Reprinted from Technical Analysis of STOCKS &
COMMODITIES magazine. © 2007 Technical Analysis Inc., (800) 832-4642, http://www. traders.com INDICATORS
Connection And Affinity
Between Betwe en Pric Price e And Volu olume me Here’s an indicator that can be used to measure the intrinsic relationship between price and volume.
W
opinions, and emotio emotions. ns. Market volume represents the number of shares traded over a given period. It is a measurement of the participation, enthusiasm, and interest in a given security. Think of volume as the force that drives the market. Volume substantiates, energizes, and empowers price. When volume increases, it confirms price direction; when volume decreases, it contradicts price direction. In theory, increases in volume generally precede significant price movements. This tenet of technical analysis, that volume precedes price, has been repeated as a mantra since the days of Charles Dow. Within these two independently derived variables, price and volume, exists an intrinsic relationship. When examined together, price and volume give indications of supply and demand that neither could provide independently.
hen securities change hands on a securities auction market, the volume of shares bought always matches the volume sold. When the price rises, the upward movement reflects that demand exceeds supply or buyers are in control. co ntrol. Likewise, when the price falls it implies that supply exceeds demand or that sellers are in control. Over time, these trends of supply and demand form accumulation and distribution patterns. What if there were a way to look deep inside price and volume trends to find out if current prices were supported by volume? This is the objective of the volume price confirmation indicator (VPCI), a methodology that measures the intrinsic relationship between price and volume. DERIVING THE COMPONENTS The basic VPCI concept is derived by examining the difference between a volume-weighted moving avVOLUME PRICE CONFIRMATION INDICATOR The VPCI exposes the relationship between the pre- erage (VWMAs) and the corresponding simple movvailing price trend and volume, as either confirming ing average (SMA). These differences expose inforor contradicting the price trend. This gives an indicaind ica- mation about the inherent relationship between price p rice tion of possible impending price movements. In this and volume. Although SMAs demonstrate a stock’s article I will discuss the derivation and components changing price levels, they do not reflect the amount of the VPCI and explain how to use it. I will also of investor participation. With VWMAs, however, review comprehensive testing of the VPCI and present price emphasis is adjusted proportionally to each further applications using the indicator. day’s volume and then compared to the average In exchange markets, price results from an agree- volume over the range of study. The VWMA is calcument between buyers and sellers despite their dif- lated by weighting each closing price with volume ferent appraisals of the exchanged item’s value. compared to the total volume during the range: One opinion may have legitimate fundamental grounds for evaluation, while the other may be pure Volume-Weighted Volume-Weigh ted Average = nonsense. To the market, however, both are equal. Sum {closing price (I) * [volume (I)/(total range)]} Price represents the convictions, emotions, and where I = given day’s action volition of investors. It is not a constant, but rather changed and influenced over time by information, Here’s an example of how to calculate a two-day moving average using both the SMA and VWMA for a security trading at $10 a share with 100,000 shares by Buff Pelz Dormeier, CMT changing hands on the first day, and at $12 a share
Reprinted from Technical Analysis of STOCKS &
INDICATORS
COMMODITIES magazine. © 2007 Technical Analysis Inc., (800) 832-4642 , http://www. traders.com CMX Last-Daily 12/22/2006 C=57.000 -6.590 -10.36% O=63.710 H=65.290 L=63.630 V=4706108 buffwave(close,0,true,0) 54.835 MovAvg 1 line(close,9,0) 53.499
57 56 55 54 53 52 51 50
with 300,000 shares changing hands on the se cond VPC = day. The SMA calculation is day 1’s price plus day VWMA – SMA VWMA 2’s price divided by the number of days, or (10+12)/ SMA 2, which equals 11. The V WMA calculation would be day 1’s price $10 multiplied by day 1’s volume, 49 which is expressed as a fraction of the total range: 48 Price-volume (100,000/400,000 = 1 / 4) plus day 2’s price $12 momentum 47 multiplied by day 2’s volume of the total range Volume 2974800.00 expressed as a fraction (300,000/400,000 = 3 / 4), 3M which equals 11.5 (2.5, day 1 + 9, day 2). 2M The VWMA measures investor commitments ex1M pressed through price, weighted by each day’s corresponding volume, compared to the total volDec 4 11 18 25 ume over time. Thus, volume-weighted averages FIGURE 1: DIFFERENCE BETWEEN SMA AND VWMA. The result of this calculation provides weight closing prices in exact proportion to the information about the relationship between price and volume. volume traded during each time period. We can begin investigating the V PCI keeping in mind how assume SMA’s short-term average volume for 10 days is 1.5 VWMAs work. The VPCI involves three calculations: million shares a day, and the long-term avera ge volume for 50 days is 750,000 shares per day. The VM equals 2 (1,500,000/ 1) Volume-price confirmation/contradiction (VPC+/-), 750,000). This calculation is then multiplied by the V PC+/after it has been multiplied by the VPR. 2) Volume-price ratio (V PR), and Now we have all the information necessary to calculate the 3) Volume multiplier (VM) VPCI. The VPC+ confirmation of +1.5 is multiplied by the V PR of 1.25, giving 1.875. Then 1.875 is multiplied by the VM of VWMA: Volume-weighted moving average VPC(+/-): Volume/price confirmation/contradiction 2, giving a VPCI of 3.75. Although this number is indicat ive of an issue under strong volume-price confirmation, this informaVPR: Volume/price ratio VM: Volume multiplier tion serves best relative to the current and prior price trend and relative to recent VPCI levels. In Figure 2 you can see the four The VPC is calculated by subtracting a long-term SMA from divisions of the price/volume relat ionship, and in Figure 4 you the same time frame’s VWMA. In essence, this calculation is the can see the VPCI in action. Next, I’ll discuss how to properly otherwise unseen nexus between price and price proportion- use the VPCI. ally weighted to volume. This difference , when positive, is the VPC+ (volume-price confirmation), and when negative, the USING THE VPCI VPC- (volume-price contradiction). This computation is the Price can be considered as the emotion, conviction, and voliintrinsic relationship between price and volume symmetrically tion of investors. Logically, you can define a price trend as the distributed over time. emotion, conviction, and volition of investors expressed over The result is revealing. For example, a 50-day SMA might be time. Generally, a buyer’s underlying emotion or motivation is $48.50, whereas the 50-day V WMA may be $50. The difference greed. Greed is the desire to obtain a profit. An uptrend could of 1.5 represents price-volume confirmation (V WMA – S MA) be viewed as an accumulation of greed over time. (see Figure 1). If the calculation were negative, it would Many times — but not always — an investor who creates represent price-volume contradiction. This calculation alone supply, a seller, is motivated by the fear of losing value in his provides purely unadorned information about the otherwise investment. Likewise, a downtrend would then be the accumuunseen relationship between price and volume. lation of fear over time. We also spoke of volume as the force The next step is to calculate the volume price ratio (VPR). that sustains price. A rising volume trend would represent a VPR accentuates the VPC+/- relative to the short-term price- buildup in energy or fuel. A decrease in volume would then volume relationship. The VPR is calculated by dividing the represent the loss of fuel — nonworking energy or entropy. short-term VWMA by the short-term S MA. For example, asGreed or an uptrend needs fuel to build and sustain itself. sume the short-term time frame is 10 days, and the 10-day Greed’s growth cannot be sustained without energy. An invesVWMA is $68.75, while the 10-day SMA is $55. The V PR would tor will lose interest and move on to better opportunities, equal 68.75/55, or 1.25. This factor will be multiplied by the whereas an investor who is a seller — maybe bea rish or fearful, VPC (+/-) calculated in the first step. Volume price ratios but not necessarily — could be motivated by greed and sell, greater than 1 increase the weight of the V PC+/-. Volume-price allowing participation in a more lucrative investment. Or the ratios less than 1 decrease the weight of the VPC+/-. seller could be motivated by greater emotions than greed, such The third and final step is to calculate the volume multiplier as lust or personal responsibilities. (VM). The VM’s objective is to overweight the V PCI when In such a case, the investor will sell his investment to buy volume is increasing and underweight the VPCI when volume material pleasures or satisfy his responsibilities. In this way, is decreasing. This is done by dividing the short-term average greed (bulls) need fuel (volume) to expa nd, but fear (bears) do volume by the long-term average volume. As an illustration, not necessarily need volume to fall. E
T
R
A
D
E
S
T
A
T
OI
N
P
R
O
S
U
TI
Reprinted from Technical Analysis of STOCKS &
COMMODITIES magazine. © 2007 Technical Analysis Inc., (800) 832-4642, http://www. traders.com Price contraction & volume contraction
Price expansion & volume expansion Trend Up & Volume Rising Greed w/ energy = Invigorated greed Phase 1
Price Volume
10 100
Strong demand
Phase 2
Phase 1
12 300
Price Volume
VWMA = 0.25 * 10 + 0.75 * 12 VPC = 11.5 (VWMA) – 11 (SMA) Price trend + 2 (rising) – VS – VPC = +0.5 (r ising) Uptrend confirmation
Bullish Trend Up & Volume Falling Greed w/ entropy = Complacency
Price Volume
Phase 1
Phase 2
10 300
12 100
Weak demand
Bearish
12 300
Price expansion & volume contraction
10 100
Downtrend contradiction
Bullish
Strong supply
Trend Down & Volume Rising Fear w/ energy = Fear
Price Volume
Uptrend contradiction
Phase 2
VWMA = 0.75 * 12 + 0.25 * 10 VPC = 11.5 (VWMA) – 11 (SMA) Price trend – 2 (falling) – VS – VPC = +0.5 (rising)
Phase 1
VWMA = 0.75 * 10 + 0.25 * 12 VPC = 10.5 (VWMA) – 11 (SMA) Price trend + 2 (rising) – VS – VPC = –0.5 (falling)
Weak supply
Trend Down & Volume Falling Fear w/ entropy = Apathy
12 100
Phase 2
10 300
VWMA = 0.25 * 12 + 0.75 * 10 VPC = 10.5 (VWMA) – 11 (SMA) Price – 2 (falling) – VS – VPC = –0.5 (falling)
Bearish
Downtrend confirmation
Price contraction & volume expansion
FIGURE 2: FOUR DIVISIONS OF PRICE/VOLUME RELATIONSHIPS. Here you can see the characteristics of the different price/volume relationships — that is, price expansion and volume expansion; price contracti on and volume contraction; price expansion and volume contraction; pri ce contraction and volume expansion.
CONFIRMING SIGNALS
CRL Last-Weekly 2/2/2007 C=45.960 -2.110 -4.39% 0=49.950 H=49.950 L=47.540 V=149200
54
Several VPCI signals may be employed in conjunction 50 with price trends and price indicators. These incl ude a 46 VPCI greater than zero, which shows whether the 42 relationship between price trends and volume con38 firms or contradicts the price trend. More importa nt, a 34 rising or falling VPCI provides the trend direction, revealing the direction of confirmation or contradic30 tion. A smoothed volume-weighted average of VPCI, 26 called “VPCI smoothed,” demonstrates how much the VPCI band (5,true,25,1.5,-1.5,DarkGreen,Red) VPCI has changed from previous V PCI levels and is used 0.00 to indicate momentum. Bollinger Bands may also be -2.00 applied to the VPCI, exposing VPCI extremes. -4.00 VPCI "V" bottom Fundamentally, the VPCI reveals the proportional imbalances between price trends and volume-adjusted Apr Jul Oct 2004 Apr Jul Oct 2005 Apr Jul Oct 2006 Apr Jul Oct price trends. An uptrend with increasing volume is a FIGURE 3: VOLUME PRICE CONFIRMATION INDICATOR (VPCI) ‘V’ BOTTOM. When the VCPI falls market characterized by greed supported by the fuel below the lower standard deviation of the Bollinger Band of the VCP I and then rises above the lower needed to grow. An uptrend without volume is compla- band, it forms a “V” bottom. cent and reveals greed deprived of the fuel needed to sustain itself. Investors without the influx of other investors (capitulation). These occurrences may be visualized by the (volume) will eventually lose interest, and the uptrend should VPCI falling below the lower standard deviation of a Bollinger eventually break down. Band of the VPCI, and then rising above the lower band, A falling price trend reveals a market driven by fear. A forming a “V” bottom (Figure 3). falling price trend without volume reveal s apathy, fear without It’s important to note when using the V PCI that volume leads increasing energy. Unlike greed, fear is self-sustaining and or precedes price action. Unlike most indicators , the VPCI often may endure for long periods without increasi ng fuel or energy. gives indications before price trends are clear. Thus, when a Adding energy to fear can be likened to adding fuel to a fire and is generally bearish until the V PCI reverses. In such cases, What if you could look inside price and weak-minded investors, overcome by fear, become irrationally fearful until the selling climax reaches a state of maxivolume trends to find out if current mum homogeneity. At this point, ownership held by weak prices were supported by volume? investors has been purged, producing a type of heat death
Reprinted from Technical Analysis of STOCKS &
COMMODITIES magazine. © 2007 Technical Analysis Inc., (800) 832-4642, http://www. traders.com
INDICATORS
140 130 120 110
TM Last-Weekly 2/16/2007 C=136.770 +11.900 +9.53% 0=124.790 H=124.790 L=123.620 V=562800
100 Uptrend VPCI signal is given in an unclear price trend, it is confirmation Uptrend Downtrend 90 contradiction Downtrend contradiction best to wait until one is evident. confirmation 80 At point 1 in Figure 4, Toyota Motor (TM) is Uptrend confirmation 2.50 breaking out of a downtrend and the V PCI confirms VPCI band (5,true,3,2,DarkGreen,Red) 0.42 1.07 Downtrend 1.50 VPCI warning signal this breakout immediately as the indicator rises, contradiction 0.50 1 crossing over the V PCI smoothed and then the zero 2 3 Zero line -0.50 bullish line. This is an example of VPCI’s bullish confir- Volume 2137390.00 VPCI VPCI bullish indication VPCI "V" bottom 3.5M confirmation mation of a price trend. Later, the V PCI begins to 2.5M 1.5M fall during the uptrend, suggesting complacency. Apr May Jun Jul Aug Sep Oct Nov Dec 2006 Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 2007 Feb By point 2, the VPCI crosses under the VPCIsmoothed warning of a possible pause within the FIGURE 4: PUTTING IT ALL TOGETHER. Here you see an example of the VPCI in action. See Fi gure new uptrend. This is a classic example of a VPCI 2 for details. bearish contradiction. Before we reach point 3, the VPCI makes a pattern forming a “V” bottom. This is a bullish VPCI smoothed. OBV crossovers of O BV smoothed would give sign, often indicating the selloff has washed out many of the indications of OBV rising relative to previous O BV levels. sellers. At point 3, the VPCI confirms the earlier bullish “V” Remember, VPCI is designed for application in a trending pattern with a bullish crossover, leading to a strong bull rally. market, with a trending indicator. Thus, we need two additional tools to complete this test. First, we’ll need an indicator to verify whether we are in a VPCI VS. OTHER PRICE trending market. A seven-period average directional moveVOLUME INDICATORS ment index (ADX) indicator fulfills this criterion by indicating The most acclaimed price volume the intensity of the trend. Next, we will need a trend indicator indicator to compare to the V PCI is to show the trend’s direction. The moving average converJoseph Granville’s original on-bal- gence/divergence (MACD) with the traditional (12, 26, 9) setance volume (OBV) indicator. Rec- tings was used to provide buy entry signals for this test. Finally, ognizing volume as the force behind we will need a test subject, which illustrates how these indicators price, Granville created OBV by assigning up days as positive work across a broad market. I applied it to the SPDR Standard & volume (measured by an up close) and then subtracting volume Poor’s 500 exchange traded fund (S PY). See Figure 5. on down days. O BV is price-directed volume, the accumulation The testing period was conducted from inception (February of +/- volume flows based upon price direction. Granville’s 1993) until the end of 2006. Standard specifications were used objective with on-balance volume was to uncover hidden coils on both indicators (O BV – 28 day and V PCI 7/28 [seven-day in an otherwise noneventful, nontrending market. With his short-term trend and (5)(4)-day long-term trend]). Results OBV indicator, Joe Granville became a renowned market were not optimized in any way. strategist. In so doing, he popularize d OBV and the wisdom of In this system, long positions are taken only when these using volume in securities analysis. conditions are met when accompanied by OBV crossovers in The VPCI differs from O BV in that it calculates the propor- the first test, or by V PCI crossovers in the second. Long tional imbalances between price trends and volume-weighted positions are exited with crossunders of O BV smoothed in the price trends. This exposes the influence volume has upon a first test or with V PCI crossunders in the second study. Alprice trend. Although both contain volume-derived data, they though this test was created for both observational and creditconvey different information. In composition, the V PCI is not ability purposes, the results are stunning. an accumulation of history like OBV but a snapshot of the influence of volume upon a price trend over SPY Last-Daily 4/22/2005 C=114.953 -32.277 -21.92% O=148.220 H=148.620 L=147.044 V=124128536 122 a specified time. This enables the VPCI to give 120 faster signals than accumulation indicators similar 118 to an oscillator. In contrast to O BV, the VPCI’s 116 objective is not to uncover hidden coils in trendless VPCI 114 sell VPCI buy VPCI (5,true,3,2,DarkGreen,Red) 0.4610 -0.2539 markets, but to evaluate the health of existing -0.15 trends.
-0.35
COMPARING VPCI
TO OBV
The most general VPCI buy signal is when the V PCI crosses above the VPCI smoothed in an uptrending market. This indicates the VPCI is rising relative to previous VPCI levels. The traditional OBV does not have a lagging trigger like the VPCI smoothed, so I amended the OBV by adding an additional eightperiod simple moving average of O BV. The net effect gives OBV a corresponding trigger to the
MACD buy
MACD (12,26,9) -1.05 -0.90 -0.15
0.60 0.00 -0.60
OBV sell 1.9M
On-balance volume SM 1631285632.00 1690542464.00
OBV buy
48.0 36.0
ADX (7) 45.95 ADX>30
24
31 Feb 7
14
21
28 Mar 7
14
21
28
1.7M 1.5M
Apr
11
18
FIGURE 5: VPCI/ON-BALANCE VOLUME (OBV) COMPARISON. Here you see the VPCI compared to the OBV, the seven-period ADX, and the MACD.
Reprinted from Technical Analysis of STOCKS &
INDICATORS
COMMODITIES magazine. © 2007 Technical Analysis Inc., (800) 832-4642 , http://www. traders.com Strategy*
Annual return
Time invested
Std dev
5-yr % Profitable Profit Sharpe factor ratio
The results of this test are displayed in Figure 6. Buy-Hold 9.94% 100.00% 17.75% 0.10 N/A N/A MACD -3.88% 24.79% 13.03% 0.27 41.79% 0.97 Excluding dividends or interest, O BV’s annualized VPCI 8.11% 35.63% 7.42% 0.74 65.15% 2.47 rate of return in the above system was -1.57%, whereas OBV -1.57% 27.02% 17.40% 0.05 42.86% 1.00 the V PC I ’s annualized return was 8.11%, an *Dividends not included outperformance of over 9.5% annualized. The V PCI improved reliability, giving profitable signals over FIGURE 6: COMPARING RETURNS OF DIFFERENT STRATEGIES. The results of applying the 65% of the time, compared to OBV at only 42.86%. various indicators are compared here. You can see the annual return of applying the VPCI is more favorable than the returns of the other strategies. Another consideration in evaluating performance is risk. The VPCI had less than half the risk as measured by volatility, 7.42 standard deviations compared to O BV with Strategy* 1993 1994 1995 1996 1997 1998 1999 17.4 standard deviations from the mean. It is not surprising Buy-Hold 3.61 -2.21 34.95 20.1 31.44 27.04 19.11 the VPCI had much better risk-adjusted rates of return. The MACD 0.31 5.42 0.88 12.63 14.72 -12.93 -30.6 VPCI 2.93 6.42 6.12 19.83 19.09 8.9 3.17 VPCI’s Sharpe ratio from inception was 0.70 and had a profit OBV -1.03 -1.24 0 18.81 7.3 12.4 -12.43 factor of 2.47, compared to O BV with a -0.09 Sharpe ratio and a profit factor of less than 1. Strategy* 2000 2001 2002 2003 2004 2005 2006 Admittedly, this testing environment is an uneven match. Buy-Hold -10.68 -12.87 -22.81 26.12 8.62 3.01 13.74 The VPCI uses information from volume-weighted prices to MACD 9.27 11.12 0.9 1.34 1.8 1.69 -11.32 VPCI -2.71 21.28 -0.65 10.4 4.27 4.8 9.29 gauge the health of existing trends, whereas O BV accumulates OBV -26.55 -28.34 12.45 -12.79 33.32 -15.58 -8.33 volume flows as directed by price changes to uncover hidden *Annual rates of return without dividends coils. Thus, the conditions setup in this system, a trending market with apparent price direction, is one in which the VPCI FIGURE 7: ANNUAL RETURNS. Here you see the annual returns of the various strategies from 1993 to 2006. Note that the worst annualized VPCI return was only is designed to succeed. Although O BV was not necessarily set –2.71%. up for failure either, this study does illustrate how less-savvy practitioners often fail to use the indicators’ information correctly or coordinate the indicators properly. calculation. This would lower the price stop when price and What if an investor had just used the M ACD buy & sell volume are in confirmation, increasing the probability of signals within this same system without utilizing the VPCI keeping an issue under accumulation. However, when price information? The investor would have lost out on nearly 12% and volume are in contradiction, dividing the stop-loss by the annualized return, the difference between the V PCI’s positive VPCI would raise the stop price, preserving more capital. 8.11% versus the MACD’s negative -3.88% rate of return while Similarly, using VPCI as a multiplier to other price, volume, significantly increasing risk. What if this investor had just and momentum indicators may not only improve relia bility but employed a buy & hold approach? Although this investor increase responsiveness as well. would have realized a slightly higher return, he or she would have been exposed to greater risks. The V PCI strategy returned NOW THAT YOU’VE DUG DEEP nearly 90% of the buy & hold strategy return while avoiding The VPCI reconciles volume and price as determined by each about 60% less risk as measured by standard deviation. of their proportional weights. This informat ion may be used to Looking at risk-adjusted returns another way, the five-year deduce likelihood of a current price trend conti nuing or reversSharpe ratio for the SPDR was only 0.1 compared to the V PCI ing. This study demonstrates that adding the V PCI to a trendsystem of 0.74. In addition, the VPCI investor would have been following system resulted in improved performance across all invested only 35% of the time, allowing the investor the major areas measured. In the hands of a proficient investor, the opportunity to invest in other investm ents. During 65% of the volume price confirmation indicator provides information time the investor was not invested, he or she would have only useful in accelerating profits, reducing risk, and empowering needed a 1.84% money market yield to exceed the buy & hold the investor toward sound investment decisions. strategy. Moreover, this investor would have experienced a much smoother performance, without such precipitous capital Armed with his proprietary indicators and investment prodrawdowns. The worst annualized V PCI return was only a grams, Buff Pelz Dormeier now advises and manages portfolios measly -2.71% compared to the underlying investments’ worst for both individual and inst itutional clients at Wachovia Secuyear of -22.81%, more than 20% difference in the rate of rities. He may be reached at
[email protected]. return! (See Figure 7.) If an investor had invested in a moneymarket instrument, while not invested in the S PDR, this VPCI The original version of this work was awarded the 2007 Charles H. Dow Award, sponsored by the Market Technicians strategy would not have experienced a single down year.
OTHER
APPLICATIONS
The raw VPCI calculation may be used as a multiplier or divider in conjunction with other indicators, such as moving averages, momentum indicators, or price and volume data. If an investor has a trailing stop-loss order set at the five-week moving average of the lows, you could divide the stop price by the V PCI
Association, as the research paper that broke new ground or made innovative use of established techniques for the year. Congratulations, Buff! See our Traders’ Tips area at http://www.traders.com/Documentation/FEEDbk_docs/Archive/072007/TradersTips/TradersTips.html for program code implementing Buff Dormeier’s technique. S&C