StockFetcher Forums · General Discussion · Do I dare utter the word "holy grail"?<< 1 2 >>Post Follow-up
BlackBars
54 posts
msg #140789
Ignore BlackBars
1/10/2018 9:53:59 AM

Using historical and statistical probability, I think I narrowed it down...

So over the past couple weeks I did some extensive autosearching using Stratasearch and I THINK I have started a multi system that yields plays of 100% accuracy each trade. (DON'T CRINGE, hear me out)

What I did was implement a very bullish momentum criteria on generic autosearch and let it run through 100,000's of combinations from the time period of 2012-2016 and I strictly chose results from it that fit the following criteria:

-Percent Profitable: 100%
-Number of Trades: 10 or more 
-Average days held: 1-3 
 
Any combinations that yielded those paramters were then tested in the year 2017 to see if they were STILL consistent with 100% accuracy and I turned the combinations that stayed 100% accurate into signals.  So after a while I gathered at least 2 dozen signals (SO FAR). Now the downside unfortunately is that new signals are generated RARLEY (some are like 10-20 trades in the span of 5 years) but according to those past years (including the year that was manually tested) they have each been statistically 100% profitable which I found a bit bizarre. Some even have been 100% profitable in the past 10 years.

Because of how rarely these signals are generated, I believe it will be a while until I find out if they stay accurate in realtime going forward. In the meantime obviously I am gradually collecting more combinations that fit the criteria. So regarding the signals I already have in my collection so far, are there any experts in this statistical feild that know any method I can do to confirm or validate if this whole thing is legit without having to wait a month or two for a signal to generate? - Not trying to emerge as impatient, I'd just like to be productive while waiting. 


Also if anyone feel the urge to post something regarding "no such thing as a home run" please resist because this is not what this is about. 


nibor100
434 posts
msg #140792
Ignore nibor100
1/10/2018 10:42:30 AM

@Blackbars,

Regarding your statement "Some even have been 100% profitable in the past 10 years."
Does that mean you tested the 20 signals over the past 10 years in StrataSearch and only some of them met your 100% profitable goal?

Did your run any backtests in SS longer than 10 years?

What stock universe are you testing over, S&P 500, Russel 2000, or bigger?

Thanks,
Ed S.





miketranz
783 posts
msg #140794
Ignore miketranz
1/10/2018 10:51:10 AM

Why not drop the percentage down to 80? You'll get more hits,plus that's a very desirable number.Legendary mutual fund trader Gil Blake created a system with 70%,in which he stated,it's very hard to have a down year at that percentage ratio..

shillllihs
4,880 posts
msg #140798
Ignore shillllihs
modified
1/10/2018 11:21:03 AM

100% of leverage inverse ETFs drop.
100% of inverse ETFs gaps up have filled.
Every stock will eventually go to zero 100% of the time.
2 trades a year for 1-3 days in a bull market is not inspiring. Create something that worked in 2008-2009.


Create a system to short both long and inverse leveraged ETFs, with a reasonable drawdown, with clearly defined entries and exits.
Decide if they only work in a bull market or both.
Ultimately, find a system that worked Jan.2008 thru May2009

I've made a killing shorting inverse ETFs, you just have to know the right time to do it.

BlackBars
54 posts
msg #140799
Ignore BlackBars
1/10/2018 11:29:10 AM

nibor100,

After configuring all of the current signals I did happen to backtest all of them even further than 10 years. The majority simply do not yield any trades before the initial 5 years. The ones that do have trades further than 5ish years (which are actually very few) do remain 100%. One of the signals even goes back 15 years prior but has only 49 trades in that 15 year span.
And to answer your last question, I use the S&P 500 Index. I am sure bigger indexes could work but because of my computer speed, I would not receive as many combinations as fast as a smaller index would offer then a bigger one would.

miketranz,

Thats a great point, but keep in mind if I end up accumulating enough 100% combos to the extent of being able to generate a signal each night for a trade, AND if this 100% strategy ends of being validated , this would be game changing, would it not? At the moment this could be seen as a fantasy but I kinda want to dig deeper for kicks.


miketranz
783 posts
msg #140802
Ignore miketranz
1/10/2018 11:48:02 AM

Hey,100% works for me,I also like the 1-3 day hold idea.Just have to see if you can get enough signals to make it tradable on a daily or weekly basis...

BlackBars
54 posts
msg #140806
Ignore BlackBars
1/10/2018 12:17:21 PM

shillllihs,

"2 trades a year for 1-3 days in a bull market is not inspiring"

In retrospective I might agree with that, BUT some of the combos worked even BEFORE the 08' market AND AFTER. Think about it like this; if I found a few that worked like so, I WILL find more. Its just a matter of time.

And I agree with you on the circumstance of finding a combo criteria that would do just as well in a bear market. I feel strongly over the fact I will come across one that has the criteria of 100% accuracy that had a trade in between that bear market period. Once again though it is just a matter of time.

Also I do not doubt the success that the ETF system you mentioned could achieve, I just personally feel if this statistical probability system that I have going on ends up being validated and I assemble enough signals to get some plays every week then I feel this would be game changing.

nibor100
434 posts
msg #140808
Ignore nibor100
1/10/2018 12:33:48 PM

@BlackBars,

Using the S&P 500 stocks might be one reason you are finding so few trades the further back you go.
Out of the current group of S&P 500 stocks, only 431 belonged to the S&P 500 in 2012, and only 316 belonged in 2002.

This won't be an issue if somehow your S&P 500 data in SS has the full compliment of stocks belonging in the S&P 500 at any given point in time in the past, but I suspect that is not the case.

If you post a couple of your 20 SS filters that you feel might be translatetable to SF code, I can take a shot of coding them into Edgerater Pro sw and run a backtest over a larger database of stocks with 20 years of historical data to see if the 100% holds up.

Thanks,
Ed S.

shillllihs
4,880 posts
msg #140809
Ignore shillllihs
modified
1/10/2018 12:39:58 PM

Yes do that. Compile a list of all systems with a 100% returns and make a super system on SF.
Just make sure it's worked in a bear market.
With SPY, you can go back and check 2 bear markets.
I have always based my trades using 1 time ETFs like
like Spy, Ung, Ure, Uyg, Vnq, Oil, Gld, Slv, etc, then you can convert them to 3x ETFs.
I started a thread a few years back trying to compile a list. of high win systems with few returns.
It's a great idea. It just has to work across all time frames, unless you want to sit out a bear market.

BlackBars
54 posts
msg #140810
Ignore BlackBars
1/10/2018 1:15:19 PM

Tad problem, there are components that are factored in the signals such as "sector strength" or "index volatility" which prevents that specific SS code from being translated into Stockfetcher's code.
Also I'd be happy to share with you an example of one of the signals, but by all means I do not want to post it publicly. You can shoot me an email however since that is a more private way to allocate information.

StockFetcher Forums · General Discussion · Do I dare utter the word "holy grail"?<< 1 2 >>Post Follow-up

*** Disclaimer *** StockFetcher.com does not endorse or suggest any of the securities which are returned in any of the searches or filters. They are provided purely for informational and research purposes. StockFetcher.com does not recommend particular securities. StockFetcher.com, Vestyl Software, L.L.C. and involved content providers shall not be liable for any errors or delays in the content, or for any actions taken based on the content.


Copyright 2018 - Vestyl Software L.L.C.Terms of Service | License | Questions or comments? Contact Us
EOD Data sources: DDFPlus & CSI Data Quotes delayed during active market hours. Delay times are at least 15 mins for NASDAQ, 20 mins for NYSE and Amex. Delayed intraday data provided by DDFPlus