Well everyone, I've patiently read through 273 pages of this thread (a few days later and with sore red eyes) I'd like to contribute. I see that this thread has split into the new development of different EAs based on 10points3, but I'd like to get back to making some progress here in this thread so I'm going to jump in and contribute. I know the 10points3 code well and I can modify as we come up with new ideas. I've programmed for a while and recently learned MQL4 to automate some of my trading ideas. I've tried a lot of the prepackaged EAs and I've also programmed a few based on technical articles. Automated trading is an interest of mine and I'm studying Artificial Intelligence in my free time. I look forward to developing this system or at least a successful variant.

So far this 10points3 seems to be very aggressive and can make/loose some serious cash fast thanks to the Martingale strategy. I like this idea because I'm in this to make money.

I've tried a lot of the variants of 10points3 including the following:
10points3 from thanks to Alejandro Galindo
10points3 with Dynamic Stop 10points 3.mq4 thanks to tururo
10points3 with Dynamic Stop RMI 10points 3.mq4 thanks to mrtools
10points3 with Dynamic Stop RMI new version
10points 3.mq4
10points 3.mq4 thanks to mrtools
EuroX2 - Mostly identified as a Scam modified version of 10points3 but I tested it anyway. You can find it, just don't pay for it.
V12 - I never test this, before my time on this board thanks to davidke20, many of the results posted on the board looked good.
RB23DETT - I never tested this, also before my time on this board thanks to davidke20

Before publishing a lot the results, my questions, observations, etc...I'd like to add a few more thanks to all of the creators and testers especially yeoeleven and davidke20.

So lets get started:
*note: these results are based on back test 90% modeling quality every tick. I know this is not the best for this type of strategy due to slippage and tick activity but I will be forward testing from now on.


10points3
Tried setting take profit from 40 to 20 on H4 so it will grab profits faster and get out, lasted 3 months longer with only slightly higher balance than default, tried with 60 died quick Tried on H1 with 20 Take profit, died two months sooner, tried with 60 and died very fast. If you decrease the take profit to much it may last longer but you will be working harder for the same money. A take profit of 40 may make you the same amount in 1 month that a take profit of 20 would do in 2 months.

If you drop maxtrades to a lower number it will typically last longer but balance before death will be much lower (meaning it will grow slower). If you decrease maxtrades too much it will not have room to work. Best setting depends on account size because you will be capped by your available funds. If you have maxtrades set to 10 but you don't have the available funds to trade past 5, you will get stuck at 5 and wait until margin call or stop out (NOT GOOD).

For the timeframe, if you go down from H4 to H1 it will typically last longer and grow faster

Pipstep is the distance between orders. The default is 15. I tried 7 and it mad money fast but died sooner. I tried 12 and it made double the money of the defaults and lasted longer (died later). The step depends on the markets volatility. In a highly volatile market you will want a larger pipstep so you don't overstep into a bad position. On a shorter timeframe H1, decreasing the pipstep caused it to die sooner. Higher pipstep tends to be better in lower timeframe.

Trailing Stop doubled to 40 on H4 time frame and got 4 months farther, tested with a 10 and dies same as default 20. Tried on H1 timeframe with 40 trailing stop and died sooner, same with 10 trailing stop.

THE CONS - This EA uses the MACD for entry. If last MACD was less than current it sets up a long position and all of the following positions will attempt to be long even if MACD turns back down. This is a very blind approach. I think davidke20 or someone else said it is about the equivalent of flipping a coin or letting a monkey put in your trades. This EA needs a sophisticated entry and constant monitoring of the trend to possibly cut losses and change direction.


10points3 with Dynamic Stop
With Initial Stop set as 1, this lasted about a month longer than the regular 10points3 before dying, noticed that with this setting the initial stop of the first step was 151 pips below entry price. If you set it to 2 then it is 152 below initial entry. I'm not sure if this design was done deliberately or not. At a later time I'll look into the code to understand it better.

Although this EA lasted longer after it took its first big hit it slowly died over the next month. Not the same exponential growth and then death like regular 10points3.

These seemed to be the optimized settings posted in the thread by yeoeleven
EURUSD, M30, Be sure to change end year on this EA or it will not work since it is default 2005

TakeProfit=38.00000000
Lots=0.10000000
InitialStop=1.00000000
TrailingStop=28.00000000
MaxTrades=5
Pips=18
SecureProfit=28
AccountProtection=0
OrderstoProtect=0
ReverseCondition=0
EURUSDPipValue=10.00000000
GBPUSDPipValue=10.00000000
USDCHFPipValue=10.00000000
USDJPYPipValue=9.71500000
StartYear=2005
StartMonth=1
EndYear=2007
EndMonth=12
EndHour=22
EndMinute=30
mm=1
risk=3
AccountisNormal=0
Magic=10201


10points3 with Dynamic Stop RMI
There are a lot of settings to this EA and I do not yet understand RMI. I'm going to research the indicator and test this version of the EA further.

Here are some of the best settings posted in the thread. I believe they were posted by the author.

TakeProfit=40.00000000
Lots=0.05000000
InitialStop=0.00000000
TrailingStop=25.00000000
MaxTrades=5
Pips=15
SecureProfit=20
AccountProtection=0
OrderstoProtect=0
ReverseCondition=0
EURUSDPipValue=10.00000000
GBPUSDPipValue=10.00000000
USDCHFPipValue=10.00000000
USDJPYPipValue=9.71500000
StartYear=2005
StartMonth=1
EndYear=2050
EndMonth=12
EndHour=22
EndMinute=30
mm=1
risk=1
AccountisNormal=1
MACD_fast_ema_period=14.00000000
MACD_slow_ema_period=26.00000000
timeperiod=240.00000000
timeperiod2=30.00000000
Par=10.00000000
CountBars=300.00000000
Length=14.00000000
Length2=7.00000000
Smooth=5.00000000
MASmooth=0.00000000
RMIPeriod=14.00000000
MomPeriod=5.00000000
Magic=10201

10points3 with Dynamic Stop RMI new version
same as above, will test after analysis of RMI and indicator used in code

EuroX2

This EA only lets you control the following 4 settings
Defaults
MaxTrades 4
Pips 5
Take Profit 40
Trailing Stop 20

Running with defaults it lasted about 3 months longer but was only about 6 thousand higher than the standard 10points3. This means you worked 3 months harder for not much additional money. When run on H1 timeframe the EA died vary fast, it is recommended on H4 but I decided to try H1 anyway.

I observed that even though the maxtrades is set to 4, it will vary the starting lot size depending on the account balance. You have no ability to disable
or override this. Before the great loss it was starting at .4, which went to .8, 1.2, then 2.4 This may have contributed to the great loss. Since this EA is locked and not truly optimized I will abandon testing this EA.


Notes and Developments - These are some of the things I'm considering coding, what do you think should be added?
Modify rather than close positions to avoid commissions
Dynamic pipstep based on ATR or predicted ATR or high activity time frames, higher for fast moving, lower pipstep for slow moving
Recheck direction using indicators before each pipstep, if changes direction look to close and switch direction (swap?)
Possibly have a maximum number of trades per bar so as to not get caught continually buying into wrong direction
Close and reverse function, BAIL/close function - thanks to V12 for this idea
Use a lot of other indicators like RSI, HMA, ASC, ISAKAS etc... use ratio, say 3 of 5 indicators agree, then go for trade
Set to close out earlier trades to eliminate dead weight. If at level four close level 1, if at level 5 close level 2 etc...
Setup a grid or some other system in code to trade news, or avoid news, then reinitialize 10points3 after news events

neta1o,

Thanks for the summary. I too have been trying to understand more about the parameters for these programs and have done some backtesting.

You have some very ambitious, but worthwhile goals, and some excellent ideas. For me, the most frustrating thing that happens as I watch the visual on backtest is when it continually goes in the wrong direction. But all of your ideas are good.

Best of Luck.

saintmo

Saintmo, that is one of the first things on my list to fix. Right now that is happening because it only decides which direction to go on the first entry. After that it just keeps going in the same direction and hoping for the best.

I'll be posting again soon...

Okay. If that is the case, we will eleminate the PipStep, and open trade according to the next signal appear. Do you want that? Eg.

Buy 119.00 - 0.10 lot Buy signal(false signal)
Sell 118.40 - 0.20 lot Sell signal(false signal)
Buy 118.50 - 0.40 lot Buy signal(this time correct)
Price got up another 20pips, we close all positions

Is this the criteria? If yes, let me know, so we start working on this 1.

Regards

David

davidke20,

Hey David, thanks for looking into this.

Here is my thoughts so far...I think that if we eliminate the pipstep we'll loose some of the strength of 10points3 and martingale in its raw form. I'd instead like to use an indicator (or two/three) as a progressive signal to audit our position.

*note - I think the pipstep should be variable depending on the volatility of the market ATR etc...

Example if we get a signal to start going long for our first entry because the RSI was > 50 and the previous RSI was < 50, we enter at .1 lots. Depending on the volatility of the market our next pipstep may be 7, once that is reached we check our RSI to see if it is rising. If it is we put in our next buy at .2 lots.

If for some reason our RSI drops below 45 or some other value (I'm still brainstorming this) then we close all long positions and start our progression in the short direction .1 lots.

Thoughts?

EDIT: The more I think about it, dropping the pipstep may be a good idea. I'm worried that if we lock ourselves into a forced pipstep we may get caught over buying into bad positions (even with the audit using indicator above) since the pipstep is such a generic buying structure. Those sideways markets would kill us if we used the pipstep.

If you use indicator's signal, you cant use the pipstep(no matter you wanted to make it dynamic or what). Vice versa if you use the PipStep, when is the next progression then? The signal or the pipstep? I'm having a hard time to do this part too, if you wanted both, I'll be having trouble now because I dont know how to make it work, may be some other may help.

Regards

David

My idea is to use indicator signal for entry and to continually reference that indicator signal while in position to make sure we don't become victim of a hard run in the opposite directoin. If that indicator reaches a certain value in the opposite direction of our position we use that to close and cut losses (this is our fail safe to avoid those big killings from 10points3). We have to give it some room to work so it cannot be a simple "If the indicator is up we keep adding, if its down we close and switch...there has to be some padding to give it room to work"

By doing this we can still use the pipstep because if we enter a long according to our indicator and it turns short for a bit, we will add to our long position according to the pipstep, we wouldn't want to close and go short immediately. We have to establish a value in the short direction far enough away to give our EA room to work, but close enough to avoid the big dives.

I'm working on the code right now and I'll post it for you to take a look at soon. I think your input and anyone else who that wants to contribute will be extremely valuable.

Will be posting again soon.

Edit: Here is a small update to the code to create a better quality entry (see attached), although there is still no protection in this code. It picks direction at the beginning.

Attached Files

FXA0 - 10points3.mq4 (10.5 KB, 143 views)

Alright, I've just about got a basic copy of the code working and I'll post it soon.

But I need your(anyones) help.

The extra volatile fast moving markets are killing this indicator because of the pipstep. If you get a 40+pip bar with a pipstep of 10, you could have 4 positions on that one bar. This is good, but most indicators lag a little and this move will happen way before the indicators get a chance to protect it.

Now I could increase the pipstep but that would make me loose a lot of the smaller profits I've won, and in those extra volatile markets it is not uncommon for 40-50pip moves in one bar.

What would be the best logic to use to make sure it doesn't buy or sell more than twice per bar? Will I need to use time for this?

Few stuff added:

Fully functional money management: Standards, Minies, NANO
StopNReverse funtion : Hope it works
Clean up codes on PipValue : So you no need to define the pip value yourself
Revised OrderstoProtect : More efficient!

Hope this help. I have no data to backtest, should you have it, please do it and feed back. Cheers

Regards

David

Attached Files

FXA0 - 10points3v0.01.mq4 (10.9 KB, 161 views)

Against the fast moving markets like news time, an idea I had sucessfully tried is just to have a fixed minimum time between two entries. The value of the next entry is not given by pipstep relative to the previous entry, but well by pipstep relative to the rate X minutes after the previous entry. By this way, the pipstep becomes dynamic following the instantaneous volatility.