EA strategy testing questions

To add comments, please log in or register
Don Steinitz
411
Don Steinitz  

Hello:

Some you may know me my name is Don Steinitz. I have a successful strategy and known as "steinitz method revealed here" over at the trading systems thread. I am here asking some questions about EA's and back testing them. First off I am learning all I can to aid my programmer in answering some of his questions and to best communicate my needs.

Here are some questions that I need answered:

1. In the strategy tester section I see a back test can be run 3 different ways

1a. Every Tick 1b. Control points (default setting) 1c. Open prices only

Is there a generally excepted way to run all EA's. The results are drastically different. 1B. produces excellent results

Also if a program is run using optimized feature is the code in any way changed internally by this function? I ask because I want to make sure I am seeing non optimized program results.

All this is important information for my EA being worked on presently. By the way I found an EA called "Vegas Tunnel Method 2" and it is showing extremely good results on 1hr and higher TF's across all currencies. I am mesmerized with the results. No compounding of lots and constant results.

TIA

lowphat
466
lowphat  

every tick should be more accurite than the other options.

having said that there is No substitue for running on a real account although the backtest is a nice tool to see if your EA is behaving correctly.

not sure i understand your optimize question but optimization just takes an external varible and runs it through a range. lets say you would like to see how your ea did with different take profits. in expert properties you will notice checkmarks next to external vars. place a check next to takeprofit enabling opt. mode on it. put in 20 for start, step 5, and end of 50.

it will run a backtest with tp settings of 20,25,30,35,40,45,50 and give you a summary if it encountered any positive results.

it will Not change any of your code.

Eureka
61
Eureka  
steinitz:

By the way I found an EA called "Vegas Tunnel Method 2" and it is showing extremely good results on 1hr and higher TF's across all currencies. I am mesmerized with the results. No compounding of lots and constant results.

TIA

Can you point us in the direction to the Vegas Tunnel Method 2? Does it use special indicators? I'd like to test it while I'm waiting for the Steinitz EA to come out.

TIA

Don Steinitz
411
Don Steinitz  
Eureka:
Can you point us in the direction to the Vegas Tunnel Method 2? Does it use special indicators? I'd like to test it while I'm waiting for the Steinitz EA to come out. TIA

Surely......

Look at this EA and tell me what you think. It is producing extremely +++ results on 1hr and up TF's. All currencies. Can somebody tell me what this EA is looking at for entries. The MQL file is attached.

The results are the best when using "control points". I have been doing just that experimenting until I get my EA running.

Files:
Eureka
61
Eureka  

Many thanks, Steinitz. Will start testing this evening.

Don Steinitz
411
Don Steinitz  
lowphat:
every tick should be more accurite than the other options.

having said that there is No substitue for running on a real account although the backtest is a nice tool to see if your EA is behaving correctly.

not sure i understand your optimize question but optimization just takes an external varible and runs it through a range. lets say you would like to see how your ea did with different take profits. in expert properties you will notice checkmarks next to external vars. place a check next to takeprofit enabling opt. mode on it. put in 20 for start, step 5, and end of 50.

it will run a backtest with tp settings of 20,25,30,35,40,45,50 and give you a summary if it encountered any positive results.

it will Not change any of your code.

Thanks for your explanation of that input screen (optimization) as I had questions about the inputs. I will attach this EA. It's remarkable. It doesn't compound trades and get's the candy! Tell me what you think it is looking at for indicators and entries. I have tried to analyise it but it is hard to figure out by looking at the chart. I need someone that can read the code.

Trust me you will be surprised....

lowphat
466
lowphat  

if the close of the (current chart period) last bar is above the current 1 hour

169 period expontial moving average ..buy

if the close of the (current chart period) last bar is below the 1 hour 169ema prev bar ..sell

its all jacked up

the original code that is commented out is

if expontial moving average 12 was below 169 and now above 169 to buy

if expontial moving average 12 was above 144 and now below 144 sell

Don Steinitz
411
Don Steinitz  
lowphat:
if the close of the (current chart period) last bar is above the current 1 hour

169 period expontial moving average ..buy

if the close of the (current chart period) last bar is below the 1 hour 169ema prev bar ..sell

its all jacked up

the original code that is commented out is

if expontial moving average 12 was below 169 and now above 169 to buy

if expontial moving average 12 was above 144 and now below 144 sell

Isn't there a 60 EMA or is that (1 hour)? For simple code it's performing like crazy. Have you looked at it?

Don Steinitz
411
Don Steinitz  
lowphat:
if the close of the (current chart period) last bar is above the current 1 hour

169 period expontial moving average ..buy

if the close of the (current chart period) last bar is below the 1 hour 169ema prev bar ..sell

its all jacked up

the original code that is commented out is

if expontial moving average 12 was below 169 and now above 169 to buy

if expontial moving average 12 was above 144 and now below 144 sell

I see it now. Designed for 1hr hence the indicators will now show up. This puppy rocks!

Thanks a lot!

lowphat
466
lowphat  
steinitz:
Have you looked at it?

the code is messed up. if by some miracle that jacked up code even ran on a live account i wouldnt use it heh

Don Steinitz
411
Don Steinitz  
lowphat:
the code is messed up. if by some miracle that jacked up code even ran on a live account i wouldnt use it heh

Jacked or not it produces results on the 1hr and higher. It's performing so far forward testing it. Just 1 position open thus far.

12
To add comments, please log in or register