strategy BarHunter DAX v1p

• 01/15/2020 at 6:33 PM #116911
  Paul

  Hi Folks!

  Here’s a new strategy BarHunter v1p, which searches for the best intradaybar on any timeframe and enter on a defined break.

  The goal is many trades on higher timeframes. It had to work with 1% stoploss and a trailing stop.

  Then if there is a reasonable good equity-curve with minimal curve-fitting it’s possible to enable & optimise trend-detection and perhaps limit the hour of entry.

  Happy Hunting!

  5 users thanked author for this post.

  noisette, Balmora74, GraHal, Francesco, bullbear

  

  Paul

  Participant

  Topics: 40

  Replies: 812

  Been thanked: 660 times
  01/15/2020 at 7:03 PM #116917
  Francesco

  Looks pretty good as always!

  Thanks for your contribution 😀

  2 users thanked author for this post.

  Paul, Florian

  

  Francesco

  Participant

  Topics: 32

  Replies: 439

  Been thanked: 200 times
  01/15/2020 at 7:13 PM #116918
  Paul

  Glad you like it!  Wanted to prove a point in the day openstraddle so I create this 🙂  Many trades!

  Also use

  once barnumberlong =3 //long once barnumbershort=2 //short

  1 2	once barnumberlong =3 //long once barnumbershort=2 //short

  and it is even better!

   

  

  Paul

  Participant

  Topics: 40

  Replies: 812

  Been thanked: 660 times
  01/15/2020 at 8:27 PM #116924
  Francesco

  Ok i made a very fast test, and i noticed that on nikkei 225 that’s a percentage of gain similar to the dax (89.36%) but with a gain loss ratio much more lower.

  Do you think could be worth trying to optimize on nikkei? Not because I’m saying your code isn’t good, but in order to diversify assets 😀

  

  Francesco

  Participant

  Topics: 32

  Replies: 439

  Been thanked: 200 times
  01/15/2020 at 9:05 PM #116927
  Paul

  Ofcourse! Other markets is my next step.Take care of the spread and I would optimise between 15-min – 1 hour  timeframe. That should go quick, not many bars in a day. That with tds off.

  But it’s always the details! The trailingstop is suited for the dax, but maybe not for the nikkei or s&p or forex.

  So you have to turn on the graph of the trailing stop and need to see that it kick’s in.

  i.e. for forex use

  atrtrail=averagetruerange[atrtrailingperiod]((close/1)*pipsize)

  1	atrtrail=averagetruerange[atrtrailingperiod]((close/1)*pipsize)

  dax

  atrtrail=averagetruerange[atrtrailingperiod]((close/10)*pipsize)/1000

  1	atrtrail=averagetruerange[atrtrailingperiod]((close/10)*pipsize)/1000

  s&p is different too.

  1 user thanked author for this post.

  Francesco

  

  Paul

  Participant

  Topics: 40

  Replies: 812

  Been thanked: 660 times
  01/16/2020 at 1:53 AM #116946
  Francesco

  Optimized for nikkei. Huge gain percentage but huge drawdown.

  

  Francesco

  Participant

  Topics: 32

  Replies: 439

  Been thanked: 200 times
  01/16/2020 at 2:05 AM #116948
  Francesco

  Sorry does not appear in the screen, the dradown is over 2k

  

  Francesco

  Participant

  Topics: 32

  Replies: 439

  Been thanked: 200 times
  01/16/2020 at 7:29 AM #116950
  Paul

  Would love to have it working on the nikkei! Drawdown is big but equity-curve is oke.

  Here another one….optimise 0-23 and set break to 20 (not optimised), no tds and spread 15.

  Such equity-curve created in less then 5 minutes.

   

  

  Paul

  Participant

  Topics: 40

  Replies: 812

  Been thanked: 660 times
  01/16/2020 at 7:30 AM #116952
  Paul

  wrong pic, here’s the good one

  

  Paul

  Participant

  Topics: 40

  Replies: 812

  Been thanked: 660 times
  01/16/2020 at 7:55 AM #116956
  Vonasi

  Paul – I see that you are back testing on only 25k units. How does it look on 100k (or 200k if you have it) with the optimised settings fixed?

  

  Vonasi

  Moderator

  Topics: 192

  Replies: 5088

  Been thanked: 1811 times
  01/16/2020 at 10:24 AM #116969
  Francesco

  @Paul love it! This system is surprisingly versatile! Can you share the version for the South Africa?
  There’s no way to “upgrade” the number of trades? Could be perfect for all day trading

  
  @Vonasi Don’t you think a backtest beyond 5 years is useless?

  

  Francesco

  Participant

  Topics: 32

  Replies: 439

  Been thanked: 200 times
  01/16/2020 at 12:22 PM #116978
  Mauro T. “Algorithm System”

  If it can be useful, done with 200K

  2 users thanked author for this post.

  Vonasi, Paul

  

  Mauro T. “Algorithm System”

  Participant

  Topics: 18

  Replies: 313

  Been thanked: 76 times
  01/16/2020 at 4:17 PM #116998
  Paul

  In theory it should be possible I guess to make 100k bars of 15 min the same as 25k bars of 1 hour. (without atr but with % profit target)

  if longtrading or (longtrading and shorttrading) then if intradaybarindex=barnumberlong then breakvaluelong=highest[4](high) //experiment! endif endif if shorttrading or (longtrading and shorttrading) then if intradaybarindex=barnumbershort then breakvalueshort=lowest[4](low) //experiment! endif endif

  1 2 3 4 5 6 7 8 9 10

  if longtrading or (longtrading and shorttrading) then if intradaybarindex=barnumberlong then breakvaluelong=highest[4](high) //experiment! endif endif if shorttrading or (longtrading and shorttrading) then if intradaybarindex=barnumbershort then breakvalueshort=lowest[4](low) //experiment! endif endif

  Lines are the same on both timeframes.

  And repeat the entry to check four bars with buy/sellshort on stop instead of one

  If that’s the cause it would take a bit the worry away from only 2500 bars.

  @Francesco The South Africa version

  1 user thanked author for this post.

  Francesco

  

  Paul

  Participant

  Topics: 40

  Replies: 812

  Been thanked: 660 times
  01/16/2020 at 5:41 PM #117005
  Vonasi

  Francesco wrote:

  Don’t you think a backtest beyond 5 years is useless?

  Far from useless as Mauro’s 200k back test shows us. Basically the strategy is a fit to a certain amount of data and it goes sideways before that and virtually sideways on recent data. The probability of it performing well going forward is unlikely.

  1 user thanked author for this post.

  Francesco

  

  Vonasi

  Moderator

  Topics: 192

  Replies: 5088

  Been thanked: 1811 times
  01/16/2020 at 6:31 PM #117008
  Paul

  @Vonasi you’re right. If a certain hour was suitable for a certain period, it could be shifted. Perhaps a quick forward test with many OOS data with would make it clear.

  In any case this posted pic above is the opposite of my other strategy 1 hour TF to show your point,

  

  Paul

  Participant

  Topics: 40

  Replies: 812

  Been thanked: 660 times
• Author
  Posts