Author Topic: solid strategy or curve fitting?  (Read 6646 times)

philbee

  • Newbie
  • *
  • Posts: 2
  • Karma: +2/-0
solid strategy or curve fitting?
« on: August 09, 2013, 02:29:52 am »
I'm about two months into active trading and wanted to share a methodology I'm trying to create strategies.  Would appreciate feedback... in particular I'm trying to determine if this back tests well only because I'm curve fitting stocks that behaved a certain way in the past.

At first, I had problems coming up with strategies that performed as well as the ones I obtained via execution consulting even playing with the optimization features.  But then I tried the simple built in bull and bear strategies that were slightly positive by themselves, and optimized for the most profitable symbols for those strategies.  I basically looked for symbols that had at least 5 trades in 30 days with 70% or higher winning percentage.

For example, here is a strategy that tests well for the past 30 days:

http://www.trade-ideas.com/Cloud.html?code=873e18f5539fdd7a0284d8fed714bfd4

It is the "Turbo Break Up" built in alert window strategy with the following tweaks: volume minimums, spread maximums, and filters to only fire alerts when the market is up for the day and for the past 15 minutes.  For the past 30 days, 26 symbols are 80%+ winners, yielding 110 trades, with a cumulative winner percentage of 75.5%.  (I realize profit factor is actually more important than winning percentage for total return, but as a new trader, I want a high winning percentage because that will let me realize if I've got a faulty strategy more quickly when comparing actual results to back testing.)  By using this methodology and tinkering around with Oddsmaker for a few hours, I can now produce alert windows that back test similarly based on quite a few of the pre-packaged strategies.  My trading in this particular strategy has been somewhat successful, but not enough trades to be statistically significant yet. 

I've attached the OM output for this strategy since 1) I'm not sure if the URL above will work since the strategy uses a symbol list in my account, and 2) the performance of this strategy will degrade over time if the symbol list is not updated frequently.

Anyway, if anyone has any questions or suggestions for improving this strategy, would love to hear it.  I'd also like to hear about other methodologies that users have found in developing their own strategies.



TradeIdeas1

  • Administrator
  • Sr. Member
  • *****
  • Posts: 272
  • Karma: +6/-0
  • You got to know when to hold em, know when to fold
    • Trade-Ideas
Re: solid strategy or curve fitting?
« Reply #1 on: August 09, 2013, 10:21:29 am »
Phil,

That strategy looks very good the only issue is that other people can't see the symbol list.  So when people look at the cloud link it shows up as blank because your symbol lists are private to your user account.

philbee

  • Newbie
  • *
  • Posts: 2
  • Karma: +2/-0
Re: solid strategy or curve fitting?
« Reply #2 on: August 09, 2013, 12:29:26 pm »
Ah thanks... I was wondering how to share strategies with symbol lists.  Below is the list as configured for the attached OM output.  I've been "re-optimizing" the symbol list every couple of days.


APU
AVAV
BIN
CKP
CTRN
DEPO
DHI
GS
HCI
HNI
HPF
IIVI
IRET
JNPR
LF
MFLX
MRTN
NLS
OWW
PCTI
PGTI
PJC
PM
PMTC
QEP
SCLN
TA
TRR
UEPS
UMH

Rockford

  • Newbie
  • *
  • Posts: 18
  • Karma: +0/-0
Re: solid strategy or curve fitting?
« Reply #3 on: July 01, 2014, 11:22:02 pm »
Phil,

If you don't mind me asking, how did you search for the symbols that produced 5 or more trades in the last 30 days? I'm using Trade-Ideas from Scottrade.

Also, it seems that 'up for the day' and for the 'past 15 minutes' are alerts rather than filters, unless I'm not understanding what you posted properly.

Thanks,
Rockford
« Last Edit: July 01, 2014, 11:25:51 pm by Rockford »

Rockford

  • Newbie
  • *
  • Posts: 18
  • Karma: +0/-0
Re: solid strategy or curve fitting?
« Reply #4 on: July 02, 2014, 07:40:20 pm »
Philbee said this on his original post above ------> I basically looked for symbols that had at least 5 trades in 30 days with 70% or higher winning percentage.

Can anyone tell me how he determined the Winning Percentage of these trades?

I'd really appreciate it, thanks,
Rockford

Barrie Einarson

  • Administrator
  • Hero Member
  • *****
  • Posts: 610
  • Karma: +20/-1
Re: solid strategy or curve fitting?
« Reply #5 on: July 02, 2014, 08:03:26 pm »
Hi .. he used the Oddsmaker - for you, the Backtester.

A webinar on the Oddsmaker:

https://www.youtube.com/watch?v=98pKiIUNqaE&index=6&list=PLNtesvHL0KGLaMCWvViZUSfCs74yGazZh


Cheers

Barrie Einarson
barrie@trade-ideas.com
Trade Ideas Trading Room Moderator
http://forums.trade-ideas.com/

TradeIdeas_DA

  • Administrator
  • Hero Member
  • *****
  • Posts: 728
  • Karma: +9/-2
    • Trade Ideas LLC
Re: solid strategy or curve fitting?
« Reply #6 on: July 03, 2014, 12:04:42 am »
In fact a whole list of videos about the OddsMaker, of various lengths, can be found here:

https://www.youtube.com/results?search_query=Oddsmaker+trade+ideas

Follow me on Twitter: http://twitter.com/TradeIdeas