SCTR dramatic change, reason ?

efaraldo

About a week ago, Chip Anderson had an article titled "Using SCTRs to Narrow your search for winning Stocks". In it, he had the following scan: [5 days ago max(15,SCTR) <20] and [todays SCTR >40]
I programmed this into a scan and had four results, one of which was SNSS with the following chart: http://stockcharts.com/h-sc/ui?s=SNSS&p=D&yr=0&mn=8&dy=0&id=p03426773224&a=402779875
As you can see, the stock has been in a fairly narrow consolidation for about 90 days with nothing much happening, EXCEPT for SCTR being around 10 for those 90 days and then, within 3 to 4 days, it shoots up to about 95.
I did a little research on the stock, and found no recent news.
My question is, how can SCTR jump that high so quickly, without a corresponding price increase or other sort of visible change?

Thank you

E Faraldo

markd

Hi @Kevo,

Well, if you look at the chart school articles for the two indicators you can see how differently they are calculated.

ROC works with only two values - the current close and the close 125 days ago. If either value changes dramatically, the result will change dramatically. So it would get a negative value for quite a long time after the big drop, because the beginning value is still marching up the bars of the preceding rally. Then after it gets to the last bar of the rally and drops way down to the next bar at the bottom of the drop, the result is suddenly positive (because price has been climbing from that low), so the indicator pops up.

Slope on the other hand works with all the values in the 125 day period, plugging them into the formula for a Raff Regression Channel. The RRC calculates an imaginary line that is a "best fit" for all the data points - meaning a line that results in the least total difference between that line and each data point. That line very gradually changes direction as old data drops out and new data pops in. That's because even if a data change is very dramatic, the effect of each data point is diluted by the 124 other data points. Slope is the rise over run of the RRC line. It changes very gradually also, unlike ROC, which as explained, suddenly pops when one value (the start, in this case) changes dramatically as the oldest data point drops out.

markd

Here's a link to the chart school explanation of the sctr calculation:

https://stockcharts.com/school/doku.php?st=sctr&id=chart_school:technical_indicators:sctr

If you look at the ROC(125) for this chart, you can see it suddenly jumped up as the fall off the cliff drop in October moved out of the calculation. There may be other things, too, but the ROC looks alot like the SCTR indicator, so I would guess that's the main reason.

Kevo

Great question @efaraldo and thanks for asking. It does look strange to suddenly jump like that. I am trying to figure it out. I have nothing definitive, but here is what I have so far:

Even though the long-term trend of SNSS is down, I noticed that many technicals have moved positive or a setup triggered among all the scales (Monthly, Weekly, Daily). Maybe the long-term trend might turn up.

Since SCTR is a Relative Strength measure within a Universe, here are more clues to the dramatic rise. In the days before April 2, there were no huge changes in Relative Strength among the Small Cap Sectors. The dramatic changes started after the April 2 close:

Small Cap Consumer Staples, Financials, and Utilities have dropped dramatically in Relative Strength since April 2, while Small Cap Healthcare has stayed relatively strong.



Here is another view:

Small Cap Utilities and Small Cap Consumer Staples have moved to lagging from leading, while Small Cap Healthcare stayed relatively strong.



Also, SNSS has not lost much Relative Strength in the Small Cap space:




@markd that is strange how ROC(125) jumped. Slope(125) gave a clue the whole time.

Kevo

I was trying to figure it out via @gord 's scanning trick, but with so much detail, i could not reach any conclusion.

I know that the SCTRs in the images above are in the ETF space and not the Small Cap space, but I figure the Small Cap stocks in those sectors that are in the Small Cap space would correspondingly lose Relative Strength.

SNSS is in Small Caps => Healthcare => Biotechnology

efaraldo

Many thanks to Markd and Kevo for their very lucid, and detailed answers.

Taking into account that I am a novice on these issues, I would agree with Markd that the calculations for ROC(125) would seem to account for the sharp rise of SCTR, since it is one of the long term indicators used in SCTR.

Using the Cycle line tool, there are exactly 125 days between the sharp drop and the sharp rise of SCTR. To me, this validates Markd's assumption that ROC(125) is the reason

If you plot SNSS daily chart and the following indicators: SNSS vs $SPX ratio, SCTR and ROC(125), as Markd suggested, this becomes very clear.

Now, does this mean that there is a glitch in the calculation of SCTR? I address this question to Chip Anderson.

P.S. I have some money riding on this issue.

Kevo

Thanks @markd and @efaraldo ; It looks like I did not make a mental note of the gap down when it came to examining ROC. I guess I am not familiar with ROC behavior after a gap.

I see that ROC(125) is 30% of the SCTR calculation. So, the SCTR spike is directly related to the ROC(125) spike like you said at the start.


SCTR has 2 parts: an indicator score, then the technical rank. When looking at the SCTR, we see the technical rank, not the indicator score. I first thought that SCTR changed because something in it's universe changed that altered it's rank. In the case of SNSS, I see this is not the main reason.

@efaraldo : Relative Strength is just one metric of TA and should not be used alone.

Just a few things I am wondering about:

Maybe when you use SCTR, you should plot it in an indicator panel to look for spikes, which would suggest a gap influenced its calculation. Even though it suddenly shot up, this does not mean it is now all of a sudden strong? I think this is what @efaraldo is getting at.

Is it coincidence that the 3 small cap sectors (mentioned in my comment above) lost SCTR (Relative Strength) at the same time, and what impact did this drop have on the technical rank of SNSS?

The following statement is not documented, but is it true? => The SCTR can rise on a stock just by other stocks dropping in rank? I think this would be true for stocks whose SCTR indicator scores were above SNSS, but those SCTR indicator scores are now below SNSS.

Kevo

SCTR considers long-term to be 6 to 9.5 months...

markd

Hi @Kevo ,

"The following statement is not documented, but is it true? => The SCTR can rise on a stock just by other stocks dropping in rank?"

Well I think not exactly. I think the SCTR score (score, not rank) is what matters. Rank is calculated after the scores.

So if yesterday's SCTR scores are:

XYZ 812
TUV 735
QRS 665
NOP 550
KLM 440

And today's scores are

TUV 780
XYZ 720
QRS 663
NOP 540
KLM 430

All the scores have fallen, but none so much that the symbols change rank.

But, if yesterday's SCTR scores are the same:

XYZ 812
TUV 735
QRS 665
NOP 550
KLM 440

And today's scores are

XYZ 720
TUV 700
QRS 663
NOP 540
KLM 430

Again, all the scores have fallen, but this time TUV's score falls below XYZ's score, so XYZ moved up in rank. Note it's not necessarily that TUV's score change was more or less than XYZ's score change, it's where each one ends up after the scores are sorted.

Kevo

Hi @markd Thank You for the comments. I feel like I am missing something after your last comment. Now I think we both might be saying the same thing. Is there a flow error in your last comment?

If a symbol with a score above SNSS yesterday now has a score below SNSS today, while SNSS maintains the same score, then the rank is now higher for SNSS even thought the score may be the same. I made the following comment above:

I think this would be true for stocks whose SCTR indicator scores were above SNSS, but those SCTR indicator scores are now below SNSS.

I made this comment after my premise:

The SCTR can rise on a stock just by other stocks dropping in rank?

Are you saying the same thing?

One case: SNSS only becomes affected when another symbol transitions across the SNSS score. The other cases: Score changes that occur above or below SNSS would not affect the rank of SNSS.

I understand the rank is determined after the score. I am using the term score in the quoted comment.

Please pardon me for beating a dead horse.

markd

I think we agree. Just a difference in phrasing.