Howdy, Stranger!
Categories
- All Categories
- 2.3K StockCharts
- 387 SharpCharts
- 143 Other Charting Tools
- 67 Saved Charts and ChartLists
- 1.4K Scanning
- 72 Data Issues
- 172 Other StockCharts Questions
- 214 Technical Analysis
- 154 Using Technical Analysis
- 2 InterMarket and International
- 19 Market and Breadth Indicators
- 39 Market Analysis
- 105 Trading
- 105 Trading Strategies
- 164 S.C.A.N the StockCharts Answer Network forum
- 65 Using this StockCharts Answer Network forum
- 99 s.c.a.n. archives
- 17 Off-Topic
- 10 The Cogitation & Rumination Emporium
- 7 Forum Test Area
New Members: Be sure to confirm your email address by clicking on the link that was sent to your email inbox. You will not be able to post messages until you click that link.
PerfChart relative performance: ratio vs. difference (Bogle "telltale chart")
The Stockcharts.com ChartSchool says: "When a PerfChart has a baseline, the performance of the baseline symbol is subtracted from the performance of all of the other symbols. This shows each symbol's performance relative to the performance of the baseline symbol." My proposal is for StockCharts to add the ability to plot relative performance as a ratio, rather than a difference.
Here's the motivation.
Charting relative performance as a difference can be misleading. For example, click "VTI" on the following PerfChart. Stretch the left end of the bottom slider back and forth between 2002 & 2007. You'll see that for a slider start date of 2002, VXF appears to *outperform* VTI between 2009-2019. But for a slider start date of 2007, VXF appears to *underperform* VTI over the same time period (2009-2019)!
This inconsistency (and dependence on the reference start date) illustrates why the relative performance is probably more aptly computed as a ratio, not a difference. For example, see the "telltale chart" of former Vanguard CEO John Bogle. For ticker symbols A & B with starting values A0 & B0, the Bogle chart shows (B/B0)/(A/A0) = B/A * A0/B0. Comparing two dates, again using a ratio, would give (B2/A2*A0/B0) / (B1/A1*A0/B0) = (B2/A2) / (B1/A1). This is exactly what we mean by relative performance, which is appropriately independent of the arbitrary reference start values A0 & B0.
In contrast, as I understand it the PerfChart currently shows B/B0 - A/A0 = (A/A0) * [(B/B0)/(A/A0) - 1] = (A/A0) * (BogleChart - 1). So the Perfchart amplifies the actual relative performance by A/A0, the amount that A has grown since the arbitrary slider start date. This is not desirable. Comparing two dates on the PerfChart, using a ratio, would give A2/A1 * (BogleChart2 - 1) / (Boglechart1 - 1), which scales the true relative performance by the performance of A. Alternatively, comparing two dates on the PerfChart using a difference would give (B2/B0 - A2/A0) - (B1/B0 - A1/A0) = (B2-B1)/B0 - (A2-A1)/A0. So a small arbitrary B0 would amplify the apparent subsequent performance of B, while a small arbitrary A0 would amplify apparent subsequent performance of A. This is why the slopes on the relative PerfChart, as currently implemented, depend so wildly on the starting date of the slider.
In short, I'd request either correcting the PerfChart to show a true Bogle-type chart of relative performance as a ratio, or at least adding a feature to do this. Again, this would be simply: 100% * [(B/B0)/(A/A0) - 1]. Would others like to see this as well?
Thanks for any help or support you can offer. (I'm a new member seeking help on SCAN, after receiving no response from Stockcharts.com for a few weeks.)
Here's the motivation.
Charting relative performance as a difference can be misleading. For example, click "VTI" on the following PerfChart. Stretch the left end of the bottom slider back and forth between 2002 & 2007. You'll see that for a slider start date of 2002, VXF appears to *outperform* VTI between 2009-2019. But for a slider start date of 2007, VXF appears to *underperform* VTI over the same time period (2009-2019)!
This inconsistency (and dependence on the reference start date) illustrates why the relative performance is probably more aptly computed as a ratio, not a difference. For example, see the "telltale chart" of former Vanguard CEO John Bogle. For ticker symbols A & B with starting values A0 & B0, the Bogle chart shows (B/B0)/(A/A0) = B/A * A0/B0. Comparing two dates, again using a ratio, would give (B2/A2*A0/B0) / (B1/A1*A0/B0) = (B2/A2) / (B1/A1). This is exactly what we mean by relative performance, which is appropriately independent of the arbitrary reference start values A0 & B0.
In contrast, as I understand it the PerfChart currently shows B/B0 - A/A0 = (A/A0) * [(B/B0)/(A/A0) - 1] = (A/A0) * (BogleChart - 1). So the Perfchart amplifies the actual relative performance by A/A0, the amount that A has grown since the arbitrary slider start date. This is not desirable. Comparing two dates on the PerfChart, using a ratio, would give A2/A1 * (BogleChart2 - 1) / (Boglechart1 - 1), which scales the true relative performance by the performance of A. Alternatively, comparing two dates on the PerfChart using a difference would give (B2/B0 - A2/A0) - (B1/B0 - A1/A0) = (B2-B1)/B0 - (A2-A1)/A0. So a small arbitrary B0 would amplify the apparent subsequent performance of B, while a small arbitrary A0 would amplify apparent subsequent performance of A. This is why the slopes on the relative PerfChart, as currently implemented, depend so wildly on the starting date of the slider.
In short, I'd request either correcting the PerfChart to show a true Bogle-type chart of relative performance as a ratio, or at least adding a feature to do this. Again, this would be simply: 100% * [(B/B0)/(A/A0) - 1]. Would others like to see this as well?
Thanks for any help or support you can offer. (I'm a new member seeking help on SCAN, after receiving no response from Stockcharts.com for a few weeks.)
0
Comments
You may be interested to explore ratio charts, which you can create by entering two symbols in the symbol window (upper left corner of the chart workbench), separated by a colon - e.g. VXF:VTI, or VTI:VXF.
You can also use the Price-Performance indicator (select from a drop down in the "Indicator" column) with symbol pairs entered in the same abc:xyz format. You can place multiple instances of this indicator on the main chart by choosing "Behind Price" in the "Position" column.
The ratio symbol calculation is explained here, a little more than half way down the page:
https://support.stockcharts.com/doku.php?id=data
Thanks for pointing to the "ratio symbol" capability (e.g. VXF:VTI) -- combining that with StockChartsACP gives me much of what I was looking for. Unfortunately it seems that the ACP doesn't (yet) offer the nice multi-symbol overlay capability of the PerfChart, so I can only view one ratio symbol at a time.
Does anyone know of a way on StockCharts to plot multiple ratio symbols, analogous to the multiple difference symbols on the PerfChart? On the ACP I tried doing VXF:VTI,VBR:VTI etc., but that didn't work -- it only used the first ratio symbol. And the PerfChart didn't recognize ratio symbols at all.
You can't use ratio symbols in the Price (Same Scale) on the overlays of charts of StockCharts either.
May be a good thing to create a chartlist with your ratio symbols included. You can view charts in summary, candleglance, etc.
RRG might be an option if you are using the same ratio denominator but, I don't think it's the visual you were looking for. If your chartlist has 50 or less symbols you can use/run RRG as one of the dropdown options for viewing a chartlist but Ratio symbols don't work there either.
Depending on your membership level you can put a number of ratio symbols in the Price and /or Price Performance indicators on the chart. Not the ACP chart, the really good chart in the StockCharts legacy system. I think you can add 4 at the extra level but maybe more if you are using the same denominator. I think you can pull 6 unique prices on one chart or something like that.