take profit. I have these set to trade 0.01 lots per $1,000 and started them at $500 deposit - read back through the thread if you want to know why.

None of the accounts have blown yet but the margin level frequently drops below 500%. I am guessing that a decent compromise would be 0.01 lots per $2,000 but it is purely guesswork.

[fbs1979]

Steve, I can assure I would'nt dare to write something on this thread before having it read entirely twice and more


For what it values I see 1000 as very aggressive/too much aggressive, 1500 aggressive, 2000 averagely safe, 3000 pretty safe, 4000 very safe, 5000 caveau (well, caveaux are sometimes robbed too...)

In heads up poker you can make a living by playin against less skilled players. Anyway to take advantage of your edge you need large numbers, so you need to have balance to sustain bad luck and consecutive/semiconsecutive losses even if you are the best player.
We had a simulator. You had to estimate your edge against your average opponent, i.e. 55% vs 45% (they are weaker players so on 100 match you win on average 55), and then set the bet, i.e. 100 dollars per match.

Then the simulator gave you the probability of going broke in function of the account balance.

For instance:

Money in play per match: 100 $
Edge: 10%

Acct. Balance /prob of failure, busto
300 $ / 99%
1000 $ / 85%
2000 $ / 60%
5000 $ / 30%
20.000$ / 1%
40.000$ / 0.01%

(numbers are wrong, just to let you understand)

I'd love a simulator like this but for Forex. I suspect I should look in the direction of Standard deviation & Sons but it seems something really difficult to obtain. In heads up poker is easy to simulate. It is like a coinflip with a coin that falls head 60% of the time. An EA and its decision are something way more sophisticated...

It is for this reason that the Coordinated Test Project is SO important.
What we can't simulate we need to observe live. When lot of data will be collected it will be possibile to solve a problem like this, not by simulation but by large sample size observation. So I really hope it is something that will be carried on in the future and I will try to contribute to it as much as i can.

[fbs1979]

Hi.

news trading with Dustin Pass software that was "auto-clicking" the but/sell button when the news come...

Yeah. I tried that also for quite a while. That was funny and exciting. But barely made some money.

Cheers
Thomas

Yes, that was reaaaally funny! But as you say, rarely significantly profitable...
Cheers

[hilavoku]

Short info from me.
I did setup a demo with Steve's set starting 2018-06-27.
Since then the account had four basket closures.

Screenshot_26.png

I contacted GlobalPrime yesterday to refund my real account and start running SPB with 500€...
The parameter that needs to be adjusted is the CashTakeProfitPerLot. 3000 is to high for a little 500€ account.

Thanks Steve and the other contributors for this EA...

[SteveHopwood]

I contacted GlobalPrime yesterday to refund my real account and start running SPB with 500€...
The parameter that needs to be adjusted is the CashTakeProfitPerLot. 3000 is to high for a little 500€ account.

No it isn't.

Give this some thought; it took me a while to get my thick skull around it.

You are trading 0.01 lots? It will take as much time to reach the basket TP on a 5,000€ account as it will on a 500€ one.

Gareth's account started to recover well as soon as I increased the TP on his account to the one in the setfile in post 1.

[hilavoku]

Thanks Steve for your thoughts.
I read the posts over DD and some were as high as $ 200 for mainly the default settings. This would be fine for the $ 500 account and if needed I will refund the account with a little more money ...
Exciting times ahead ...

[fbs1979]

Obiously the main variable to control risk is lotsize and thus dollops... Anyway I think it is important to consider also TP and gridsize.

TP:
At first it seems that global risk/exposure is not dependant on TP. Smaller TP means frequent but smaller closures, higer TP means rarer but richer closures. Risk should remain the same ...
In reality, given how the EA is designed, higer TP means eventually more open trades so more exposure if things turn wrong. Think about the case where things fall just before you almost reached an ambitious TP....

Grid size/distances:
They directly influence the velocity to reach TP. The closer they are, the faster everything goes, the more exposure you have. If things go in the wrong direction, having a narrow grid can easily lead to lots of negative positions, because even smaller retracements would activate new trades and then bring them down again. With lighter grid this would be much more difficult.

To put it in perspective, taking the Hilavoku 500$ account case, it is true that if lotsize is the same, the “velocity” to reach a 30TP is the same for a 500 and a 5000 account. But it also true that to reach that TP, SPB would need to open the same number of multiple positions, thus actually its exposure would be 10 times higher. Unless he widens up the gridsize by a factor of 10... At that point risk would be the same, but everything would go slower... or reduce the TP at 3 to have same risk and velocity but less returns...

(then we obvioulsy dont desire/expect the same level of risk from a 500 and a 5000 account, it was just to elaborate with a concrete example)

So, starting from the assumption that that the logic behind SPB works and thus give us a basic edge (otherwise everything is useless), it will be crucial to find the optimum combination between these variables to maximize results and avoid wrong combinations that could even lead a pontentially winning EA into a losing one. Once again, coordinated tests are and will be gold.

[fbs1979]

Thanks Steve for your thoughts.
I read the posts over DD and some were as high as $ 200 for mainly the default settings. This would be fine for the $ 500 account and if needed I will refund the account with a little more money ...
Exciting times ahead ...

I also supposed that the max drawdown experienced until now was 200$ on 0.01 lots. (20% on a 1000 account). But Steve recently told me that it was 500. I’m a bit confused on this.

Anyway, considering that we are still on a small small small sample size I suppose 200$ on 0.01 positions is kind of an average drawdown (users here experienced it various times in few months. Me too, after 2 or 3 weeks of testing...). So for 0.01 lotsize I expect easily 400/500 $ drawdown, rare 800/1000+. If experienced drawdown is already 500 instead of 200, then figures would be even higher.
But at the moment it is pure guessing.

[fbs1979]

Oh, and yes, I will try to keep my mouth shut now for a while...

[hilavoku]

We need all information we can get and thanks for thinking out loud fbs1979...