The Forex Robot World Cup

[raa]

Open Trades:

00:00buy0.50eurchf1.46341.45861.4834 1.46490.000.000.0069.40 20100105s5 [LMD_Multicurrency] 5xC EA405500562010.02.23
12:25buy0.50eurusd1.36131.34031.3625 1.35560.000.000.00-285.00 20100104s2 [HiRIDER] 5xC EA405500732010.02.23
12:25buy0.50eurusd1.36151.34051.3627 1.35560.000.000.00-295.00 20100104s2 [HiRIDER] 5xC EA405500982010.02.23
12:25buy0.50eurusd1.36151.34051.3627 1.35560.000.000.00-295.00 20100104s2 [HiRIDER] 5xC EA405501772010.02.23
12:27buy0.50eurusd1.36151.34051.3627 1.35560.000.000.00-295.00 20100104s2 [HiRIDER] 5xC EA405501912010.02.23
12:27buy0.50eurusd1.36141.34041.3626 1.35560.000.000.00-290.00 20100104s2 [HiRIDER] 5xC EA405502042010.02.23
12:27buy0.50eurusd1.36141.34041.3626 1.35560.000.000.00-290.00 20100104s2 [HiRIDER] 5xC EA405502142010.02.23
12:27buy0.50eurusd1.36151.34051.3627 1.35560.000.000.00-295.00 20100104s2 [HiRIDER] 5xC EA405502232010.02.23
12:27buy0.50eurusd1.36141.34041.3626 1.35560.000.000.00-290.00 20100104s2 [HiRIDER] 5xC EA405502422010.02.23
12:27buy0.50eurusd1.36131.34031.3625 1.35560.000.000.00-285.00 20100104s2 [HiRIDER] 5xC EA405502612010.02.23
12:27buy0.50eurusd1.36121.34021.3624 1.35560.000.000.00-280.00 20100104s2 [HiRIDER] 5xC EA404810052010.02.19
01:50buy0.50gbpjpy141.980.000.00 139.120.000.00-1.75-1 585.02 20100101[SuperVolcano] 5xC EA404812772010.02.19
01:52buy0.50gbpjpy141.870.000.00 139.120.000.00-1.75-1 524.05 20100101[SuperVolcano] 5xC EA404813482010.02.19
01:53buy0.50gbpjpy141.880.000.00 139.120.000.00-1.75-1 529.60 20100101[SuperVolcano] 5xC EA405332062010.02.23
00:00sell0.50usdjpy91.1491.6988.84 90.240.000.000.00498.67 20100105s1 [LMD_Multicurrency] 5xC EA 0.000.00-5.25
-6 970.60

Floating P/L:-6 975.85


I have a lot to say about Fusion from my observations and I have orginally posted it, but decided not to be in haste and wait a bit. But in short: Fusion has no risk management in regards to opened positions and stop-losses used, or at least I did not see it (as they have recommended not to use money management feature) working. And it is very hard to judge it manually as you can see some of the positions now running close to 300 pips of loss!!!!!! I would not be surprised to see no stop loss there what so ever, even internal one. I.e. your equity will keep going down until you actually will get a margin call and be wiped out.

If we look at the lot size set as per new "conservative" recommendations of FRWC, i.e. 0.02, we are risking maximum simultaneous loss or floating equity draw down to 300 pips on up to 10 positions 0.02 lot size each, i.e. 600 USD, which is a 60% of account it is recommended for. Judge that for yourself. The figures from my account above are self-explanatory.

To my opinion to the moment, even if my Fusion account will get recovered, I still would consider it a VERY DANGEROUS EA, as it lacking control on positions, integrity and risk management. Sticking to a single EAs like HiRider, LMD might be a better choice as you would have lots more control over it. SuperVulcano to my opinion (and it will not change whatever huge wins it may or will deliver) is a VERY DANGEROUS thing. I simply would never accept an EA running me into 300 pips of loss without taking a loss much earlier. It is a hell roller-coaster. Once these SV positions are closed whatever they will be I am shooting SV down. It did not pass a test.

[raa]

Continuing to back test robots while Fusion is posting dreadful results on live and test accounts... More I back test HiRider more I like it, actually MM feature works in HiRider, when it does not in Fusion (even when only HiRider is selected, it still uses default lot sizes, weirdo), however strategy 1 in HiRider is always set to 0.01 lot, which may have some sence for 1000 account they have been using in competition, but even for 5-10K it makes no sense... With fixed to 0.1 lot and 10 lots for 2009 it turns 10,000 into somewhat 25,000 with a gentle curve, but it has a better potential, as as account is growing returns from fixed size lot are diminishing. In real life it is not a problem though, as account is growing not in seconds but in weeks and month and there is a plenty of time to adjust it to current sized account, though. So, it makes more sense to use sole HiRider robot for a back test (MM works, although Strategy 1 uses always 0.01 lots), for real trading it could make more sense to use HiRider embedded into Fusion (having all other EAs there switched off), as there you can actually change a size of Strategy 1 lot. MM function won't work but as I mentioned before in real-time trading it is not a big issue to change the lot size once in a while. Hopefully FRWC people will address those bugs/issues. I will collect my early remars by the end of the week and will write them some suggestions.

[raa]

You know guys, I have spent a lot of time yesterday back testing HiRider and Fusion with only Hirider enabled (it is the same only in Fusion MM feature is not working at all, when enabled all trades are set to 0.01 no matter what account is). HiRider shows impossibly perfect balance curve for the year 2009 until november, it actually grew 10,000 into 12,000,000. Impossible, you say. And I agree. In November and December HiRider posts two huge falls, and January and February this year is the same. What a f...k? Oh yes, smart developer heavily optimized it for the last year (-s) data, till November 2009 and we can see it. Signals, stop losses, all is optimized for the "perfect wave". What does it mean for me? Well, it means two things:
1. This is not what I (and I believe we) expected and is very far from "all market conditions weapon"
2. I have demanded a refund from Plimus, through whom FRWC does sales. We will see if the refund is working.
3. Santa does not exists

Back to the hard way of growing account.

[capsmart]

I wish you all the best with the refund.
Thank you for the hard work you did informing us about the robot.
Take care

[dannyd]

Gentlemen

Daily update:- Fusion took 14 positons yesterday 2 buy and 2 sell on euro/gbp, scalaping trades and all four won (£6 per trade) and 10 buy trades on euro/usd!! each of these positions had a 220 S/L and has gone from being some £1100 down to £600 down. Interesting to see when it decides to get out. Lousy risk reward ration as t/p level set at 12 pips above entry.

Running position therefor

Starting Balance- £5000
Current balance - £6170
Equity - £5552

As a point of note my Metatrader crashed during the night although limited moves mean it should not have effected the trading positions.

RAA, How can you tell which robot is transacting the trades? In addition I opened a new profile on Metatrader with the charts requested for the robot but during the day I have a differant profile open running Marc's 'My chart setup'. Would this effect the robots effectiveness at all?

Dannyd

[raa]

Dannyd, right-click on terminal log and select comments. In comments it is described who's trading.

While waiting on a reply from FRWC (I suspect they would tell me to send a 30-days report on Fusion activity), I continue to back test HiRider. I have found out a weird thing that results based on M5 chart (as recommended for HiRider traded solely, and since M1 are unavailable for back testing) are so very different from results based on H1 charts. Generally H1 results are good even for this years Jan and Feb, while M5 results are devastatingly poor. I am wondering what difference it makes, could be robot working out of candles' closing times rather from tick values? Looking for the safe sensible values, the standard values by default (10 trades with 5% margin position size) are for the quick trip to get wiped out on data from Nov2009 to Feb2010 (those values give fantastic results for Jan2009-Nov2009, I believe robots is heavily optimized for that period).

[raa]

And an another proof of HiRider's optimization for 2009: I run a test from 2005 to 2010. From 2005 to the end of 2008 there is no profit - base amount of 1000 goes up and down but there is no winning curve, by the beginning of 2009 it is about 2000 usd. Yet starting with 2000 usd by Nov2009 it turns it to nearly one million dollars, then loosed 40% of it mid-Nov, winning everything back by the end of the 2009. The curve for Jan and Feb 2010 is nothing like a curve for any month from Jan to Nov 2009.

After running dozens of simulations I can recommend these "safe" settings for HiRider, that during period of 2005-2009 and first two months of 2010 do not generate draw down more than 40% (which most I decided to tolerate), yet generate profit (around 30% from Jan-Feb 2010):

MaxTrades = 5
UseMM=true
LotriskReductor=4
EquityLimit=600 (when starting with 1000 account, i.e. 60% of initial deposit)

These are settings for separate HiRider, not part of the Fusion (MM does not work in Fusion, and changing of basic lot for HiRider Strategy 1 also does not work, I have earlier said, mistakinly, that it does work, it doesn't, despite Fusion having a setting for it). Also a general setting in Fusion for max number of orders does not work either, I had there 10, despite that my live account right now having 15 open positions, 14 of those loosing. I.e. risk management with Fusion is nearly impossible.

[dannyd]

Cheers RAA.

Only one new trade conducted today on HiRider, a scalp looking for 4 pips! The old 8 trades instigated by HiRider are still running with a current deficit of £600. This despite being only one pip from it's original T/p at 15.30hrs today. Not impressed.

End of week 1

Fusion has now been running for exactly a week and balances are as follows:-

Starting Balance - £5000
Current Balance - £6194
Open trades - £570 negative

Current net position £5624.

So net position up just over 11% but if HiRider does not get out of these euro/usd trades shortly that could all go.

Re settings

All as per recomended in manual except the following

Lots left at 0.01 rather than 0.01 per £1000
Lots Risk reductor set to 2.5 v 1.5 rec
Max lots set at 10

[raa]

I have been back testing HiRider versus MegaDroid and guess what, MegaDroid posted at least twice of a prifit comparing to HiRider both on Jan2009-Feb2010 testing and Nov2009-Feb2010, as well as Jan2010-Feb2010. Settings for both set in a way to have a similar maximum draw down during the testing period.

===========Nov09-Feb10=================
START=5000
11600===Megadroid: Aggr=true Recov=true Risk=0.5
08045===HiRider: MM=true MaxTrades=7 LRR=4.3 Chart H1 (one huge loss)
06227===HiRider: MM=true MaxTrades=7 LRR=4.3 Chart M5 (two huge losses, barely recovered)

I am now trying to live test HR using H1 chart, it gave me a warning message, but seems to be working, however HR with the same setup but different price feed on M5 chart has opened one trade, on H1 chart it did not. So maybe it won't work on H1 at all. Anyways, after all that testing I honestly don't see a big value in HR. Certainly less than MD.
Hard to say anything about other robots, but my Fusion setup on a live account is currently running about -20% floating loss with recommended settings and it does not obey settings regarding max number of positions (15 open when 10 set as max), and any risk management is nearly impossible. Waiting for a refund.

[raa]

Hi, got a reply from FRWC:

I am sorry to hear that. We truly want to help you with FRWC trading. First of all let me explain the Money Back Policy: Trade ANY of the FRWC's Royal Trader robots on a DEMO account (they are FREE to open with any FX broker) or live account, as you wish, for 30 days. If you are not happy with the results, NO MATTER what they are, send us your account statement within 60 days of purchasing and we will give you a FULL refund - no questions asked. We do not want your money if you are not happy with the product! Please, however, bear in mind that upon receiving a refund on your purchase, you will no longer be able to use any of our products If you have technical problems with the software (i.e. installations issues, activations troubles) please let us know and we will be more than happy to assign a personal manager that will assist you!

As expected, here is my reply to that:

Hello, I understand why you are asking for a month testing, but here is the main reason why I don't want this product: this is not because it had lost something right away, this is because any of your robots are not yet in a stage of a commercial product. I was interested in Fusion, however, it is simply does not work as described: there is a setting for Max number of concurrent trades, it does not work, I have a 10 set up there and my Fusion account has 15 opened trades currently, does it work? No. There is a setting for a money management feature, if switched on all trades are set to 0.01, on account of 50,000. In Fusion for HiRider there is a setting to change a size of a minimum lot for strategy 1, no matter what is set there it still uses 0.01 lots, which is totally senseless on any account bigger than say 5000. So it does not work too. It is simply impossible to have any risk management with this product. It might have some clever ideas regarding trading, but you guys did not develop it yet to the stage where people should pay for it. I will simply not trade with any of these robots just because it is nearly impossible to justify risks involved and possible losses. I have set up Fusion on real account according to your basic recommendations (0.01 lot per each 1000), which is even less that your "conservative" settings which double that, and because robot has opened too many trades and SuperVulcano has positions running into somewhat 300 pips of loss , the floating loss is more than 20%. How conservative is that?

This is the reason that no matter what results it will produce during the month I don't want to use this product. You need to work more on this, guys, to bring it to commercial level. And then I (and others) may come back. But for now I want my money back and now.

Thank you.