Price Ratio, Price Comparison and Price Differential

Chart Forum » Indicators » Price Ratio, Price Comparison and Price Differential

«« Previous Next »»

Thread Start New Thread
Last Poster Posts Pages Last Post

Author

Message

colin_twiggs

Member

Post Number: 671
Registered: 09-2002

Rating: N/A
Votes: 0


Tuesday, September 23, 2003 - 10:46 am:

At present the Price Ratio, Price Comparison and Price Differential indicators are calculated using the ratio of closing prices between two securities/indices at the start of the chart. This has confused a number of users because the ratio varies with each chart time period selected. We are considering changing the formulae, to use the ratio of closing prices at the start of the data: the first day on which both closing prices are quoted.

If you would like to express your opinion on the subject, please post to this thread.

Colin

hazluk

Member

Post Number: 75
Registered: 09-2002

Rating: N/A
Votes: 0


Tuesday, September 23, 2003 - 11:45 am:

Colin: If I understand the issue correctly, I definitely would NOT want it changed because..

"...the first day on which both closing prices are quoted." can already be easily obtained by simply clicking on the "Scroll to Beginning" button.

Only $12 (USD) per month

mosaic1996

Member

Post Number: 213
Registered: 01-2003

Rating:
Votes: 2


Tuesday, September 23, 2003 - 12:00 pm:

I would like to compare relative performance from a particular point in time. So keep it as is in preference to the 'first common data point'. This is how it is presented in all other packages that I have used.

Another option would be to include a parameter to set the start date for comparisons. Options for 'comparison start date' would be
a) start of chart (current)
b) specify a start date
c) first common data point. This option doesn't make sense to me as you wouldn't really know from which date the comparison starts.

In general I would use option a) most of the time, option b) often, and option c) never.

Cheers,
Graeme

colin_twiggs

Member

Post Number: 672
Registered: 09-2002

Rating: N/A
Votes: 0


Tuesday, September 23, 2003 - 04:58 pm:

Graeme & Gino,

For Relative Strength/Price Ratio, the start date is not really significant. What is important is the slope of the line.

With IC, the Price Ratio starts at zero, at the start of the chart. The Ratio is re-calculated every time that you change the Time-Period, to keep the start at zero. This is unnecessary. Compare the two BSO charts below.

On the 2-year chart the point marked [!!] is at 5%.

2year

On the 6-month chart the same point is at -3.8%.

6month

The ratio has been re-calculated because the chart time-period has changed. I don't see it as that important that both charts start at zero.

What I see as more important is: (a) new users are confused by the shifting indicator line, and (b) re-calculation prevents IC from adding moving averages and trendlines to the indicator.

Metastock calculates Relative Strength (Comparative) based on the the ratio at the start of the data. Mansfield Charts (used by Stan Weinstein) appear to do the same. I believe that we should follow suit.

The Price Comparison indicator merely adjusts the price line of a second security to start at the same point on the chart as the first. It is more important for this indicator to start at a common point - to fit the two lines to the same price scale.

If necessary, we may be able to offer both options: (1) start of data; and (2) start of chart (the present option).

Colin

(Message edited by colin_twiggs on September 23, 2003)

mosaic1996

Member

Post Number: 214
Registered: 01-2003

Rating: N/A
Votes: 0


Tuesday, September 23, 2003 - 09:09 pm:

Hi Colin,

A new related requirement!
I thought that I would be able to used the Zoom function to demonstrate what I saw as a potential problem with your proposal (assuming that I am understanding it correctly).

I was surprised to see both price comparison and price ratio recalculated to start at 0 when I zoomed in. It is quite a bizarre visual as you scroll left/right with the relative performance measures being recalculated continually. I could understand why people would complain about that. To my mind I see the Zoom in function as a way to have a closer look at a period of interest.

I would like to see relative performance measures anchored at the start of the time period selected.

Now having said that, I can see a real advantage in maintaining the status quo. Specifically, if I wanted to check relative performance of stocks, sectors, etc. during a previous similar part of the economic cycle. In other words, zoom into a middle section of the chart. Could be useful for timing switches between sectors, or to identify likely leaders from a sector, or sectors that lead out of a bear market, etc. based on past cycles.

Now there's an each way bet!

Your Proposal!
My understanding of what you are proposing is that you will have relative performance calculated from the first data point that the selected stocks have in common. This means that the relative performance measures would always be the same at the same point in time irrespective of the time period displayed.

I have attached 5 charts showing relative performance.
SMS vs SOH and KAZ = 3 charts
a) real long term before KAZ existed
b) medium term
c) short term (the period in which I am interested)

WBC vs NAB, ANZ, CBA
a) long term
b) short term (the period in which I am interested)

I believe that it is a lot easier to ascertain the relative performance over the short term period that I am interested in using the current system than under the proposed changes. Whilst the compressed x-axis has an impact when trying to visualise what the proposed chart would look like, I believe that it is a problem for the human eye to make the appropriate adjustments. There are countless very simple examples where the eye is tricked. For example which line is longer?

-----------------
>-----------------<
or is it this one? -----------------

Incidentally if you are planning to precalculate relative performance measures, I think that you may be underestimating what would be involved.

Hopefully I have explained myself reasonably clearly,
Cheers,
Mosaic

mosaic1996

Member

Post Number: 215
Registered: 01-2003

Rating: N/A
Votes: 0


Wednesday, September 24, 2003 - 08:30 am:

Hi Colin,
I have resubmitted the WBC vs other Banks Chart as the start point that I selected didn't highlight the problems that I see.

Cheers,
Mosaic

colin_twiggs

Member

Post Number: 674
Registered: 09-2002

Rating: N/A
Votes: 0


Wednesday, September 24, 2003 - 10:05 am:

Mosaic

If I understand you correctly, you say that the lines need to start at a common point on the short-term chart for comparison to be meaningful.
I agree that, for the Price Comparison indicator, the price lines must start at zero.

You may be better served by plotting comparative performance on the price Chart using Price Comparison, rather than in a separate slot using the Price Ratio. The Price Ratio indicator's purpose is really to measure relative strength between two securities or indexes. The slope of the line is important and not the value. I believe that Relative Strength is best measured from a fixed point at the start of the data, rather than re-calculating the line with each change in time period.

If there is a need, we will offer both options.

Colin

mosaic1996

Member

Post Number: 220
Registered: 01-2003

Rating: N/A
Votes: 0


Wednesday, September 24, 2003 - 12:07 pm:

Colin,

Getting confused in my old age. Posted response in the wrong place.

http://forum.incrediblecharts.com/messages/8/142740.html

Cheers,
Mosaic

mosaic1996

Member

Post Number: 219
Registered: 01-2003

Rating: N/A
Votes: 0


Wednesday, September 24, 2003 - 12:03 pm:

Hi Colin,

Firstly I should point out that I'm not paying customer (yet).

I am happy to call you to discuss these sorts of issues if you provide a phone number, and you are happy to discuss on the phone.

I'm far more inclined to look at Price Comparison than Price Ratio. Don't really see what the difference is (am I missing something?), except that the relationship is reversed (who's outperforming who), and PR is expressed in percentage terms (which has some appeal for price comparison as well, although using the log scale has a similar effect).

In your original post you grouped price ratio, comparison and differential all together for the same change. I gather that you have had a change of heart on Price Comparison.

I've included a SMS-KAZ price ratio chart to try and demonstrate what I perceive as a potential problem. It is self evident what the relative performance is when the reference point is 0. My concern is that it wouldn't be as self evident if the price ratio started at around 350% and retraced to around 150%.

You have mentioned that the slope of the line is the most important thing, and that the slope remains the same. Personally I have trouble gauging the slope of the line accurately. If the there is something very relevant about the slope of the line, would it be better to plot the first derivative of the price ratio which would make it clear what the slope was actually doing. This would result in a very disjoint plot as the PR slope jumps around (negative, positive from day to day) due to normal market noise. It may be necessary to use a moving average of price ratio.

Cheers,
Mosaic

colin_twiggs

Member

Post Number: 675
Registered: 09-2002

Rating: N/A
Votes: 0


Wednesday, September 24, 2003 - 12:38 pm:

Mosaic

Thanks for your input.
Yes. On closer scrutiny it is probably best to leave the Price Comparison and Price Differential as they are.

We will offer both options for the Relative Strength/Price Ratio indicator:
(1) Price Ratio - as it is now; and
(2) Relative Strength - using the ratio of closing prices at the start of the data.

This will enable us to offer moving averages on option (2). They cannot be calculated on option (1).

I prefer to communicate via the Forum so that all readers are in the picture. If I discuss something on the phone, I never get around to posting an update on the Forum.

Colin

new2waix

Member

Post Number: 135
Registered: 10-2003

Rating: N/A
Votes: 0


Friday, November 14, 2003 - 12:50 am:

Has the issue of the Price Ratio indicator been resolved or changed?

Having read through the early chapters of Weinstein's book he preaches the use of relative strength, however when I adjust the time period my results vary dramatically (as described above).

How can I set this up to work in IC without the distortion?

What settings does Weinstein use?

What time period, weekly/daily? I'm using weekly charts - use weekly RS?

(I apologise if this has been answered... I have searched the IC forum's but may of missed the solution).

Thanks :-)

colin_twiggs

Member

Post Number: 788
Registered: 09-2002

Rating: N/A
Votes: 0


Friday, November 14, 2003 - 03:49 am:

N2W,

The Mansfield Charts from Stan's book appear to calculate RS from the start of the data - and do not normalize to zero at the start of the chart.

I use a 2 or 3-year Price Ratio (compared to XAO or XJO) as an equivalent. Stan uses the cross of RS above/below zero as a signal. To me this is irrelevant as the position (above/below zero) depends on how much data history you have. I focus on the slope (or trend) of the RS line, this does not change, no matter what time period you use.

We plan to introduce a static RS, measured from the start of the data, as an alternative. But this has not been implemented yet.

Colin

max_kemlo

Member

Post Number: 1
Registered: 12-2002

Rating: N/A
Votes: 0


Thursday, August 03, 2006 - 09:25 am:

Colin Twiggs said

The Mansfield Charts from Stan's book appear to calculate RS from the start of the data - and do not normalize to zero at the start of the chart……………We plan to introduce a static RS, measured from the start of the data, as an alternative. But this has not been implemented yet.

Colin there is an explanation of how the relative strength line in calculated around the zero line by Mansfield chart if you still intend to supply this feature on I/C. The problem is that the site is in French, so I had to decipher it, and will gladly supply it to you if you are interested

http://perso.orange.fr/cyberpapy/Mansfield.xls

Mansfield charts are actually still in business here

http://www.mansfield-charts.com/

And example charts are here

http://www.mansfield-charts.com/Targets/Samples/FWeeklySamp.htm

Cheers Max

(Message edited by robin on August 03, 2006)

colin_twiggs

Member

Post Number: 2593
Registered: 09-2002

Rating: N/A
Votes: 0


Thursday, August 03, 2006 - 08:00 pm:

Max,

A translation would be useful.

Thanks, Colin

max_kemlo

Member

Post Number: 2
Registered: 12-2002

Rating: N/A
Votes: 0


Friday, August 04, 2006 - 09:42 am:

Hi Colin
The RS line in Mansfield charts or the "performance ratio" as they call it is plotted by first calculating the “base price” which would appear to be the “zero” line in the performance ratio pane of the chart

This is a loose translation of the explanation in page 1 of the spreadsheet http://perso.orange.fr/cyberpapy/Mansfield.xls but page 2 of the spreadsheet probably sheds the most light on the process as it has a working example.

Méthode de calcul du "base price"
Method of calculating “base price”

Mansfield fait chaque semaine le rapport du cours de clôture hebdo de l'action, au cours de clôture hebdo du S&P500.

Every week Mansfield makes (finds?) the return (or yield) at the weekly closing price of the S&P500.

(Market price (t) / S&Pprice (t) ) que j'appelerai R (t)

(Market price (t) / S&Pprice (t) ) what appeals to me R (t)

Ils en effectuent ensuite la moyenne arithmétique des 52 dermières semaines (le graphe complet en comporte 130 soit deux ans et demi)

It is, in effect, the arithmetical average of the last 52 weeks (the complete graph is composed of
130 weeks, i.e. 2.5 years

Cette valeur donne le "Base price"

This value (equation) gives the “Base price”

Base price = BP = ( R (1)+ R (2) +.... + R (t)+....+R (52) ) / 52

c'est une constante (ou presque...et c'est là mon problème) pour tout le graphe

it is a constant (or almost…and that is my problem) for all the graphs

Le graphe :
le Graphe "performance ratio" est donc la représentation de l'équation:

the graph "performance ratio" is therefore represented by the equation

PR (t) = (( R (t) / BP) -1 ) x 10

le coefficient 10 est un facteur d'échelle choisi arbitrairement par Mansfield.

The coefficient 10 is a factor of scale chosen arbitrarily by Mansfield

From the Mansfield website
Performance ratio scale
This is a percent scale with a zero omitted. Two for example, indicates that the issue is outperforming the market by 20%

I hope this offers some assistance

Cheers max

Ps if you do offer this feature in I/C would it be possible to incorporate it into the stock screening module? As this would be a very beneficial filter

mosaic1996

Member

Post Number: 1346
Registered: 01-2003

Rating: N/A
Votes: 0


Friday, August 04, 2006 - 04:01 pm:

Mansfield calculates the ratio of the weekly close of the share (de l'action) to the weekly close of the S&P500.

I call that (or let that be) R(t)

The rest of the translation is close enough.

However, the link provided goes on to say that his formula does not replicate the Mansfield chart, and that the differences/divergences are very significant, and are not consistent between charts/shares.

Cheers,
Mosaic

colin_twiggs

Member

Post Number: 2597
Registered: 09-2002

Rating: N/A
Votes: 0


Friday, August 04, 2006 - 10:08 pm:

Max,
Thank you for the translation.
Including Relative Strength in the Stock Screens has always been our intention.

Regards,
Colin

mosaic1996

Member

Post Number: 1347
Registered: 01-2003

Rating: N/A
Votes: 0


Saturday, August 05, 2006 - 10:57 am:

Colin,

If you include a Relative Strength to the Stock Screen would it be against XJO, XAO, and/or other indicies such as the relevant sector, ASX300, XSO (small ordinaries), top 20, 50, 100, etc.

Cheers,
Graeme

colin_twiggs

Member

Post Number: 2598
Registered: 09-2002

Rating: N/A
Votes: 0


Saturday, August 05, 2006 - 04:05 pm:

Graeme,
For the ASX we would use the XAO (the results would be very similar if we used XJO).
Regards,
Colin

mosaic1996

Member

Post Number: 1350
Registered: 01-2003

Rating: N/A
Votes: 0


Saturday, August 05, 2006 - 05:21 pm:

Colin,

I suspect that the proposed RS stock screen would be pre-calculated. If this is correct, then do you plan to store a single value for each stock based on the RS from the start of the data.

This approach would be of limited use in my opinion. I believe that if is essential to be able to search on RS over the most recent past, e.g., 1 week (too short?), 2 weeks, 1, 2 and 3 months, 6 months (too long?).

Cheers,
Graeme

froggy

Member

Post Number: 13
Registered: 06-2006

Rating: N/A
Votes: 0


Sunday, August 06, 2006 - 11:50 am:

HI COLIN, MAX, JUST WANTED TO LET YOU KNOW THAT I'M A FRENCH MAN, SO IF I CAN HELP YOU WITH TRANSLATION, I WOULD LOVE TO HELP YOU IF MORE COMPLICATED TEXTS.

CHEERS

JEAN (FROGGY)

colin_twiggs

Member

Post Number: 2601
Registered: 09-2002

Rating: N/A
Votes: 0


Sunday, August 06, 2006 - 03:56 pm:

Jean,
Thank you for your offer of help.

Regards,
Colin

mosaic1996

Member

Post Number: 1351
Registered: 01-2003

Rating: N/A
Votes: 0


Sunday, August 06, 2006 - 05:15 pm:

Colin,

On reflection I can't see that adding RS to stock screens would add any value, given that you already have screening on absolute Percentage Price Movement (+/-) available.

I think that they would result in the same ordered list being generated.

Currently the smallest timeframe is 1 month, and the longest 5 years. personally, I would like to see 1, 2 and 3 weeks added, and personally I can't see the value of 3 and 5 years. Clearly, I can't speak for others.

Cheers,
Graeme

Ca va M. Brandebourg.

max_kemlo

Member

Post Number: 3
Registered: 12-2002

Rating: N/A
Votes: 0


Sunday, August 06, 2006 - 04:55 pm:

Thank for the input guys as this one seems worth pursuing and thanks Froggy for your offer of assistance as my translation was a bit rough. Colin have you given any thought as to how the screening module would work or is a bit early for that yet.

Cheers max

macca888

Member

Post Number: 24
Registered: 10-2002

Rating: N/A
Votes: 0


Monday, August 07, 2006 - 11:31 am:

G'day Colin

I would like to add my weight (and it's increasing) to mosaic1996's comments re RS.

I often add XAO or XJO to a saved watchlist from the screening module and then screen it again for a Percentage Price Movement comparison to the stocks on the list.

The 3 and 5 year criteria are almost irrelevant for me because of my expected trading timeframe and the fact that many stocks change name,list or delist over this period of time and hence don't come up on my scans if I include these criteria in the Percentage Price Movement.

Any shorter timeframes than 1 year in addition to those already available would be more useful to me and I think mosaic's suggestion re 1,2,or 3 weeks is excellent.

On the rare occasions, more out of interest rather than necessity, I could calculate the Percentage Price Movement over 3 and 5 years from the chart or on a spreadsheet.

macca888

Incredible Charts
Free Charting Software

Auto-Fit Trendlines

Trend Channels

Linear Regression
Channels

Raff Regression
Channels

Standard Deviation
Channels

"Buy low, sell high is a cliche, not a blueprint for action. It blinds investors to the professionals' approach of buying high and selling higher." Stan Weinstein