I started with testing what would be better, SingleSideComplexOffsetting 'ON' or 'OFF'. I used EURAUD for this, the results are from the last 4 weeks. Magic #0 is 'ON' and magic #1 is 'OFF'. 'OFF' generates appr. 50% more profit than 'ON';
Since GBPJPY is still the MajorProviderOfPips, I started a new test today on this pair. SSCO 'off' and 3 charts;
- 5 pips grid
- 10 pips grid
- 20 pips grid
I will report on this next week.
In the mean time, I'm running RnR live and love it. I woke up this morning to nearly 1% growth, and this on a UK bankholiday......
.
Gertje,
Are you referring to Single Complex Sell Offset/Single Complex Buy Offset when you say SSCO? Checked the code and there's nothing such as SingleSideComplexOffsetting.
Are you referring to Single Complex Sell Offset/Single Complex Buy Offset when you say SSCO? Checked the code and there's nothing such as SingleSideComplexOffsetting.
Regards,
Haggadah
Haggadah,
If you have followed the whole thread, you would have noticed a test between the CSSO 'on' and 'off', which ended in favour of 'off'.
I don't know about the code, but I do now about this specific setting, which is this one;
image.png
.
You do not have the required permissions to view the files attached to this post.
TraderJoeForex » Fri Jul 29, 2016 1:35 pm wrote:
Hi Gertje, I've been looking at your FXBlue stats.. is it possible to see the max drawdown for each of the magic# ? I couldn't see a way to break this down...
It's not obvious to me either! The best way I've found is by looking at the balance and equity charts for each magic. You have to filter the magic, then select stats and then charts. You then need to pick "banked and floating P&L (hourly)" chart from the drop down.
Here's magic 1 for example. Hope that helps.
You do not have the required permissions to view the files attached to this post.
Bruster400 » Fri Jul 29, 2016 12:45 pm wrote:
It's not obvious to me either! The best way I've found is by looking at the balance and equity charts for each magic. You have to filter the magic, then select stats and then charts. You then need to pick "banked and floating P&L (hourly)" chart from the drop down.
I've been looking at your FXBlue stats.. is it possible to see the max drawdown for each of the magic# ? I couldn't see a way to break this down...
Cheers and hope you are well.
Hi TJF,
I'm doing fine, thank you. Just returned home from a 3 week holiday in Austria.
I'll post the "investor password" for the account as soon as my Empty4 releases it to me....
Last week was a nice test with lots of volatility.
Results are;
#0; plain vanilla
$962,21 profit
918 trades
Max DD $1.069,07
Current DD $255,27
#1 M15 MA
$395,74 profit
294 trades
Max DD $514,45
Current DD $295,68
#2 M5 MA
$533,77 profit
449 trades
Max DD 565,81
Current DD 91,58
All in all it looks like the EXP version using the MA can make a difference regarding DD.
My live account however, running #1 is in $2.035,51 DD as I'm writing this. Profit last week was $394,14
No idea why the huge difference in DD, didn't had a chance to look into it yet.
I ran 3 RnR Exp demo's running at GP over the last 2 weeks. Settings based on a TJF setfile with: MinOpenTradesToStartOffset=4, OffsetPipProfitStandard=15, UseRecoveryMode=true, TradeImbalanceForRecovery=5.0, TradeImbalanceToReset=3.0, HedgedBasketPips=15, ClearBasketPips=15
Results:
Magic, Closed profit, Open DD, Max DD, #Trades
Magic14, 620, -358, - 487, 1002, MA simple typical 10, CSSO off
Magic15, 729, -284, -1293, 1002, MA Simple typical 10, CSSO on
Magic16, 1257, -199, -621, 1995, no MA used, CSSO off
Conclusions so far: - My MA settings (MA Simple, Typical, 10) on magic 14/15 resulted in worse results compared not using the MA in Magic 16. For me, with these settings, trying to give directional bias did not help.
#trades was cut in half, but so the results and during the weeks DD was worse than without using MA.
-CSSO 'on/off' was interesting. CSSO did very bad in week 1 with results of magic15 at 50% compared to magic 14. Also in that week CSSO 'on' suffered a 1293 max DD. However in week 2 CSSO 'on' made up all ground and more.
I think the consensus is to leave CSSO 'off', but I will test in longer running trial anyway. IMO you need at least a couple of months forward testing to draw valid conclusions.
- Magic 16 without using its MA did best over the 2 weeks. Would love to do this on real account.
But looking at Gertje's remark in the previous post, I better wait for his analysis why his live account is suddenly in such a DD.
I love this bot, but still need a bit more optimizing:
- The 5 pip profit per trade is very low. Is does not take much to erode this away. Lets hope a real account will give better fills/spreads than Demo.
- I remember TJF and Gertje were at 7 cents per trade with 5 pip grids, but will have to reread if that was with 'original' RnR, and what their settings were. And of course not every week is the same.
- Will need testing to find how much can you increase OffsetPipProfitStandard, OffsetPipProfitRecovery, HedgedBasketPips, and ClearBasketPips before hurting the 'Winning trade percentage' too much.
- The larger grid experiment of Gertje is also interesting ... much higher profit per trade trade (>10 pips with 10 grid). A larger grid approach could be interesting for Cent accounts with high spreads for the the less capitalized among us.
During my tests I encountered several instances where RnR got itself stuck in an endless loop due to an 'invalid ticket'. I understand that it is just RnR trying to close a ticket that already was closed, but that communication about it got lost or not fast enough to react.
Anyway, RnR freezes, but keep sending warnings about the 'invalid ticket' until it is removed from the chart.
I am not good enough with code to find why it gets into a loop and just stops. I see there is a MopUpTradeClosureFailure but that apparently does not address the issue above.
I have been trialing Demo's at different brokers, and tried different terminal versions but so far I am not sure yet if there is a full-proof combination ...or if there ever will be with Empty4.
Interestingly different brokers (all ECN Demo) gave different problem rates:
- Broker1 Demo - 2 loop incidents (in 2 weeks with 6 RnRs). Not bad, but what if this happens live on a real account which I am about to start? If you don't see that the EA is stuck you can potentially lose a lot of money.
- Broker2 Demo - 100% failure in 16 tries. Could not run any RnR longer than about 4-9 hours before it looped (wanted to try broker for their cent account availability).
- Broker3 Demo - just couple of days, but so far so good.
What is your experience? And how did real/live compare to Demo?
Or is because my ping to the servers at 350ms is too slow?
Edit: Moved my question about Master EA for warning about stuck EA to the coding section
Last edited by WhoKnows on Tue Aug 02, 2016 10:36 am, edited 1 time in total.
WhoKnows » Sun Jul 31, 2016 12:16 am wrote:I ran 3 RnR Exp demo's running at GP over the last 2 weeks. Settings based on a TJF setfile with: MinOpenTradesToStartOffset=4, OffsetPipProfitStandard=15, UseRecoveryMode=true, TradeImbalanceForRecovery=5.0, TradeImbalanceToReset=3.0, HedgedBasketPips=15, ClearBasketPips=15
Results:
Magic, Closed profit, Open DD, Max DD, #Trades
Magic14, 620, -358, - 487, 1002, MA simple typical 10, CSSO off
Magic15, 729, -284, -1293, 1002, MA Simple typical 10, CSSO on
Magic16, 1257, -199, -621, 1995, no MA used, CSSO off
Conclusions so far: - My MA settings (MA Simple, Typical, 10) on magic 14/15 resulted in worse results compared not using the MA in Magic 16. For me, with these settings, trying to give directional bias did not help.
#trades was cut in half, but so the results and during the weeks DD was worse than without using MA.
-CSSO 'on/off' was interesting. CSSO did very bad in week 1 with results of magic15 at 50% compared to magic 14. Also in that week CSSO 'on' suffered a 1293 max DD. However in week 2 CSSO 'on' made up all ground and more.
I think the consensus is to leave CSSO 'off', but I will test in longer running trial anyway. IMO you need at least a couple of months forward testing to draw valid conclusions.
- Magic 16 without using its MA did best over the 2 weeks. Would love to do this on real account.
But looking at Gertje's remark in the previous post, I better wait for his analysis why his live account is suddenly in such a DD.
Interesting results. I think the MA filter may work best on it's own i.e. with recovery mode OFF so not using the recovery set file I posted as there may be a conflict there between all the filters.
Could not run any RnR longer than about 4-9 hours before it looped (wanted to try broker for their cent account availability).