SSS EA v1.0 - Auto Trader Bot

mos def in that sense

This looks like an evolution, hopefully I'll post a set file today so we'll have a baseline to start with.

My whole approach to testing/forex in general is to find a new use for something that's been done before. If you think about it almost every invention ever created is just a new use for something that was invented before.

Glad to hear someone finds those sets worthwhile, hopefully someone else will take those and make them work in the real world.

Hi,many thanks for proposing this EA.I've open a demo account this morning and first microlot short trade launch at 15H GMT+1.As a newbie,I've been around on many interresting threads and learned a lot during these past 15 days.

Regards to all.
P.S My english is a bit approx,so sorry.
Michel.

Please see the attached graph, EUR/USD 1 Hr, all default settings. Ignore the the dip in the end, that was just the end of the test. I noticed that there was quite a bit of drawdown. One of the trades that were forced closed at the end of the test had been open since May '09. The EA is profitable, but it might be hard to stomach the drawdown in a live account.

I will try and add a stoploss and see what happens, but I have noticed doing so will reduce profitability.

Hey Roller,

Thanks for the efforts. What source data/broker are you using for you backtesting ?

There is a trailing stop loss active in the default settings.

Good to have you on the SSS test team.

Thanks,

getdown71

I downloaded Metatrader from the Metaquotes website, and the data is coming from the "Metaquotes Software Corp." The demo account assigned to me is Alpari US, but I haven't started demo testing yet.

Looks intriguing. I've seen some very good work from these guys before. Can you give us a basic idea of what it's doing under the hood, though? Just a quick lowdown on the style of trading perhaps, not the actual logic that's being used, of course. I'd just like to know what to expect when testing it.

Basic overview;

The Primary Indy is the Price channel stop. Trends are picked up by this great piece of work.

Confirmation filters;

A) Damani - a rough measure of liquidity and movement of that liquid. Once it states "TRADE" it will look to final filter. In times of low liquidity and lesser market activity it will print "DO NOT TRADE" and therefore blocks any entry.

B) DSS Bressert - This final filter is THE sanity check prior to passing an entry on to the shell EA for execution. (if the shell logic permits - see the great manual for the shell core functionality)

Exits,

We have 4 methods of taking profits

2 hard levels in pips for TP1 and TP2
2 rubbery levels that vary with ATR of that pair (set via multiplier)

We have 1 method of stop loss

1 rubbery level that varies with ATR of that pair (set via multiplier)

In a nutshell, that is it. There are a swathe of other xux99 secret sauce bits in the shell EA which turn this from a mostly winning/but some bigger losing moments EA into a consistent pip machine with less than 12% DD on EURUSD of the last 2 years.

I think I had closer to 39% drawdown in my backtest. Were your settings a little different?

The settings are default.

Your 2 variations to my backtest are;

1) I use FXDD as my demo broker

2) I went back 2 years to present day (and had more profit in the first 5-6 months and therefore lower net drawdown)

Seem not bad made some tests on many paird. Will install it now on a server with a good broker and add it to almost ALL pairs. Timeframe 1H ?

Regards
FX13

For details post 1

Hey FF I spent last night (yea a friday night...:/) backtesting this EA till i was seeing about 5 or 6 balance lines and a bunch of swiggly green equity lines, mostly winding up in the zero position, as usual.... It was here that i had an epiphany....

All of my previous setups with SL's and TP's failed miserably so most of my newer setups had some sort of TP, no SL, whether a batch %, ATR, trailing stop whatever.... And all of them did ok until some trade comes along and either wipes or threatens to wipe my precious backtesting dollars. So I thought to myself...

What if using a TP, even ATR based or otherwise, is just as bad for your equity as using a SL? What is the backtester not showing me in terms of pips left on the table?

Which led me to this conclusion: Any system based on any sort to TP or SL will ultimately fail...

Why? Because eventually you will hit a string of losses that will damage your account beyond repair or wipe it all together. Run the math, its not that hard.

The traditional counter argument to that has been "use good MM". Two problems. Good MM just delays the inevitable and secondly if you set it too low the EA will make like 1% a month. I can get that at my local bank... (I live in Brazil, outrageously high intrest rates...)

Some have suggested grid trading as an alternative. Well I could sit here provide you with mathematical explanations of why grid trading wont work long term; or you could scroll a EURUSD chart to Dec 4, 2009 and see for yourself.

I could go on and on critiquing the different methods of overcoming this problem but none of them address the fundamental underlying problem.

If not TP's and SL's whats the new paradigm shift? We have to start thinking like the brokers, b/c after all 90% of brokers make money and 90% of traders don't.

How do brokers make money? Short answer: Spread. How are we going to make money? Spread.

I hope the set file attached begins a new way of thinking about EA's here on FF.

On the surface it looks insane. Its a martingale with no SL, which should be enough to scare the sweet jesus out of the newbies.

Basically we're using the indis and trades to determine market direction. Kind of how a poker player places "tester bets" to see how his opponent reacts. If we hit a new high or low, we increase our trade size hoping to capitalize on an overall move. We make money at the end of the week on the spread between winners and loosers kind of like a broker.

If the markets go up, we win, markets go down, we win. We'll break even if the markets go sideways. That's the general working idea. Honestly i don't understand all the math involved behind a loosing week, but it has to do with fast runs late in the week against us/news, indi failure and errors in backtesting. We can also loose if we overexpose ourselves to the market. I don't run news filter on backtests bc i find it unreliable and slows down the process significantly.

I'm using GBPUSD and EURUSD m5 in my backtesting and closing weekly but this concept can and should be scaled for the appropriate pair. We can trade on m1 and close every 4 hours for GBPJPY for example or trade EURCHF on h1 and close monthly. Finding the right TF combinations to look at is key for this strategy.

As currently built, SSS ea is probably not the best tool for this and probably not where its creators intended it to go. But I hope this set provides us with a roadmap of where to go in the future. Sorry for the short backtest (07.01.09 to 01.29.10) but i only have 8 gb of HD, I'd love to see this run longer term on other pairs.

My conclusion is - any predetermined TP/SL EA based on mechanical system which consist of conventional or custom made indicators fails. Look at THV EA thread, system looks good, EA fails. If you experienced trader you would never open those trades on low volatility or news or hundreds other factors like EA does. Surprisingly almost only system which works with indicator based EAs is martingale. They would tell you it doesn't, but it does. You have to work very hard to determine how it works though. Forget about 2% of equity per trade, back test to identify the max movement w/o retracement so you could determine your pip step between trades and how much of $ you need to survive. This is completely different approach. You have to withdraw from your account often too, so if black swan event happens eventually (and it will, believe me) you would lose as little as possible. Did you see many successful live EA statements?
Here is one of them (not mine).

http://www.onix-trade.net/?act=monit...stat&xid=14780

Totally agree.

The ONLY thing that works as SL and TP with any accuracy over time and market conditions are fibonacci retracements and extensions. Price respects fib levels more than it does abitrary TakeProfit and Stoploss values.

Hi Naynay,

Can you detail your idea out, so we could add it to SSS v1.01 ?

I think I understand your statement, but I want to be 100% sure.

Thanks,

getdown71

Fibs are mos def an improvement.

But only opening orders based on indis and closing them all at a given interval is still better.

If your indis are accurate good entries > bad entries = profit. Barring any sort of late move agaisnt our majority position....

After testing pretty much every EA on FF, TSD etc... and many commerical EAs I´m fairly convinced this is the only way to make consitent profits with an EA in MT4.

Oh and my motherboard fried last night but i improved orginal set i posted here by reducing min order distance to 0 on GBPUSD although this sent DD through the roof... this should probably be ATR based imo... If 10 pips worked well on GBPUSD, the equivalent ATR should work on other pairs

Thanks Joe - Another item for v1.01 - ATR based min order distance.

Since I don't really know how your EA trades, I'm not sure it will be much help to you...

Having said that, what I am suggesting is an EA that "reads" support and resistance and price action.

In manual trading, I can manage a high degree of accuracy when the price breaks these levels. Once the level is broken, you KNOW where the price is eventually going to go. This is where fib retracements come in. When you know where the price is going to go (or where it might go thorugh), then you have your Takeprofit. You stoploss then, would be just behind a fib level where the price would not go (and would be a strange event if it did).
I have done some manual trades with some accurate stop loss placement like this, resulting in profitable trades with very little drawdown.

Like this, and EA can trade a lot more accurately, but half the battle is knowing WHERE the levels are, and how to read the price action when it occurs... and THAT is not only hard to determine, but hard to code as well (although I am working on trying to do it myself).

I am still learning myself.

Does your EA use the levels on the DSS or just the colored dots?

Thanks

Lux