RSI crossover EA - page 3

New comment
 

Hi Pingarcia,

Sorry, I am just learning to program and I can not help more. perhaps someone else would like to help.

Lotincrease: will increase or decrease lots in relation with account balance. If you use money management it should be put to false.

Recupfactor: if set to more than 1 will increase the number of lots by this number after a loosing trade. If set to 1 will do nothing

I agree that it does not seem to follow the rules completely. However, I have been forward testing at M15 timeframe with eurusd. and it is doing astonishingly well.

Please see attached statement. (note that some trades have been done with 56 lots due to my mistake in the settings of the MM which should be put to 1)

I think that it is worthwhile to continue forward testing. Back testing results are different.

Files:
pingarcia.zip  8 kb
 
Pinggarcia:
UPDATE:

After testing the 2 EAs presented here, I regret to inform that they don't seem to satisfy the criteria set - wrong entry.

Maybe someone can make the EA for this - kindly see post #4 for the rules, and post #19 for the actual entry/exit of trades.

Hi Pinggarcia - sorry buddy - I'm in the same boat and still in the early stages of learning the code. Your requirements seemed simply enough...

...still learning,

Mike

 

I use RSI-5 and RSI-14

1H time frame

Here's the entry point, for BUY:

When the value of RSI-14 go lower than 40, then on its way up, wait for RSI-5 to cross it. The value that we need is at the CLOSE of each candle only.

To lessen the false signals, the value of RSI-5 and RSI-15, when it crossed at the CLOSE of the candlestick must have a difference of 5 points or above.

To close the BUY position, wait for RSI-14 to reach 60 level or above. Then wait for the RSI-5 to cross it on its way down. Wait for the candlestick to finish and see if indeed RSI-5 will have crossed the RSI-14, whatever the difference of their value is, close the trade.

Here's the entry point, for SELL:

When the RSI-14 go higher than 60, then on its way down, wait for RSI-5 to cross it. The value that we need is at the CLOSE of each candle only. To lessen the false signals, the value of RSI-5 and RSI-14, when it crossed at the CLOSE of the candlestick must have a difference of 5 points or above.

To close the SELL position, wait for RSI-14 to go down to 40 level or below. Then wait for the RSI-5 to cross it on its way up. Wait for the candlestick to finish and see if indeed RSI-5 will have crossed the RSI-14, whatever the difference of their value is, close the trade.

This seems to be working fine with EURUSD. I wanted to back test it with other pairs too. Is that possible?

I dont know how to back test. And somehow, checking my manual testing, BUY seems to be doing better. Anyway, can you please see if the automatic back test for the month of April yielded the same results?

April 2007 manual test results:

BUY 4/4 (00H) - 1.3335 (+34)

CLOSE 4/5 (02H) - 1.3369

BUY 4/9 (07h) - 1.3365 (+42)

4/10 (00h) - 1.3378 (+29)

CLOSE 4/10 (07H) - 1.3407

BUY 4/23 (14H) - 1.3565 (+61)

4/24 (05H) - 1.3566 (+60)

CLOSE 4/25 (00H) - 1.3626

BUY 4/26 (17H) - 1.3604 (+19)

4/27 (00H) - 1.3597 (+26)

4/27 (07H) - 1.3598 (+25)

CLOSE 4/27 (15H) - 1.3623

A total of +296 pips.

The reason why there are multiple open trades, is because RSI-14 did not reach 60 level, and the market ranges up and down, and then up again.

Please feel free to test, and put your comments.

Some ideas:

1. As it is, there's no SL set, which is very dangerous, thus I was thinking of setting the SL at 30pips

2. Maybe the RSI threshold value can also be changed 30 (from 40), and to 70 (from 60), Hopefully, the EA can make it "adjustable" for purposes of testing and see the optimization level.

3. Currently, it is set at 60M timeframe, I am also interested on how it will fare with shorter time frames.

 
rarango:
Hi Pingarcia,

I agree that it does not seem to follow the rules completely. However, I have been forward testing at M15 timeframe with eurusd. and it is doing astonishingly well.

I think that it is worthwhile to continue forward testing. Back testing results are different.

Nice to know you are getting nice results testing this one on EURUSD, too. But, are you also doing this manually, according to the rules I set out in post #4? Or are you using one of the EAs posted in this thread?

 

I'm also demo trading it manually, and it profitable with some few loss trades.

But using the EA, it seems the EA don't enter at the criteria you have suggested. I hope creator of the EA or someone with programming knowledge can help us fix it.

 

I'm done - take a look into another thread

actually I posted into the wrong thread 3 times in a row. Dain Bramaged...

RSI crossover:

The idea itself is interesting but again

Lets see:

1. buy when

RSI-14 < 40

RSI-14 of this bar is higher than RSI-14 of the previous bar

RSI-5 is at least 5 points higher than RSI-14

If that is all, it can be done

2. close buy when

RSI-14 > 60

RSI-14 of this bar is lower than RSI-14 of the previous bar

RSI-5 is LOWER than RSI-14 at least 1 point

3. vice-versa for sell, meaning we sell when we close buy and we buy when we close sell

done. no success in trading though.

Files:
rsi_cross_shinigami.mq4  6 kb
 

The rules you posted change some things. I'll be back to you when I'm done re-writing some things in the code.

 

Thanks!

 

Here are the conditions:

if(iRSI(NULL,0,14,PRICE_CLOSE,0)<RSI14_Value_Buy_Open) //open buy where rsi is lower than some value

{

if(iRSI(NULL,0,14,PRICE_CLOSE,0)>iRSI(NULL,0,14,PRICE_CLOSE,1)) //and RSI is going up

{

if(iRSI(NULL,0,5,PRICE_CLOSE,0)>iRSI(NULL,0,14,PRICE_CLOSE,0)+RSI5_Diff_Buy) //and rsi5 is higher n points

{

if(iRSI(NULL,0,5,PRICE_CLOSE,0)<iRSI(NULL,0,5,PRICE_CLOSE,1))

{

if(iRSI(NULL,0,5,PRICE_CLOSE,1)<iRSI(NULL,0,14,PRICE_CLOSE,1))

With these conditions we have NO buys during 2007. Are you sure you are getting everything right?

I've used your rules:

RSI crossover:

1. buy when

RSI-14 < 40 (1st condition)

RSI-14 of this bar is higher than RSI-14 of the previous bar (second condition)

RSI-5 is at least 5 points higher than RSI-14 (third)

... and, RSI-5 of this bar is lower than RSI-5 of the previous bar (added this, fourth condition)

Like what happened on Apr 4 (0H), when the value of RSI-5 is 44, and RSI-14 is 37 (difference of at least 5 points). And the previous value at Apr 3 (23H) of RSI-5 is 22, and RSI-14 is 29, to make a RSI "cross".

The value of which are taken at the close of the candle, at the candle where the cross just happened. If it happened on the next candle, we don't open any trade at all.

Please re-write that passage in typical English, I do not quite understand "the value of which are taken at the close of the candle". We use last value and current value. I've checked before-last value and last value. In both cases no buys at all.

Stoploss of 20, kindly added by me, kindly eats your depo in 10 deals. Use 100 instead.

If you actually meant that RSI5 should be going UP for buy, then we have buy orders this year.

 

Thats it for now, I'm going to sleep.

Only Buy orders for now.

General income 2006-now is 8076 pips with profit factor of 1.76

Are you sure you want to share this system?

If you want, I can further contact you on e-mail (pm me)

Or I can simply post it here. But I do not want some individuals selling it as a holy grail for thousands of dollars.

12345
New comment