StockFetcher Forums · Filter Exchange · The Ultimate Filter<< 1 2 >>Post Follow-up
268 posts
msg #127771
Ignore novacane32000
3/30/2016 6:22:20 PM

I have yet to find a filter that works consistently year after year under every condition.

Some filters work well in trending markets, others in sideways markets but if I could have just one filter it would be a predictive filter for SPY .

If you know the overall direction of the market and can see when turning points are coming, you could argue for not needing any other filter.

I looked at a few SPY filters from the past on here which included --Advancing -Declining ratios, SARS and the VIX but after backtesting all three they did not work very well--too many whipsaws.

Any suggestions?

805 posts
msg #127783
Ignore marine2
3/31/2016 3:03:00 PM

There really is no ultimate filter only in the eyes of the beholder. The market, and it's equities supporting it respond to general indicators and there really isn't a perfect filter predicting 100% success each and every time. But, I do have a pretty solid trend filter that used with SPY (S&P 500 ETF) can give you very nice results. Hopefully this helps you predict with better results in what the S&P 500 will do. Here is the filter:

close above 10
and weekly slow stochastic(5,5) %K is below 70 previous 1 day
and weekly slow stochastic(5,5) %K is crossing above weekly slow stochastic(5,5) %D previous 1 day
and slow stochastic(15,3) is below 75 previous 1 day
and aroon oscillator(25) is below 0 previous 1 day
and ma(20) is below ma(40) previous 1 day

268 posts
msg #127795
Ignore novacane32000
4/1/2016 7:07:48 PM

so you buy these stocks when SPY is trending up?

805 posts
msg #127802
Ignore marine2
4/2/2016 8:05:35 PM

I'm assuming you wanted to buy into SPY when it showed a very good percentage that it was now about to trend upward. You would then use the filter code I provided in the previous communication I sent in this thread.

If not, and you wanted a filter that would tell you when SPY was about to trend downward then you would write code doing the opposite. Where the Weekly Slow Stochastic(5,5) %K is crossing above 80, etc., etc.

805 posts
msg #127808
Ignore marine2
4/3/2016 3:52:33 AM

Correction on the downward trend, I meant to say the Weekly Slow Stochastic(5,5) %K is crossing below 80 previous 1 day.

86 posts
msg #127828
Ignore jimvin
4/4/2016 2:13:40 AM

For what it’s worth...
I’ve tried a little over 400 stock models here over the last 5 years and found none that survive the “death cross” Market correction.
In addition, as marine2 noted, the “perfect model” like the “perfect mate” will always be a matter of subjective preference.
My preference (for stock models) is one that offers a reasonable return on a maximum 2-day turn-around (M/T) with gains well outpacing losses.
That said, one of my favorite filters is based on the work of the long-silent estetson (last post circa 2007).
The original is given below and can also be found here:

Fetcher[set{avgliquidity, Price * Average Volume(30)}
set{avgliquiditymil, avgliquidity / 1000000}

set{signtoday, sign(Day Change)}
set{signyesterday, sign(Day Change 1 day ago)}
set{sign2days, signtoday + signyesterday}
set{normalizedsign2days, sign2days / 2}
set{relativechange, Day Change / Average Day Range(30)}
set{consecutivedays, relativechange * normalizedsign2days}
set{baddayprelim, signtoday * signyesterday}
set{badday, baddayprelim - 1}
set{normalizedbadday, badday / 2}
set{nonconsecutivedays, abs(relativechange) * normalizedbadday}
set{alldays, consecutivedays + nonconsecutivedays}
set{dailygainpctavgrange, cma(alldays,100)}
set{consecutiveup100, count(sign2days > 1, 100)}
set{consecutivedown100, count(sign2days < -1, 100)}
set{trendiness100, consecutiveup100 + consecutivedown100}
set{consecutiveup20, count(sign2days > 1, 20)}
set{consecutivedown20, count(sign2days < -1, 20)}
set{trendiness20prelim, consecutiveup20 + consecutivedown20}
set{trendiness20, trendiness20prelim * 5}

Price > 2
avgliquiditymil > 1
Average Day Range(30) > 4

add column avgliquiditymil
add column Average Day Range(30)
add column trendiness100
add column trendiness20
add column dailygainpctavgrange

draw trendiness20
draw cma(trendiness20,10)
draw trendiness100
draw cma(trendiness100,10)
draw dailygainpctavgrange

sort by column 7 descending

In use I modify it slightly in my application to fit some of my additional trading preferences.
I also use a Pring KST overlay from in doing my analyses of the selections it returns
Statistics for the last 48 weeks of play are:
Average gain per trading round: .80% (that’s point-8%, not 80%)
Average Gain: $2,949.64
Average loss: -$1,779.79
Accuracy: 57%
Experience shows that I will on average only achieve about 50% of the actual closing gain (or loss) since the big players wilt get their millions already made will get trades serviced first.
Numbers here reflect playing with 200K in a Roth IRA, in Monday at Market open and out no later than Tuesday at 3:30 acknowledging the free-rider rule so trading can begin again on the following Monday.
Before anyone dismisses the apparently pitiful average gain, run a spreadsheet calculating 100K at .8% gain compounded for 10 years; the result should be a final tally of a little over 5 million.
One of the keys is the non-taxable Roth IRA noted above; without that taxes will crush compounding.

268 posts
msg #127850
Ignore novacane32000
4/5/2016 2:52:12 PM

Not sure how you track and trade all those results. Do you sort them and then look for what?

268 posts
msg #127851
Ignore novacane32000
4/5/2016 2:54:30 PM

Looking for a simple filter to track the movement of the overall market as represented by SPY.

Nothing will be perfect but maybe something comes close.

805 posts
msg #127858
Ignore marine2
4/5/2016 4:49:16 PM

As a suggestion, you might try going into the search field for past threads and see what has been recorded so far on this question you have. I am sure someone has answered this before. Or, just keep plugging away at creating more filters that hopefully one of them will be your answer to the truth. May your investing be rich and your intentions be noble. Trade On fellow investor / trader.

268 posts
msg #127859
Ignore novacane32000
4/5/2016 6:17:03 PM

I did that search before posting this message. I found 3 that I backtested on excel and none of them worked.

StockFetcher Forums · Filter Exchange · The Ultimate Filter<< 1 2 >>Post Follow-up

*** Disclaimer *** 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. does not recommend particular securities., 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 2016 - 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