Ok, so I am a persistent son of a gun and so have fiddled with robot building for over 3 years, you could say I have a few good ones now, I have in built so many in the past that my head spins sometimes lol ;)

Anyway, this morning I was "fiddling" with an old account, which has an old version on it and decided to bring a few of its features up to date a bit. 

It made me think about this forum, as  I have very specific rules for all my robots (ie, my single trade robots ONLY make profits, ie 100% - thats a good one right there ), and I would like to share them with you, perhaps it might help you if your into this crazy type of coding.

MY ROBOT BUILDING RULES :-

The main rule is this -, if it did a trade in live trading or in a backtest, I expect to see the EXACT trade in both lof them, no compromises, ther can be no exceptions. Now I often see it writen that backtests are not teh same as forward tests - WHAT A LOT OF GUFF - the truth is the robot has simply been "curve fitted". ( Thats just plain cheating in my language folks haha. )

Moving on, another of my funny rules is that I NEVER download back history, what I get from my broker, is what I got, and thats it. MY broker may delay my trades 1700ms and be slower than a packmule going up a mountain trail, they may charge me 3 pips spread, they may heavily filter their feeds, they may "snag" the occasional pip from me ( I have software that checks haha), but by gum they work with my robots and thats what counts, the reason my robots work is entirely due to not playing about with the data ! Why not you may ask ... well folks .. its because it makes my brokers very PREDICTABLE - and thats what indicators are good at, trying to predict where the market is going, based on wheres its been. So next time your tempted to get 3 years of data -  just wait 3 years haha.

Another golden  rule is that I only backtest for the past 6 months or so, the market is not the same as last year, so why bother with it, I tested then anyway, so why redo it all, after you have been building, tweaking and testing for a few years you will see it makes sense, stop worrying about the past and try to figure out where the market will go in say .. the next 5-15 minutes .. that shouldn't require to much backtesting what :)

Ok, so we have predictable data, robots that do what they say on the tin, up to date testing, and what else ...

Well, if I was to give anyone any advice on what to add to a robot, it would be this ...

   if ( DayOfWeek()==1 || DayOfWeek()==5) {return(0);}

Now don't get me wrong, it took me several years to even dare to add that, as it kinda contravens one of my other rules ( well apart from DON'T FIDDLE WITH INDICATOR SETTINGS that is haha - almost forgot that one - always use the default values - or your just "curve fitting" , theres that expression again folks haha , well its so easy to do and to fool yourself - if you have just realised that is what you have done recently - my work here is done :) ), the rule being ...

NO TIME RESTRICTIONS !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

So you can understand my "trepidation, would I be breaking this rule by not trading certain days, well I struggled for several years with absolutely no time or day restrictions, no news restrictions of any kind, nothing, but plain old indicators, and I do alright, but WOWWWWEEEEEEEE .. when I added that ( ok not exactly that, I did put start n finish times for those days in haha, but het I'm not here to build your robot for you lmao )you can imagine my surprise when a robot with a drawdown of 15-18% MAX, suddenly went down to 2-3% max drawdown, even doing a longer test (for me) of this year only brought it up to 5-8%, and finally 4-6% drawdown.

Nice you may say, but so what, you simply do less trading... DO I ????

NO !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

The robot openned and closed MORE TRADES !!!

Well you can imagine my ASTONISHMENT

Enjoy :)

I look forward to reading all your robot rules, or if you try that bit of code and find .. NIRVANA, like I did...