Hi David,
I've been tied down with computer problems and other distractions for several days. Sorry for not getting back to you. I should not have downloaded the micro data after all. I think I'll stay away from micro.

What news is there, if any? Have you made any changes? Shall we restart our testing? Shall I get back to optimization?

autumn

David,
Trailing Stop is not working.
Will be good if will works only on first open trade,
if positive.
B.

Thats what I meant. When doing micro account backtest, the MT4 will eat up all the memory available, so I have to gradually remove all the history data without jeopardized the memory usage. All we need is optimization. Micro testing is not a must, because whatever reflected on normal account should resemble on micro. The only difference is normal account needed an extra zero on capital in order to sustain the drawdown, whilst micro account doesn't need that much. I do micro testing because I'm hoping to run these high risk EA on my small money instead of my main account. Running it on a 100 bucks micro account doesn't hurt me. Besides, if this kind of high risk EA can make good return with minimum deposit, why not start small. Everybody in this business know this is a high risk game. I would like to see how this EA perform with different settings on different pairs.

Regards

David

__________________
"All through time, people have basically acted and reacted the same way in the market as a result of: greed, fear, ignorance, and hope. That is why the numerical formations and patterns recur on a constant basis." - Jesse Lauriston Livermore
I'm blogging my EA creation journal

Hi David,

I wanted to stay away from micro, but a friend of mine keeps insisting that I test on micro, so I decided to risk it. I'm optimizing v1.01 on GBPUSD 15min 3.5 years of downloaded (not Alpari) data on Interbank terminal disconnected from Internet. The chart below shows you a preliminary result. Do you think it is worth continuing even with large drawdowns (80%).

It seems to me that the goal is not to accumulate several million dollars in an account but to boost the account to a level that generates a steady and adequate income. I suppose we all have a different definition of what adequate is, but in any case the revenue should be stable. Based on this preliminary result, I think I would be quite happy letting the EA run for a few hundred trades, and when it started to invest the maximum lots (apparently 50 in this case) I would start removing the profits every week. You might even want to start pulling out profits before hitting the 50 lot level. What do you think?

Let me know if you have any suggestions. Thanks again for producing a great tool.

autumn

Attached Images

10p3v101gbpusdmicro3plusyears15min.gif (22.5 KB, 450 views)

I am afraid that I mistakenly used regular data for the test that I thought was a test on mini data. I suppose that the results are therefore invalid.

autumn

I have check the backtest data provided by several brokers - all supplied the same stuff, so I would be surprised if there is a difference between standard and mini accounts...

If it is disconnected, it canot be test as mini. 80% DD definitely not what I'm looking for, may be others like it. I'm more then happy to maintain as low as 6% drawdown with consistant profits. The yearly return of 20% is the key. And the 20% of profit must shows on the equity chart as smooth slope instead of spikes everywhere. This is due to spikes shows that the backtest is curve fitted. With spikes everywhere in the backtest, shows the test is fitted with TakeProfit/StopLosses that fitted the history data. With this kind of setup, if there is any bizzare things happened to the market that created movement exceeded the curve fitted value, then we're screwed.

Reason for 6% drawdown is the simple mind of capital protection. I know it seems insane to talk about capital protection at trading martingale, but who knows there is a way. 20% profit annual is a simulation of 6% yearly inflation + capital growth expanses(electricity, bandwidth, technology maintenance) that will caused me about 8% of my capital. Bare in mind 8% is for my capital size. I'm not sure about yours. Overall, my annual expanses will be 14%. Now we have a drawdown of 6% or profit of 6%. The net asset value for the fund will be down to 94% balance after a drawdown year. With the same step of bad performance consequtively running, it will take 12 years to halve the capital(which I think we're not that dumb to continue lossing the next 12 years straight). Vice versa, with the correct ROI rate, we'll be able to double the account in 3.6 years time, without considering the loss of inflation and or other expenses.

Its all about money management. Making an EA is not that hard. To make an EA work for you is not that horrible too. To manage the EAs to work with you, that is the crusial part and canot be ignore at anyone time. If you fail to portfolio your fund with correct risk exposure, you'll be quickly dry up the account instead of building wealth. With the wrong calculation on the expected ROI, even you balanced a 6% profit in the end of the year, you're still lossing interest because the actual profit you made canot cope up with the current inflation. You may find that last year you bought your lunch for $5, but its costing you $5.50 this year. That is a 10% difference. How can your 6% profit counted as a profit? It is a loss!!!

Difference is the lot size. Testing on MINI account will get a faster growth because the lot size can be vary for 3 decimal correction. Compound interest is faster and more accurate compare to standard account. An EA gone through both standard and mini test, you'll find similar pattern on the equity curve. The only difference is the rate of capital growth on mini will outrun the standard account.

Regards

David

__________________
"All through time, people have basically acted and reacted the same way in the market as a result of: greed, fear, ignorance, and hope. That is why the numerical formations and patterns recur on a constant basis." - Jesse Lauriston Livermore
I'm blogging my EA creation journal

Aha, didn't consider that - good to know!

"Its all about money management. Making an EA is not that hard. To make an EA work for you is not that horrible too. To manage the EAs to work with you, that is the crusial part and canot be ignore at anyone time. If you fail to portfolio your fund with correct risk exposure, you'll be quickly dry up the account instead of building wealth. With the wrong calculation on the expected ROI, even you balanced a 6% profit in the end of the year, you're still lossing interest because the actual profit you made canot cope up with the current inflation. You may find that last year you bought your lunch for $5, but its costing you $5.50 this year. That is a 10% difference. How can your 6% profit counted as a profit? It is a loss!!!"


Hi David,

I've been following your progress (with Yeo) from as far back as 2006. I must commend you guys for the great job so far especially on versions 2 & 3. It ain't easy for sure. I thought I should show you the template below; perhaps you can create a quick genius work from it any time .

Expert Advisor Builder for MetaTrader 4

I must say a big "thanks" for all your efforts regarding the 10p3v0.02 EA. I've been testing it back and forth and I'm certain that it'll rank among the best EAs of the year (no joke).

One question though, pls let me know why 10p3v0.02 & v0.03 both work with other brokers' platforms but not with FX Pro's. I can see that their numbers have an extra digit although they don't allow microlot trading (starts from 0.1).


Thankfully,

Piper.

Thanks David for the maxlot code. Have you come up with any other pairs that may have some potential besides eur/usd for ver 04? It is a multi-currency EA.....