GridDancer / GridMaster

[renexxxx]

Is it possible to stop it from adding trades to steps that already have open positions? (The new additional trades were a few pips off from the existing open trades as the ATR had widened so the step gap had increased).

GridDancer does this already. If/when a new grid level is reached, it checks the most recent HGI-signal. If HGI says 'BUY', it will check if there is already an open BUY at this (new) level. If so, it will exit silently. Similarly, if HGI says 'SELL'. How does it 'know' if there is already a trade at this level? It looks at the trade comment, particularly the bit between the []-brackets.

Whilst GridDancer has an open sequence of trades, the grid size (or distance between trades) will remain constant, even the ATR changes dramatically during this time. The grid size is even persisted into a GlobalVariable, so even if you restart your Empty4 platform and/or change any of the parameters, it will only modify the grid size if its stack of trades is empty (i.e. it is about to start a new sequence).

So, I don't know what happened on your account, but this is not designed to happen.

[Timmy]

Hi Rene,

Thanks for your reply. I understand all that you have mentioned here. However I can assure you that I had a double up on positions (unfortunately the screen save that I did got lost).

Its not a big deal but can you comment on what I suspect has happened.

So all I can explain for this is - as the Equity TP was hit ($50) during the spike in GBPUSD and the automatic rinse n repeat was triggered, then the trades began to close when the GM activated (MasterTimerSeconds = 10), and the grid step Global Variable was cleared. At that point the spike pulled back and so the Dancers assumed they could set new trades as they saw no trades open in the two steps that it doubled up on.... because the GM had cleared the GlobalVariable for the grid step and reset a new one.
Could this be the case....?

Anyways, moving on - I wish to discuss risk management with GD/GM. And get a discussion going on this super important topic.

My strategy will centre around risk management. For me it is all about limiting drawdown through limiting overall risk. So firstly I am setting a limit of 1:5 on my leverage (not the brokers leverage). I will be using a demo test account of $15k as this will be the size of my initial live trial account if demo testing shows promise.

Hence my maximal risk at 1:5 leverage will be 5 * $15k = $75,000 of open contracts at any one time. Or 0.75 of a full lot contract.

So if I trade positions of 0.02 lots per trade then this will allow me to open 37 positions (0.75/0.02) or 75 positions at 0.01 trade size.

So now the question becomes how do I limit the Dancers to operating within this defined level of open positions. There are several ways we can do this:
1/ By setting a specific Equity TP target - I have gone with $50 per cycle
2/ By setting the DistanceBetweenTradesATRratio - I have used 0.2
3/ Limit the number of instruments being traded - I trade 4 pairs (EU, UJ, GU, AU)

(I have not considered MinProfit or Basket Profit yet. And I will not be using SL's)

At $50 per cycle I will be making 0.33% per cycle (50/15000) and as I am targeting approx 5 to 6% per month then that is on target for the account size if I am able to close one cycle per day. On my second day of testing it closed two cycles for 0.66% so I may have to dumb down the profit (as silly as that sounds - but it is important to get the lowest risk profile for my on-target return).

I have used 0.2 ATRratio so this will roughly equate to 5 steps per 1 days range. So this means I will get about 20 positions open if all four pairs max out their range for a day.

My logic here is that if the days range is met for the four major pairs then at least one cycle should be able to completed. In fact they only need to each move 50 pips or approximately half a days long term range to open two steps and close out a cycle of $50 TP. This is why I am going with 0.02 per trade and four pairs.

Now if the markets reverse and the trade positions move into DD then what do i have to play with to get me out of the crap... We have established that I have 37 trades to do this. Will it be enough? If we assume that across all four instruments the average step is 15 pips and the trades are spread evenly over each pair as the drawdown progresses (of course this may not happen) then we have ~9 trades per pair or around 135 pips movement to force the 9 contracts in play for each pair.

This is my starting point. So now I need to understand how the GM will close out individual losing positions around the MinProfit & Basket Profit processes - and how this will effect or be effected by risk.

Comments please guys.

Tim

[acostafulano]

René, a question Re ATR Grid Size.

Once the EA is started and the grid size goes into the Global Variables, does it remain that way ETERNALLY until you manually delete the GV+Restart the EA? Or is the grid size re-computed (And the GV updated) with every new EquityTP/SL closure? The latter would make more sense to me from a trading perspective, but if this is not the case, it'd be good to know so that I can re-set the GV+EA every week or so.

Thanks for your comments on this

Diego

[murysco]

Secondly, we can alter GridMaster so that it does look at HGI-signals and ignores BasketTP and EquitySL ... but this is a totally different concept. It would make more sense (if we wanted to do this), to (1) get rid of GridMaster and (2) alter GridDancer to close its own (SingleSymbolMultiTrade) basket if/when the HGI-signal reverses. GridDancer would be a standalone single-symbol EA, much like one of the HGB EAs Steve has created.

Don't get me wrong: I don't want to discourage people from posting suggestions for improvements to this combo. Indeed, I am constantly looking to curb the inherent drawdown we are getting with this strategy, as I believe this has so much potential. But I don't think we should walk into avenues that have been walked in before.

Thank you for your kind reply.

I just thought it is worth giving a try but if you think otherwise, then no problem. This version is good as it is now.

Cheers.
MM

[renexxxx]

So all I can explain for this is - as the Equity TP was hit ($50) during the spike in GBPUSD and the automatic rinse n repeat was triggered, then the trades began to close when the GM activated (MasterTimerSeconds = 10), and the grid step Global Variable was cleared. At that point the spike pulled back and so the Dancers assumed they could set new trades as they saw no trades open in the two steps that it doubled up on.... because the GM had cleared the GlobalVariable for the grid step and reset a new one.
Could this be the case....?

Not exactly, but you are describing a scenario that is undesirable and should be fixed (i.e. whilst GM is closing trades, one or more of the GD's are opening new ones) This is one of the side-effects of multi-threading and we need a mutex to stop the GDs from trading whilst the master is at work. I will add this to the next version.

This is my starting point. So now I need to understand how the GM will close out individual losing positions around the MinProfit & Basket Profit processes - and how this will effect or be effected by risk.

You can refer to this post, where I tried to explain this.

Comments please guys.

I like the way you're thinking: You are treating this thing as a business and try to estimate your highest business expenses. I think that there will be times that your calculated 37 trades won't be enough to get you out of the crap -- but I can't tell you how many would be enough ... On my demo (that has done extremely well this week), the margin level was at one point as low as 600% ... this is too low for my liking. (This demo account has a 1:200 leverage). I wish that I could tell you that in the worst case the account still survives, but I can't.

[renexxxx]

Once the EA is started and the grid size goes into the Global Variables, does it remain that way ETERNALLY until you manually delete the GV+Restart the EA? Or is the grid size re-computed (And the GV updated) with every new EquityTP/SL closure? The latter would make more sense to me from a trading perspective, but if this is not the case, it'd be good to know so that I can re-set the GV+EA every week or so.

Yes, the latter is the case. As and when all trades of a sequence are closed, the GridDancer's calculate a new ATR-based grid size (if it was ATR-based), and store this new value in a GlobalVariable. For as long as trades of this sequence remain open, it will use that grid size (to avoid having a stack of trades open with variable distances between them).

Cheers.

[renexxxx]

Here are my results of my demo account, as updated until now:

Capture.JPG

As you can see, as soon as it got out of 'the crap' it started raking in profits again. There was hardly any noticeable effect of NFP. I will keep this going during the US Elections ... would be fun to watch.

[kwchau]

Here are my testing results comparison. I started both Dotty and Dancer both on Oct 27 with the same 18 pairs.
I like what dancer has performed in this period. The return is smaller but the DD is also much smaller, make me feel more comfortable if there is a black swan events emerged.

The first chart is Dancer and the second is Dotty. At the same time, I also started RnR with less fewer pairs, which is attached in the third picture.

[trader689]

i've had great results over the last 2 weeks, running the recommended pairs pretty much ootb:

hgidancer.jpg

[acostafulano]

Hi everyone, here are my initial results and myfxbook links of the HGI GM I'm running

I tried a bunch of other stuff but killed those demos early on.

I may be adding new setups soon!

https://www.myfxbook.com/members/acosta ... 15/1840248
https://www.myfxbook.com/members/acosta ... 15/1840251
https://www.myfxbook.com/members/acosta ... -3/1847536
https://www.myfxbook.com/members/acosta ... -3/1847537

Basically Using TREND and RAD only.

EQUITY TP-RINSE TP-SL in the demos title.
It's either All pairs running or René's 7 Pairs (on the title as well!)

Thank you René for this wonderful EA's

Diego