The investors who sell because nothing happened
In 2017 the American stock market was so calm that a whole class of institutional portfolios held more equities than their own capital, simply because the rules they follow told them to. Reconstructed with the model behind this indicator, their estimated equity exposure reached 135 per cent of capital at its peak that year and still stood at 111 per cent on the last session before 5 February 2018. By 14 February it was 47 per cent. None of these portfolios had changed its view of the economy, of earnings or of valuations, and most of them have no view at all. What had changed was a single number, the volatility of the past few weeks, and the rule translated that number into selling. This indicator exists to make such buying and selling, driven by rules rather than by opinions, visible to anyone who looks at a chart, because it moves prices whether or not one knows about it.
A rule instead of an opinion
A volatility control portfolio sets itself a target, typically a level of annual volatility such as ten per cent, and then sizes its equity position so that the portfolio as a whole moves roughly that much. The arithmetic is disarmingly simple. If the market has recently been moving at an annualised twenty per cent, a ten per cent target means holding half the capital in equities and the rest in cash, and if the market calms down to eight per cent, the same target calls for 125 per cent, which is achieved with futures or borrowed money, usually up to a ceiling of 150 per cent. Exposure is therefore target volatility divided by realised volatility, capped at a maximum leverage, and recalculated every day.
Pine Script®
Such rules are not a fringe product. S&P Dow Jones Indices publishes a whole family of risk control indices built on exactly this formula, and its own research describes how pension plans with defined benefits and insurers selling variable annuities use them to obtain a smoother path of returns that fits their liabilities better (Cheng et al., 2021). There is academic support for the idea as well. Moreira and Muir (2017) showed that portfolios which take less risk when volatility is high earned higher returns relative to their risk across a range of equity factors, and Harvey et al. (2018), testing more than sixty assets, found that volatility targeting raises the Sharpe ratio for equities and credit and reduces the severity of extreme losses across all asset classes, since the worst days tend to arrive when volatility is already high and the portfolio run to a volatility target is already small.
Why a sensible rule becomes everybody's problem
What is prudent for a single pension fund acquires a different character once a large amount of money follows the same rule, because the rule is procyclical by construction. It buys when markets have been quiet, which usually means after they have risen, and it sells when volatility jumps, which usually means while they are falling. Volatility, in turn, does not arrive and leave in single days but clusters, a property that Engle (1982) formalised when he showed that the recent past carries information about the variance to come, so that a violent day is rarely followed by calm and the selling continues for as long as the volatility estimate keeps rising. A detail that surprises most people is that the rule does not care about direction. Realised volatility is computed from squared returns, so a three per cent rally raises the estimate just as a three per cent decline does, and a sharp rebound in a nervous market can trigger further selling from these portfolios.
The consequence for everybody else is that part of the supply and demand in equity markets is mechanical and, within limits, predictable. Bank research follows this group of investors closely, Deutsche Bank's weekly Investor Positioning and Flows report for instance carries its own section on the equity allocation of volatility control funds (Deutsche Bank, 2026, cited in Finvaulta, 2026), yet neither the methods behind such estimates nor the amounts of money involved are public.
Rebuilding the machine from public rules
Because the rule depends on nothing but prices, the positioning can nonetheless be rebuilt from the chart itself. This script is not affiliated with or endorsed by Deutsche Bank or any other bank, and it does not attempt to reproduce their figures, which are unavailable, but reconstructs the mechanism from rules that are documented in public. The most detailed of these is the S&P methodology, in which realised volatility is the higher of two exponentially weighted estimates of daily log returns, one with a decay factor of 0.94 that reacts quickly and one with 0.97 that reacts more slowly, annualised with 252 trading days, and in which two days pass between the observation of volatility and the rebalancing of the index (S&P Dow Jones Indices, 2014, 2024). The same pair of decay factors appeared as early as 1996 in J.P. Morgan's RiskMetrics, where 0.94 served for trading horizons and 0.97 for investment horizons (J.P. Morgan/Reuters, 1996).
Pine Script®
Real portfolios are not all built alike, and so the indicator blends four stylised designs, weighted equally because nobody outside these funds knows how the money is actually distributed across them.
The weights can be changed under the Custom preset, and a design without data, such as the bond portfolio before IEF started trading in 2002, simply drops out of the average. For readers who want to follow one type of fund rather than the blend, a preset menu replaces the four designs with a single rule.
On SPY every preset cut its exposure in at least nine of the ten shock episodes, whereas the top zone was followed by a clearly larger reduction for the blend, the simple weighted and the balanced design, less clearly for the pure S&P rule and not at all for the fast decay rule, whose slow second factor keeps exposure from building up in calm markets in the first place. The blend therefore remains the default.
What distinguishes this from an ordinary volatility indicator is that it does not stop at measuring volatility but translates it into the positions a specific group of investors must hold, with their leverage ceiling, their implementation delay and, in the fourth design, the diversification their bonds provide. From that translation follow the outputs a volatility line cannot give, namely the adjustment already fixed for the next session, the size of move that would force these portfolios to sell, and the scale of that selling for moves of a given size.
What appears on the chart
The main line shows where today's estimated exposure stands relative to the past five years, as a percentile between 0 and 100. A reading of 85 means that exposure was at or below today's level on 85 per cent of the days in that window. Alternatively the line can show the exposure itself in per cent of capital, together with the leverage ceiling. The colour of the line shifts gradually from green at the bottom of the range over blue in the middle to red at the top, or in five steps if the stepped colouring is preferred, the same colour is used for the candles if candle colouring is on, and a shaded area between the line and the midpoint makes the distance from normal visible at a glance. Five zones, from very low below the 10th percentile to very high above the 90th, are marked by bands and faint lines, and by default only the two extremes tint the background. Dots mark the days on which the reading enters one of the extremes. The zones are descriptive labels, not buy or sell instructions, and the thresholds can be moved.
The dashboard adds what the line cannot show and comes in two versions. The simple one answers the questions most readers have.
The professional version goes further and reports the estimated exposure, its percentile and its change over one and five days. Because of the lag of two days, tomorrow's adjustment is already determined by closes that lie in the past, and the row labelled scheduled next session shows it, positive when the portfolios will add equities and negative when they will reduce them. The four volatility estimates are listed side by side. Two further rows state the size of the next day's move, downwards and upwards, beyond which target exposure starts to fall, and an optional table shows how much target exposure would change after moves of one, two or three per cent in either direction, with the bond leg assumed unchanged. All changes are expressed in percentage points of exposure. A field for assets under management converts them into dollars, but it is set to zero by default, since no reliable public figure for the size of these strategies exists and any number entered there would lend the output a precision it does not have.
Four alerts are available.
A status row at the bottom of the dashboard reports whether the model is still warming up, whether data for the symbol or the bond leg is missing, and whether the chart is intraday or a crypto market, where the convention of 252 trading days per year understates volatility levels although the percentile remains comparable. On continuous futures such as ES1! the price jumps at every contract roll unless the chart is adjusted for contract rolls, and the model would read that jump as a market move, so the index itself or an adjusted chart is the better choice.
What history says about the reading
The model was tested on daily data for the SPDR S&P 500 ETF from 1993 to September 2026 and the iShares Treasury bond fund IEF from 2002, following a hypothesis written down before the first calculation. Ten shock dates were fixed in advance, from the Lehman collapse in September 2008 to the tariff shock of April 2025, and in nine of them the estimated exposure fell by between 25 and 61 per cent within twenty sessions. After 5 August 2024 it fell from 74 to 55 per cent within ten sessions, and during March 2020 from an already low 35 to 14 per cent. The one miss is instructive. In June 2022 exposure had been falling for months and stood at 38 per cent before the shock, so there was comparatively little left to sell and the cut amounted to 18 per cent. The test confirms direction and timing, whereas the size of these episodes in dollars cannot be checked with public data.
The more useful finding concerns the percentile. When the reading was in its top decile, the following 21 sessions brought an average reduction of 6.0 to 7.6 percentage points of exposure, against an average of practically zero across all days, measured on periods that do not overlap so that the same episode was not counted several times, with p values between 0.002 and 0.006. Across the zones the average change over the following month ran from plus 3.4 points in the lowest zone to minus 6.9 in the highest, although the two lower zones hardly differed from each other, which is the reason they are presented as labels rather than as signals. Part of this pattern is mechanical and should be read as such, because volatility tends to return to its average and exposure is its inverse. A very high reading means calm, and calm has historically been followed by less calm more often than by more.
Two further results shaped the design. Modelling several targets at once, say five, ten, twelve and fifteen per cent, changes the percentile only where the leverage ceiling binds, and the percentile of such a blend correlated at 0.9995 with that of the single ten per cent model, so the indicator uses one target. The balanced stock and bond portfolio, by contrast, behaves differently enough to matter, as its daily changes correlated at only 0.72 with those of the design that holds equities alone, the gap being largest in years like 2013 when bonds diversified and smallest in 2022 when stocks and bonds fell together.
Using it without overestimating it
The indicator is best read as a description of how fragile or how supported the market is from the side of investors who follow rules, and it is best read once a day after the close. With the default lag the exposure, its percentile and the scheduled change are fixed by earlier closes and do not move during the session, whereas the move thresholds and the scenarios depend on the current daily bar and keep changing until it closes. A very high reading says that a large amount of money is invested because markets have been quiet and will leave if they stop being quiet. This is a reason to be modest with leverage, to consider protection while volatility is still low, and to expect a shock to be amplified rather than absorbed. A very low reading after a sharp decline says that much of the forced selling has already happened and that the same rule will buy back gradually once volatility subsides, a support that depends entirely on that subsidence actually taking place. The scheduled change and the move thresholds help to put a single day into context, for instance to understand why a market that has just fallen one or two per cent may meet additional mechanical supply over the following sessions. Perhaps the most lasting use has little to do with timing at all, namely to size one's own positions by recent volatility instead of by a fixed amount, which is the very idea the academic work cited above supports.
What the indicator does not do should be stated as plainly. It makes no claim about future returns, and no trading rule built on its zones has been tested. The four designs are a stylised sample of a market whose real composition, targets, rebalancing bands and use of implied volatility vary from manager to manager. The same flows are estimated by bank research desks as well, so the indicator offers a picture that professionals already have rather than an advantage over them. Its bond leg consists of US Treasuries on every chart, which suits American indices and fits European or Asian ones poorly, and in those cases a preset without the bond portfolio, such as the S&P risk control preset, is the better choice. Finally, the model works on daily data regardless of the chart's timeframe, and the changes over one and five days are counted in trading days on weekly and intraday charts too.
References
Cheng, T., Yi, C., Banerjee, A. and Srivastava, V. (2021) Limiting Risk Exposure with S&P Risk Control Indices. New York: S&P Dow Jones Indices. Available at: spglobal.com/spdji/en/documents/research/research-limiting-risk-exposure-with-sp-risk-control-indices.pdf (Accessed: 27 September 2026).
Deutsche Bank (2026) Investor Positioning and Flows, 28 August 2026. Frankfurt am Main: Deutsche Bank. Cited in: Finvaulta (2026) Sideways at Modestly Overweight. Available at: finvaulta.com/research/deutsche-bank/sideways-at-modestly-overweight-2026-08-28 (Accessed: 27 September 2026).
Engle, R.F. (1982) 'Autoregressive conditional heteroscedasticity with estimates of the variance of United Kingdom inflation', Econometrica, 50(4), pp. 987-1007.
Harvey, C.R., Hoyle, E., Korgaonkar, R., Rattray, S., Sargaison, M. and Van Hemert, O. (2018) 'The impact of volatility targeting', Journal of Portfolio Management, 45(1), pp. 14-33. doi:10.3905/jpm.2018.45.1.014.
JPMorgan Chase & Co. (2026) Notes Linked to the S&P Risk Control Index Series. Underlying supplement no. 2-I, dated 17 April 2026, filed under Rule 424(b)(2). Washington, DC: U.S. Securities and Exchange Commission. Available at: sec.gov/Archives/edgar/data/19617/000121390026045213/ea0285802-17_424b2.pdf (Accessed: 27 September 2026).
J.P. Morgan/Reuters (1996) RiskMetrics Technical Document. 4th edn. New York: Morgan Guaranty Trust Company. Available at: msci.com/documents/10199/5915b101-4206-4ba0-aee2-3449d5c7e95a (Accessed: 27 September 2026).
Moreira, A. and Muir, T. (2017) 'Volatility-managed portfolios', Journal of Finance, 72(4), pp. 1611-1644. doi:10.1111/jofi.12513.
S&P Dow Jones Indices (2014) Index Mathematics Methodology. New York: S&P Dow Jones Indices. Available at: spice-indices.com/idpfiles/spice-assets/resources/public/documents/methodology-index-math.pdf (Accessed: 27 September 2026).
S&P Dow Jones Indices (2024) S&P Risk Control 2.0 Indices Methodology. New York: S&P Dow Jones Indices. Available at: spglobal.com/spdji/pt/documents/methodologies/methodology-sp-risk-control-2-indices.pdf (Accessed: 27 September 2026).
Disclaimer
In 2017 the American stock market was so calm that a whole class of institutional portfolios held more equities than their own capital, simply because the rules they follow told them to. Reconstructed with the model behind this indicator, their estimated equity exposure reached 135 per cent of capital at its peak that year and still stood at 111 per cent on the last session before 5 February 2018. By 14 February it was 47 per cent. None of these portfolios had changed its view of the economy, of earnings or of valuations, and most of them have no view at all. What had changed was a single number, the volatility of the past few weeks, and the rule translated that number into selling. This indicator exists to make such buying and selling, driven by rules rather than by opinions, visible to anyone who looks at a chart, because it moves prices whether or not one knows about it.
A rule instead of an opinion
A volatility control portfolio sets itself a target, typically a level of annual volatility such as ten per cent, and then sizes its equity position so that the portfolio as a whole moves roughly that much. The arithmetic is disarmingly simple. If the market has recently been moving at an annualised twenty per cent, a ten per cent target means holding half the capital in equities and the rest in cash, and if the market calms down to eight per cent, the same target calls for 125 per cent, which is achieved with futures or borrowed money, usually up to a ceiling of 150 per cent. Exposure is therefore target volatility divided by realised volatility, capped at a maximum leverage, and recalculated every day.
Such rules are not a fringe product. S&P Dow Jones Indices publishes a whole family of risk control indices built on exactly this formula, and its own research describes how pension plans with defined benefits and insurers selling variable annuities use them to obtain a smoother path of returns that fits their liabilities better (Cheng et al., 2021). There is academic support for the idea as well. Moreira and Muir (2017) showed that portfolios which take less risk when volatility is high earned higher returns relative to their risk across a range of equity factors, and Harvey et al. (2018), testing more than sixty assets, found that volatility targeting raises the Sharpe ratio for equities and credit and reduces the severity of extreme losses across all asset classes, since the worst days tend to arrive when volatility is already high and the portfolio run to a volatility target is already small.
Why a sensible rule becomes everybody's problem
What is prudent for a single pension fund acquires a different character once a large amount of money follows the same rule, because the rule is procyclical by construction. It buys when markets have been quiet, which usually means after they have risen, and it sells when volatility jumps, which usually means while they are falling. Volatility, in turn, does not arrive and leave in single days but clusters, a property that Engle (1982) formalised when he showed that the recent past carries information about the variance to come, so that a violent day is rarely followed by calm and the selling continues for as long as the volatility estimate keeps rising. A detail that surprises most people is that the rule does not care about direction. Realised volatility is computed from squared returns, so a three per cent rally raises the estimate just as a three per cent decline does, and a sharp rebound in a nervous market can trigger further selling from these portfolios.
The consequence for everybody else is that part of the supply and demand in equity markets is mechanical and, within limits, predictable. Bank research follows this group of investors closely, Deutsche Bank's weekly Investor Positioning and Flows report for instance carries its own section on the equity allocation of volatility control funds (Deutsche Bank, 2026, cited in Finvaulta, 2026), yet neither the methods behind such estimates nor the amounts of money involved are public.
Rebuilding the machine from public rules
Because the rule depends on nothing but prices, the positioning can nonetheless be rebuilt from the chart itself. This script is not affiliated with or endorsed by Deutsche Bank or any other bank, and it does not attempt to reproduce their figures, which are unavailable, but reconstructs the mechanism from rules that are documented in public. The most detailed of these is the S&P methodology, in which realised volatility is the higher of two exponentially weighted estimates of daily log returns, one with a decay factor of 0.94 that reacts quickly and one with 0.97 that reacts more slowly, annualised with 252 trading days, and in which two days pass between the observation of volatility and the rebalancing of the index (S&P Dow Jones Indices, 2014, 2024). The same pair of decay factors appeared as early as 1996 in J.P. Morgan's RiskMetrics, where 0.94 served for trading horizons and 0.97 for investment horizons (J.P. Morgan/Reuters, 1996).
Real portfolios are not all built alike, and so the indicator blends four stylised designs, weighted equally because nobody outside these funds knows how the money is actually distributed across them.
- The S&P convention just described, with the two exponentially weighted estimates.
- A plain average of squared returns over one month, since S&P permits equally weighted estimates as an alternative to exponential weighting (Cheng et al., 2021), with a window of my choosing, as S&P does not publish one.
- The same over three months, again a window of my choosing.
- A balanced portfolio of sixty per cent equities and forty per cent US Treasuries, represented by the IEF fund and scaled to the same target on the volatility of the whole portfolio, so that the correlation between stocks and bonds enters the calculation. Its equity exposure is the equity share multiplied by the scaling factor. Risk control products combining an equity and a bond index exist in the S&P family (S&P Dow Jones Indices, 2024), which is why this design is included by default.
The weights can be changed under the Custom preset, and a design without data, such as the bond portfolio before IEF started trading in 2002, simply drops out of the average. For readers who want to follow one type of fund rather than the blend, a preset menu replaces the four designs with a single rule.
- S&P risk control uses the decay factors of 0.94 and 0.97 that govern the S&P 500 Daily Risk Control indices.
- Fast decay uses the pair of 0.8705 and 0.99725 that S&P applies to one of its global risk control indices (JPMorgan Chase & Co., 2026).
- Simple weighted takes the higher of the volatility over one month and over three months.
- Balanced 60/40 follows only the stock and bond portfolio.
- Custom opens every setting of the four designs.
On SPY every preset cut its exposure in at least nine of the ten shock episodes, whereas the top zone was followed by a clearly larger reduction for the blend, the simple weighted and the balanced design, less clearly for the pure S&P rule and not at all for the fast decay rule, whose slow second factor keeps exposure from building up in calm markets in the first place. The blend therefore remains the default.
What distinguishes this from an ordinary volatility indicator is that it does not stop at measuring volatility but translates it into the positions a specific group of investors must hold, with their leverage ceiling, their implementation delay and, in the fourth design, the diversification their bonds provide. From that translation follow the outputs a volatility line cannot give, namely the adjustment already fixed for the next session, the size of move that would force these portfolios to sell, and the scale of that selling for moves of a given size.
What appears on the chart
The main line shows where today's estimated exposure stands relative to the past five years, as a percentile between 0 and 100. A reading of 85 means that exposure was at or below today's level on 85 per cent of the days in that window. Alternatively the line can show the exposure itself in per cent of capital, together with the leverage ceiling. The colour of the line shifts gradually from green at the bottom of the range over blue in the middle to red at the top, or in five steps if the stepped colouring is preferred, the same colour is used for the candles if candle colouring is on, and a shaded area between the line and the midpoint makes the distance from normal visible at a glance. Five zones, from very low below the 10th percentile to very high above the 90th, are marked by bands and faint lines, and by default only the two extremes tint the background. Dots mark the days on which the reading enters one of the extremes. The zones are descriptive labels, not buy or sell instructions, and the thresholds can be moved.
A reading pinned at 100 for weeks, as in 2017, is no error. When volatility is low enough, every design reaches its leverage ceiling and the estimated exposure stays at the same maximum day after day, and since the percentile counts days at or below today's level, it remains at 100 until the ceiling is left.
The dashboard adds what the line cannot show and comes in two versions. The simple one answers the questions most readers have.
- Where positioning stands and how much of their capital these portfolios hold in equities.
- How that compares with the past five years, shown as a rank out of 100.
- Whether they will buy or sell at the next session.
- How large a move would make them sell.
The professional version goes further and reports the estimated exposure, its percentile and its change over one and five days. Because of the lag of two days, tomorrow's adjustment is already determined by closes that lie in the past, and the row labelled scheduled next session shows it, positive when the portfolios will add equities and negative when they will reduce them. The four volatility estimates are listed side by side. Two further rows state the size of the next day's move, downwards and upwards, beyond which target exposure starts to fall, and an optional table shows how much target exposure would change after moves of one, two or three per cent in either direction, with the bond leg assumed unchanged. All changes are expressed in percentage points of exposure. A field for assets under management converts them into dollars, but it is set to zero by default, since no reliable public figure for the size of these strategies exists and any number entered there would lend the output a precision it does not have.
Four alerts are available.
- Very High fires when the reading enters the top zone.
- Very Low fires when it enters the bottom zone.
- Mid Range Entry fires when it returns to the middle.
- Deleveraging Event fires when estimated exposure is cut by more than a chosen number of percentage points in a single day, ten by default.
A status row at the bottom of the dashboard reports whether the model is still warming up, whether data for the symbol or the bond leg is missing, and whether the chart is intraday or a crypto market, where the convention of 252 trading days per year understates volatility levels although the percentile remains comparable. On continuous futures such as ES1! the price jumps at every contract roll unless the chart is adjusted for contract rolls, and the model would read that jump as a market move, so the index itself or an adjusted chart is the better choice.
What history says about the reading
The model was tested on daily data for the SPDR S&P 500 ETF from 1993 to September 2026 and the iShares Treasury bond fund IEF from 2002, following a hypothesis written down before the first calculation. Ten shock dates were fixed in advance, from the Lehman collapse in September 2008 to the tariff shock of April 2025, and in nine of them the estimated exposure fell by between 25 and 61 per cent within twenty sessions. After 5 August 2024 it fell from 74 to 55 per cent within ten sessions, and during March 2020 from an already low 35 to 14 per cent. The one miss is instructive. In June 2022 exposure had been falling for months and stood at 38 per cent before the shock, so there was comparatively little left to sell and the cut amounted to 18 per cent. The test confirms direction and timing, whereas the size of these episodes in dollars cannot be checked with public data.
The more useful finding concerns the percentile. When the reading was in its top decile, the following 21 sessions brought an average reduction of 6.0 to 7.6 percentage points of exposure, against an average of practically zero across all days, measured on periods that do not overlap so that the same episode was not counted several times, with p values between 0.002 and 0.006. Across the zones the average change over the following month ran from plus 3.4 points in the lowest zone to minus 6.9 in the highest, although the two lower zones hardly differed from each other, which is the reason they are presented as labels rather than as signals. Part of this pattern is mechanical and should be read as such, because volatility tends to return to its average and exposure is its inverse. A very high reading means calm, and calm has historically been followed by less calm more often than by more.
Two further results shaped the design. Modelling several targets at once, say five, ten, twelve and fifteen per cent, changes the percentile only where the leverage ceiling binds, and the percentile of such a blend correlated at 0.9995 with that of the single ten per cent model, so the indicator uses one target. The balanced stock and bond portfolio, by contrast, behaves differently enough to matter, as its daily changes correlated at only 0.72 with those of the design that holds equities alone, the gap being largest in years like 2013 when bonds diversified and smallest in 2022 when stocks and bonds fell together.
Using it without overestimating it
The indicator is best read as a description of how fragile or how supported the market is from the side of investors who follow rules, and it is best read once a day after the close. With the default lag the exposure, its percentile and the scheduled change are fixed by earlier closes and do not move during the session, whereas the move thresholds and the scenarios depend on the current daily bar and keep changing until it closes. A very high reading says that a large amount of money is invested because markets have been quiet and will leave if they stop being quiet. This is a reason to be modest with leverage, to consider protection while volatility is still low, and to expect a shock to be amplified rather than absorbed. A very low reading after a sharp decline says that much of the forced selling has already happened and that the same rule will buy back gradually once volatility subsides, a support that depends entirely on that subsidence actually taking place. The scheduled change and the move thresholds help to put a single day into context, for instance to understand why a market that has just fallen one or two per cent may meet additional mechanical supply over the following sessions. Perhaps the most lasting use has little to do with timing at all, namely to size one's own positions by recent volatility instead of by a fixed amount, which is the very idea the academic work cited above supports.
What the indicator does not do should be stated as plainly. It makes no claim about future returns, and no trading rule built on its zones has been tested. The four designs are a stylised sample of a market whose real composition, targets, rebalancing bands and use of implied volatility vary from manager to manager. The same flows are estimated by bank research desks as well, so the indicator offers a picture that professionals already have rather than an advantage over them. Its bond leg consists of US Treasuries on every chart, which suits American indices and fits European or Asian ones poorly, and in those cases a preset without the bond portfolio, such as the S&P risk control preset, is the better choice. Finally, the model works on daily data regardless of the chart's timeframe, and the changes over one and five days are counted in trading days on weekly and intraday charts too.
References
Cheng, T., Yi, C., Banerjee, A. and Srivastava, V. (2021) Limiting Risk Exposure with S&P Risk Control Indices. New York: S&P Dow Jones Indices. Available at: spglobal.com/spdji/en/documents/research/research-limiting-risk-exposure-with-sp-risk-control-indices.pdf (Accessed: 27 September 2026).
Deutsche Bank (2026) Investor Positioning and Flows, 28 August 2026. Frankfurt am Main: Deutsche Bank. Cited in: Finvaulta (2026) Sideways at Modestly Overweight. Available at: finvaulta.com/research/deutsche-bank/sideways-at-modestly-overweight-2026-08-28 (Accessed: 27 September 2026).
Engle, R.F. (1982) 'Autoregressive conditional heteroscedasticity with estimates of the variance of United Kingdom inflation', Econometrica, 50(4), pp. 987-1007.
Harvey, C.R., Hoyle, E., Korgaonkar, R., Rattray, S., Sargaison, M. and Van Hemert, O. (2018) 'The impact of volatility targeting', Journal of Portfolio Management, 45(1), pp. 14-33. doi:10.3905/jpm.2018.45.1.014.
JPMorgan Chase & Co. (2026) Notes Linked to the S&P Risk Control Index Series. Underlying supplement no. 2-I, dated 17 April 2026, filed under Rule 424(b)(2). Washington, DC: U.S. Securities and Exchange Commission. Available at: sec.gov/Archives/edgar/data/19617/000121390026045213/ea0285802-17_424b2.pdf (Accessed: 27 September 2026).
J.P. Morgan/Reuters (1996) RiskMetrics Technical Document. 4th edn. New York: Morgan Guaranty Trust Company. Available at: msci.com/documents/10199/5915b101-4206-4ba0-aee2-3449d5c7e95a (Accessed: 27 September 2026).
Moreira, A. and Muir, T. (2017) 'Volatility-managed portfolios', Journal of Finance, 72(4), pp. 1611-1644. doi:10.1111/jofi.12513.
S&P Dow Jones Indices (2014) Index Mathematics Methodology. New York: S&P Dow Jones Indices. Available at: spice-indices.com/idpfiles/spice-assets/resources/public/documents/methodology-index-math.pdf (Accessed: 27 September 2026).
S&P Dow Jones Indices (2024) S&P Risk Control 2.0 Indices Methodology. New York: S&P Dow Jones Indices. Available at: spglobal.com/spdji/pt/documents/methodologies/methodology-sp-risk-control-2-indices.pdf (Accessed: 27 September 2026).
Disclaimer
This indicator is provided for educational and analytical purposes. It estimates the positioning of a class of investors from public rules and price data and does not constitute investment advice or a recommendation to buy or sell any security.
開源腳本
秉持TradingView一貫精神，這個腳本的創作者將其設為開源，以便交易者檢視並驗證其功能。向作者致敬！您可以免費使用此腳本，但請注意，重新發佈代碼需遵守我們的社群規範。
Where others speculate, we systematize.
edgetools.org
edgetools.org
免責聲明
這些資訊和出版物並非旨在提供，也不構成TradingView提供或認可的任何形式的財務、投資、交易或其他類型的建議或推薦。請閱讀使用條款以了解更多資訊。
開源腳本
秉持TradingView一貫精神，這個腳本的創作者將其設為開源，以便交易者檢視並驗證其功能。向作者致敬！您可以免費使用此腳本，但請注意，重新發佈代碼需遵守我們的社群規範。
Where others speculate, we systematize.
edgetools.org
edgetools.org
免責聲明
這些資訊和出版物並非旨在提供，也不構成TradingView提供或認可的任何形式的財務、投資、交易或其他類型的建議或推薦。請閱讀使用條款以了解更多資訊。