MegaDroid

[Batchboy]

Originally Posted by Maree43

Changed risk to .2 and left reduce lots to false as this is meant to greatly increase profitability as per the manual.

Maree,
I took a quick look at the back-test years on their website, looked no further back than 2008 to find a nearly 3 losers in a row moment. After you click 2008 scroll about 3/4 down to find, think it was something like 10/14. First loss is big as lots up to 18, but notice the reduced lots for next day's losses on 15th & 16th? (There is a couple winners before second day's loser) So authors testing with reduced lots on, thought that I didn't remember such being false for a reason, but then I'm not totally sure as not looking at manual as you may be, but it is obvious from these tests it's on. I'd hate to see results if setting at false!
jb

[pipdetector]

Originally Posted by Batchboy

How 'bout $500-to-$34K in half a year going safely with reduce-lots=true? (.5 risk) Have best of both worlds! The $500 account (purposely set first date of tester just before a double whamee of losers) gets down to around $185, but it puills out and heads to $34K by today. I dare not set to false as it would eat the account down to the margin , but I am at .5, not .2, so operating different.
jb

I backtested it with Risk = .5 with Reduce lots = False in the period starting from 2008.01.01 till today the account was annihilated in Oct. 2008.

With Risk = 0.2 & with Reduce lots = False it tripled in that period.
With Risk = 0.2 & with Reduce lots = True it made a bit more than with Reduce lots False

I am more for RiskLevel between 0.1 and 0.2.

[Batchboy]

Originally Posted by pipdetector

I backtested it with Risk = .5 with Reduce lots = False in the period starting from 2008.01.01 till today the account was annihilated in Oct. 2008.

With Risk = 0.2 & with Reduce lots = False it tripled in that period.
With Risk = 0.2 & with Reduce lots = True it made a bit more than with Reduce lots False

I am more for RiskLevel between 0.1 and 0.2.

pipdetector,
I don't see if you did .5 & true, did you? From what you say it doesn't seem you did this combo.
My 1min data on platform I'm testing on only goes back to 6/1/08, so 9 months is most I got with good resolution. You have 1min back-data up to 1-1-08?
jb

[Batchboy]

I just did even longer period back-test, back from 2/7/08, result = $59K from $500 and .5 risk with true for reduce lots, since 7/7/08 it's $44K.
I get to feeling like rogue user of an EA here , but honestly, it's simple math, if you increase proportionately the lot sizes (which are also proportionate to account size) then the same size (!!) SL digs AND (!!) the grade of climb during a storm of winners will all be bigger, so yes digging in deeper on the rare double/triple losers, you can especially can feel it when first starting with a certain balance, but you "feel it less" after account grows, thus what was before at say .2 a lighter down 'n up, here we have a heavier down 'n up, but all just a magnified picture of THE SAME THING (!!)............. so you ultimately end-up with the larger bank-roll in the end....... perhaps adage "no pain no gain" applies?
When I just did from 2/7/08 there was a real heavy bout into the beginning account that quadruple dug-in, couple of losers followed later by couple more, but it only digs in proportionate to what account there is left (!! thus like in "Deal or no Deal", a "safety net"), and it is only a magnification of what you would have if you did .2, it's only mathematically so, if one couldn't stomach it plus you need to go conservatively with what owners suggest, naturally don't do.
I think they have surpassed/excelled unknowingly even themselkves in what they achieved, and thus something like this COULD be discovered, most WON'T use it this way for above reasons, but hell, you'll never know unless you try! Anyone for $500 start capital on both this agressive mode AND conservatively agreeable mode? Why lose out on the possibility?
jb

[Batchboy]

Brought myself down from feeling high about having good test results on AlpariUS, so went for IBFX tester platform:
demo-tester, mini:
risk=.5, reduce lots = true
Note: I now test in two separate segments by changing the DST from 0 to -1 at DST change date of 3/8/09, I do this by placing in test both in EA and beginning balance the balance result from test up to 3/8, ok? Makes for some work but worth it to make sure of things:
------------------------------------------------------------------------------
$500 initial balance 11/21/08 > 3/8/09 gmtoffset=0 profit = $4100
(accidentally forgot to add that to the $500, so I only put $4100)
3/8 > 4/4 gmtoffset = -1 profit final = $8K,
so 16X in 3-1/2 months!
--------------------------------------------------------------------------------
on live$ mini IBFX platform tester from which I placed investor url in above post, it has live$ 1min back-data from date 2/16/09:
$500 initial balance, same settings as above tests:
2/16 > 3/8 gmtoffset= 0, 1 loser in this short period, profit = $144,
3/8 > 4/4 gmtoffset= -1, no losers, profit from (added total balance this time!) $644 = $668
Total balance = $1168
more than double the account in about a month 'n a half

Just wanted to make sure since I'll be putting my money on this in awhile that this is working on IBFX too, and it appears so, tester seen to replicate live$ also.
-----------------------------------------------------------------------------------
jb

[pipdetector]

I reviewed my results and it seems that the GMT offset that I used was wrong...

When I set it to GMT offset to 1, the backtest results started to behave.

Quite good. Even with your incredible 0.5 risk level...

[Batchboy]

Originally Posted by pipdetector

I reviewed my results and it seems that the GMT offset that I used was wrong...

When I set it to GMT offset to 1, the backtest results started to behave.

Quite good. Even with your incredible 0.5 risk level...

Alrighty then! Let's hope she continues to replicate how tester is showing! I just need to watch my $3.50 live$ running of it for a bit to verify all is a go and then I step to $100 as an interim on way to $500. I figure if those big losers come while doing these initial start-up levels then I simply wait for it to "supposedly" recover "like as in tests".
jb

Watching discussion with interest. Need to absorb it all for a bit!

I intend to start with $500 live when happy with settings and performance.

Keep up the good work testers - much appreciated.

[sergiel]

I started using MegaDroid on my Alpari-us demo account day after it was released and was starting to wonder why it wasn't making any trades. Any idea if they will upgrade to Build 222 soon?

Downloaded the UK version but wont perform a backtest for some reason. Is GMT setting on UK 1 as well?

Thanks all

Originally Posted by Batchboy

re: Heavens-to-Mergatroid........Eee-ven!

I had it on AlpariUS for a few days, no trades! That I discovered was because they are not updating their MT4 to new build-versions, 220 per manual says it's not sufficient. AlpariUS chat was not a nice one in this regard! I'm no longer migrating my server to NY just to get 1ms delay to connect with them!
Soon as I went to IBFX build 223 a trade was opening. Even better is down to the minute ("tick") the tester was replicating what happened live$ on that trade (!), except you have to wait for the hour bar to pass. I like that, tester replication! They have on screen a digit counter, since IBFX is going 5 digit mid April I suspect this EA will be no problem as it will detect this and do the right math. It feels like they done their homework!
Their website now has many years past performance tests. Some years have unfortunate digs into the equity, some not, yet all years do well on the initial Thou by years end.
I use 25% risk; I'm not worried, as I have "reduce lots" = true so the EA will not continue trading previous size lots and kill the account, and authors mention in emails of late that 10 or 20 is ok.
This account number 57021 had $3.51 and was the published link for the micro-mini P.SAR when I started using it instead now with "Mergatroid" yesterday, it's up $.15 now:
http://206.71.180.236/statements/statement.htm
OkYdOkY! May we be "blessed" with "Mergy"
jb

[pipdetector]

The GMTOffset is crucial in backtesting. It seems that even one hour shift makes a world of difference.

Alpari UK offset is 1. This is the setting that I use.