StockFetcher Forums · View by Author: (297 messages)  [ Display By: Date / Subject ]<< 1 2 3 4 5 ... 30 >> 
General Discussion · sectors and industries - ranking by momentum?
msg #156532
4/9/2021 9:47:37 AM

Stockfetcher has Industries:
Stockfetcher has Sectors:

But there's no way to (momentum, relative/absolute) rank them?

It should be easier to:
-assess the market regime/environment
-identify momentum in sectors and industries (and rank them)

This kind of functionality should be native to Stockfetcher. There should be more statistical tools, and it should be possible sort filter results by some broad momentum criteria tied to the industry or sector itself. Then it should be possible to scan the components of those outperforming sectors/industries.

In TC2000, this is easy to do. You have a watchlist of industries or sectors. You sort by, say, ROC20. You click on the industry or sector row. Another watchlist is populated with every symbol in that sector. You sort THOSE by, say, by ROC20, or filter them by breakout. Now you're buying very tickers that are actually _providing the momentum that makes that sector or industry the leader_. WHY isn't something so (proven, logical, readily available in data) a bedrock feature here?

Why does this matter? Momentum is one only factors available to retail traders for consistent outperformance. So why aren't retail-grade tools making this easier?

Does anyone have a way to do this, kludge or not? I have tried using indexes as the proxy, but the problem is that the components are not provided.


Filter Exchange · Help transitioning from TC2000
msg #156531
4/9/2021 9:35:18 AM

Also, this TC2000 breakout code looks similar to the "double bottom" pattern output on SF, but it also appears to have some continuation after the last resistance level. Just eyeballing.

Here's this morning's output using ETFs:

Symbols from TC2000

Filter Exchange · Help transitioning from TC2000
msg #156530
4/9/2021 9:17:31 AM

Ed, there is lots of TC2000 code not in the forums, however (and it's shame they shut the forums down). Bruce in the forums was an absolute legend, one of the most competent "bridges" between the customer and a software platform that I've personally come across in decades of using commercial tech.

I have a lot of TC2000 code and there are snippets throughout the web.

General Discussion · MORE than 50 trades a month?
msg #156149
3/15/2021 9:51:41 AM

more frequently, yes

General Discussion · Does the Candlestick Matching Tool really exist?
msg #156143
3/14/2021 1:17:27 PM

Stockfetcher used to have something called "user definable candlestick patterns." I think it ran on an Applet?

There was this:

And there's this current tool:

General Discussion · MACD
msg #156136
3/13/2021 11:25:22 AM

I think a much better idea is to post your code in full and allow people to help you improve it. If you consider your options, this would be the much more enriching choice.

General Discussion · How to count appearances in a filter with ranking? xarlor?
msg #156131
3/12/2021 7:19:59 PM

This further explains why this count is needed. See how VDE suddenly appears among the "mainstays". Soon thereafter, a few days, VDE was among the top 3. And it has been hovering there since.

The main problem with my solution is that the "count" rolls-forward with the counting lookback window, in this case 100 days. So, "31" will not necessarily increment to "32" on the next day, since the end of the window moves forward one day and the new window might not contain a hit. Therefore, you might not see the numbers increment until several days later -- at least this is why I think it happens this way.

Still, progress kind of. Thanks to all who shared help along the way.

General Discussion · How to count appearances in a filter with ranking? xarlor?
msg #156130
3/12/2021 7:13:23 PM

I want to share the solution I came up with for counting hits on a filter where there aren't conditions. It's based on what xarlor shared. And my version very simple so it probably could be enhanced.

This is a rotation filter, so it is important for me to know how long a ticker has been appearing, or if it is newly among the highest-ranked symbols.

The way I solved it is to simply count the ranking/sorting column if it has value greater than 1. You can change the target value to identify tickers that are popping quickly, or to see those that have been around a lot. The base code is commented in the "hit machine" comment line.

I'll leave the filter as I use it. I tend to copy/paste from the commented-out criteria rather than create separate filters.

apply to watchlist(vanguardETFs)
apply to watchlist(ETF20215k)
apply to watchlist(indexes2021)
market is ETF
market is not OTCBB
Average Volume(30) > 1234567
apply to symlist(bnd, dbc, veu, vnq, vti)
close crossing above ma(20) within the last week

apply to watchlist( )

simple rotation filter -- dos the code match the rules?
fotchstecker 3/8/2020 1:15:53 PM
System -- Buy the Rank Score leader
1. Overall Rank Score = ( 20-Day Return ) *.5 + ( 3-Month Return ) *.5
2. Close > Average( Close, 100 days);

set{20Return, ROC(20,1)}
set{ROCrank1, 20Return * .25}
set{60Crank, ROC(60,1)}
set{ROCrank2, 60Crank * .75}
set{ROCrotate, ROCrank1 + ROCrank2}
sort on column 9 descending

ADD column ROCrotate 3 months ago
ADD column ROCrotate 1 month ago
ADD column ROCrotate 1 week ago
ADD column ROCrotate

/* hit machine -- for filters that use only sorted columns and no conditions, count by the sorted column hits */
set{x1,count(ROCrotate > 0,1)}
set{long1,x1 * 1}
long1 > 0
set{hits,count(long1 > 0,100)}
add column hits {hits/100days}

and draw MA(100)
/* and close > MA(100) */

add column Historical Volatility(30,1)
add column Historical Volatility(100,1)

add column relative strength(spy,30) > -1

set{ma100a,count(close > ma(100),1)}
ADD COLUMN ma100a {above ma100}

set{ma20a,count(close > ma(20),1)}
ADD COLUMN ma20a {above ma20}

set{20x, count (close crossing above ma(20) within the last week,1)}
add column 20x

and add column industry
and add column sector

draw ma(20)


General Discussion · MACD
msg #156127
3/12/2021 3:43:36 PM

Thanks, Cheese. You always bring good vibes around here.

General Discussion · MACD
msg #156118
3/12/2021 9:19:46 AM

Mac, to add to your comments:
"Don't use real money until you have a system that win's more times than not and even that isn't perfect."

You can have a low winning percentage only if your average winners outweigh your average losers, but low winrate systems require even more planned sizing and money management. You have to take your acceptable loses in a long enough string to win over time.

Trading = a series of a trades.

I recently started trading a completely new way using Stockfetcher and it is related to the above. I export signal lists and take a large number of trades every day using software with the targets already configured and running on 1 minute data. The targets are based on an involved process of looking at the 40-day forward performance of every ticker that showed up starting 6 months ago. I posted about this elsewhere in the forums.

It's the main reason I keep asking Stockfetcher to completely revise the analysis component of this tool. There are two giant gaping flaws with SF:
1. Signal performance based on close when it should be calculated from that close but also the following day's OPEN (2 different metrics), since most people will use signals generated at night the prior day.
[related, if you ever calculate performance using the following OPEN, not the prior CLOSE like SF does, you will be quite surprised at the difference the overnight makes. Beware!)
--->I know one user is entering limit orders after close in the after-hours market to deal with this. This is *sort* of a solution, but a non-specific one because the prices filled after close and before open have a high degree of variance. It's not clear whether doing this is more like the close or more like the open. So while it's a way to 'get orders in' it's not really clear, analytically, whether one is trading in a way that aligns with close prices for signal generation day. Also, it's a pain in the ass and can't easily be done programatically. I don't want to enter 10-50 limit orders each night if it's not clear that doing so will make me fabulously wealthy. ;)

2. A true testing tool that essentially does what I mention above -- it gives true performance metrics for any given ticker generated by the filter going forward from signal date to some inputted end-date expressed in number of days. So, Ideally, you should be able to pick a 'start date' for the analysis. The tool will output every signal generated and, using OHLC data, provide average peaks, lows, and days to peak and lows. The user should have an idea of where to take stops, how long to hold, and where to scale out profits. All of this data is READILY available and is used by SF for calculations RIGHT NOW, and it would not be a heavy lift to drastically improve the usability of this resource. It's not like I'm talking about building a radical backtesting engine.

StockFetcher Forums ·  · << 1 2 3 4 5 ... 30 >>

*** 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 2022 - 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

This site is protected by reCAPTCHA and the Google Privacy Policy and Terms of Service apply.