Optimization of UT1.3: HOWTO

Here is my method to optimize UT1.3. 

This will allow you to get nice results very fast, then you can improve the generated set by testing each month separately, use forwarding test and optimization. 

The settings tab:

1. TIMEFRAME: If you start the optimization for a new pair from nothing existing you can use M15 or M30
2. EXECUTION: Use Open Prices Only
3. OPTIMIZATION: Balance + Min Drawdown

 The Inputs tab, set:

1. CompletedBars =  true
2. GlobalTF = Current
3. useWeekDay = false
4. UseTradingTime = false
5. ManageOpenedTradesDuringTimeOff = false
6. MaxVolumePosition = 10
7. MaximumPercentageatRisk = 20
8. UseLotManagement = true
9. LotManagementType = 1
10. LotPerTrade = 10
11. TakeProfit, StopLoss, TrailingStop = 0

Your inputs tab should look like this:

Now select inputs for first optimization

then, launch optimization....

30 minutes to 2 hours later, you should get something like this:

Select the best result (less drawdown) and run it by clicking on it.

As you can see, it's very choppy; hopefully it's only the first optimization step..

Now, let's select the following bloc of parameters:
Uncheck first all past checked parameters.
Then select only TradingTimeStart and End Start 0 Step 1 Stop 23
Run the optimization like this.

You should get something like this:

As you can see, the result is much higher and drawdown is really less...
Let's run the best result now

Still Choppy but looks better...
It's important to run the best result each time an optimization step is done because it loading the values in the inputs allowing us to start the new step with best parameters found.
Now, let's uncheck all boxes and select only what's about money...

It's important to optimize TP,SL and TS because they will be the watchdogs during non trading times and will avoid a big loss if fast market moves may occur (and they often occurs)
Let's run the optimization now...

Here is what we get...

As you can see, profit was multiplied by 2 and drawdown divided by 3.

Lets run the better result:

Getting better again...

Now we will make a fine adjustment of indicators.
Let's uncheck everything and check only cci, atr and both SARx
and run optimization again.
Each time you will do that you will get better results and less DD.
It's possible to get a dd <= 5 most of the time.

To train, you can use a a pair with less spread like USDJPY.

At the end of the process, you will just have to select each tick instead of Open Prices to be sure that the results are the same. (and you will see that they are !)

Normally, you managed to get a good starting set in less than 1 day. without having to use the ruinous MQL5 cloud agent. 
A 4 core system is highly recommended.
Now the boring job will start like check every month, use forward testing and all these things that will make your set is a real winner.
As you can see on this blog, this bot it absolutely not a "guess-what" EA and everything finds a logical explanation through procedures.
The results you find with the tester are reliable and as it's shown on other pages, real trading has same results as the tester.
Only indicators and fine coding are used here, and this bot works with trend tendencies, this is not a scalper designed only to work in backtests.
With some small efforts every one is able to get some good results from this bot and the possibilities to use it are really wide.
Within newt week, I will release this V1.3 on MQL5 market with one default set working for the main well-known brokers.

If you have questions, feel free to ask...