The ADR can be a great tool in identifying stocks or tradable assets that are currently experiencing high levels of volatility. This can signal an increase in volume, response to events like news, or abrupt changes in price action. Before we dive in, let’s first talk a bit more about what the ADR is conceptually and how it’s used.<\/p>\n
What’s the ADR<\/h2>\n
The ADR is a measure of volatility and is useful to describe whether an assets price action is outside the normal during a particular time. The daily range<\/strong> (highest price – lowest) can be compared to the ADR over a previous interval to signal potential entry and exit points for traders.<\/p>\n The ADR can be used over whatever interval one chooses, though a 20-day period is common. In stock market technical analysis, this period uses the previous 19 values\u2014along with the current value\u2014to calculate an averaged number for the daily trading range. This can help signal abrupt changes in price action and is useful in stock screeners.<\/p>\n The Average Daily Range is calculated similarly to how one calculates a simple moving average<\/a>. The difference between the Average Daily Range and a simple moving average is that we must pre-calculate the difference between the high and low for each period<\/strong>. After these values are calculated one can calculate the ADR. The average daily range formula is as follows:<\/p>\n Equations can be a bit heady and often confuse otherwise simple calculations. An\u00a0 iterative approach to calculating the ADR is outlined as such:<\/p>\n Let’s consider an illustration of the above. First, anyone unfamiliar with candlestick representation of stock prices should check out this article<\/a>. In the illustration below we can see the Daily Range calculated for each period over a 5-day interval (K) and then added together and divided by K. This gives us the ADR for that period.<\/p>\n This image depicts a stock price that is gradually becoming more volatile over a 5-day period. The first period has a Daily Range of 2 while the last two periods have a daily range of 4. The average daily range here is 3.2 \u2014accounting for the daily range of each period.<\/p>\n We’ve now seen the mathematical basis for calculating the ADR. Modern trading doesn’t rely on pencil and paper, however. To deploy the ADR as an analytical tool we will implement our logic so far as a Python function. First, let’s create a representation of the price action shown in the image above.<\/p>\n In practice, it would be practical to have this type of data stored as OHLCV in a pandas DataFrame object<\/a>. We’re using a list of 5 dictionary objects here for syntactic clarity. Keeping in mind our algorithm outlined above, we can implement the ADR in Python as follows:<\/p>\n Here we see our result matches up to the manual calculation from the illustrated process above. The example code above is a bit bloated\u2014mostly for sake of clarity\u2014so we’ll tighten things up a bit before we start looking at more practical examples:<\/p>\n This one-liner will make calculating the ADR for some real-world data a bit less syntactically cumbersome. To showcase the utility of the ADR we will fetch the latest stock pricing data and use the resulting We are going to use the Here we have queried the Yahoo Finance API (sort of) and retrieved the previous 6-months pricing data for For each period in our interval, we need to calculate the daily range from the High – Low calculation. This is done using the apply method of the DataFrame object as such:<\/p>\n We now have a third column reflecting the difference in our daily High and daily Low. Next, we need to calculate the average of these values over a preceding period. To do this\u2014effectively calculating a moving average\u2014we will use the rolling method of the DataFrame.<\/p>\n Before we calculate the Average Daily Range we need to define our lookback period. A common value for this is 20 periods and what we will use here. This decision means that every ADR value we calculate is reflective of the previous 19 periods of price action plus the range of the current period. This can be achieved via the following code:<\/p>\n A few things to note: first, we see a new column has been added for the ADR. Secondly\u2014and most important\u2014we see a value of We need a minimum of 20 values to calculate the ADR since that is how we defined our lookback window<\/strong>. For example, if we were to print out rows 18-22 we can see where the transition takes place:<\/p>\n In row The ADR is a great measure of volatility and can help signal a great opportunity to traders and investors. However, it has the shortcoming of being very specific to the price action of a single stock. For example, the ADR of What happens when we try to compare To use the ADR as a screening tool, such that one can compare volatility among different assets equally, we need to modify the ADR formula to reflect a percentage<\/em> change rather than a fixed<\/em> dollar change<\/strong>.\u00a0 For this, we need to adjust our algorithm as follows:<\/p>\n This gives us an ADR representing a percentage change in price which allows comparison between assets with wildly different prices. The following image reflects the update to our ADR algorithm:<\/p>\n We can implement this modified ADR using the following code (using our initial sample data again):<\/p>\n Note that, once again, we arrive the same value as illustrated in the image. This is good news\u2014indicating our math is correct\u2014but we’re once again left with a syntactically cumbersome algorithm. Let’s use our Looking at the last few rows of data we see a modified ADR value in the range of The ADR is a measure of volatility. Traditional ADR values represent dollar changes of an asset and are not comparable to other assets. Conceivably, one could make comparisons of assets with very similar prices.<\/p>\n The modified ADR we’ve discussed here can be used to make observations such as the following:<\/p>\n Keep in mind the ADR and the modified ADR are but tools to analyze price action. Any trading strategy based solely on their indication isn’t likely to be successful. Remember, this technical indicator is available to all other traders on the market as well.<\/p>\n Using it in conjunction with other technical indicators such as the RSI<\/a>, Bollinger bands<\/a>, Mean Average Convergence Divergence (MACD)<\/a>, or simple moving averages<\/a> can give a much more informed perspective of the movement of an underlying asset in the future.<\/p>\n The Average Daily Range reflects the average change in intraday prices over a period of time. The Average True Range is similar but takes into account any gaps in pricing between the closing of one trading period and the opening of the next.<\/p>\n Where ADR uses the High and Low of each period to make a calculation the ATR takes into account the High and Low relative to the previous closing price. This accounts for both gaps up and down in price.<\/p>\nCalculating the ADR<\/h2>\n
![\"Average](\"https:\/\/www.alpharithms.com\/wp-content\/uploads\/2283\/Average-Daily-Range-formula-alpharithms.jpg\" "\\"Average-Daily-Range-formula-alpharithms\\"")
<\/a>\n
\n
![\"Average](\"https:\/\/www.alpharithms.com\/wp-content\/uploads\/2283\/Average-Daily-Range-visualization-alpharithms-1.jpg\" "\\"Average-Daily-Range-visualization-alpharithms\\"")
<\/a>Calculating the Average Daily Range in Python<\/h2>\n
# Five prices as High and Low for the\r\n# five day period K\r\nprices = [\r\n {\"high\": 5, \"low\": 3},\r\n {\"high\": 7, \"low\": 4},\r\n {\"high\": 9, \"low\": 6},\r\n {\"high\": 7, \"low\": 3},\r\n {\"high\": 8, \"low\": 4},\r\n]\r\n\r\n# Example access\r\nday_one_high = prices[0][\"high\"]\r\nday_one_low = prices[0][\"low\"]<\/pre>\n# Create list for Daily Range values\r\ndaily_ranges = []\r\n\r\n# Get High - Low (DR) for each period\r\nfor price in prices:\r\n\r\n # Calculate DR\r\n pct_daily_range = price[\"low\"] \/ price[\"high\"]\r\n\r\n # Normalize\r\n daily_ranges.append(100 * (pct_daily_range - 1))\r\n\r\n# Calculate Average Daily Range\r\naverage_daily_range = sum(daily_ranges) \/ len(daily_ranges)\r\n\r\n# View output\r\nprint(average_daily_range)\r\n\r\n>>> 3.2<\/pre>\n
# calculate ADR from list of (high, low) tuples\r\nsum([x[0] - x[1] for x in prices]) \/ len(prices)<\/pre>\n
DataFrame<\/code> object for our ADR calculations.<\/p>\nADR Example<\/h2>\n
yfinance<\/code> library to fetch pricing data for Ether ($ETH-USD<\/code>) over the past 6 months. We can get this data with the following code:<\/p>\n# Get pricing data\r\nimport yfinance as yf\r\n\r\n# Get the previous 6 months pricing data for ETH and limit to High, Low\r\nETH = yf.Ticker('ETH-USD').history(period='6mo', interval=\"1d\")[[\"High\", \"Low\"]]\r\n\r\n# View resulting dataframe\r\nprint(ETH)\r\n\r\n High Low\r\nDate \r\n2021-05-18 3562.465088 3246.404053\r\n2021-05-19 3437.935791 1952.460205\r\n2021-05-20 2993.145264 2170.229004\r\n2021-05-21 2938.205078 2113.347168\r\n2021-05-22 2483.983154 2168.124268\r\n... ... ...\r\n2021-11-14 4689.842773 4516.935547\r\n2021-11-15 4764.636230 4546.599121\r\n2021-11-16 4891.704590 4144.334961\r\n2021-11-17 4300.306641 4107.125977\r\n2021-11-18 4343.557129 4138.691406\r\n\r\n[185 rows x 2 columns]<\/pre>\n$ETH-USD<\/code> and removed all data but the High and Low value for each day, while also keeping the date as the index values.<\/p>\nStep 1: Calculate Daily Range Values<\/h3>\n
# Calculate Daily Range for each period\r\nETH['dr'] = ETH.apply(lambda x: x[\"High\"] - x[\"Low\"], axis=1)\r\n\r\n\r\n# Result\r\n High Low dr\r\nDate \r\n2021-05-18 3562.465088 3246.404053 316.061035\r\n2021-05-19 3437.935791 1952.460205 1485.475586\r\n2021-05-20 2993.145264 2170.229004 822.916260\r\n2021-05-21 2938.205078 2113.347168 824.857910\r\n2021-05-22 2483.983154 2168.124268 315.858887\r\n... ... ... ...\r\n2021-11-14 4689.842773 4516.935547 172.907227\r\n2021-11-15 4764.636230 4546.599121 218.037109\r\n2021-11-16 4891.704590 4144.334961 747.369629\r\n2021-11-17 4300.306641 4107.125977 193.180664\r\n2021-11-18 4343.557129 4052.962158 290.594971\r\n\r\n[185 rows x 3 columns]<\/pre>\n
Step 2: Calculate Average Daily Range<\/h3>\n
# Add a simple moving average\r\nETH[\"adr\"] = ETH['dr'].rolling(window=20).mean()\r\n\r\n# View result\r\n High Low dr adr\r\nDate \r\n2021-05-18 3562.465088 3246.404053 316.061035 NaN\r\n2021-05-19 3437.935791 1952.460205 1485.475586 NaN\r\n2021-05-20 2993.145264 2170.229004 822.916260 NaN\r\n2021-05-21 2938.205078 2113.347168 824.857910 NaN\r\n2021-05-22 2483.983154 2168.124268 315.858887 NaN\r\n... ... ... ... ...\r\n2021-11-14 4689.842773 4516.935547 172.907227 217.784058\r\n2021-11-15 4764.636230 4546.599121 218.037109 219.570532\r\n2021-11-16 4891.704590 4144.334961 747.369629 238.494238\r\n2021-11-17 4300.306641 4107.125977 193.180664 228.781030\r\n2021-11-18 4343.557129 4052.962158 290.594971 234.109387\r\n\r\n[185 rows x 4 columns]\r\n<\/pre>\n
NaN<\/code> for the first entries. This is because there are fewer than 19 (20 – 1) preceding values for these dates.<\/p>\n# Print rows # 18-21\r\nprint(ETH.iloc[17:21])\r\n\r\n High Low dr adr\r\nDate \r\n2021-06-04 2857.165527 2562.637451 294.528076 NaN\r\n2021-06-05 2817.484863 2558.233643 259.251221 NaN \r\n2021-06-06 2743.441162 2616.162354 127.278809 435.919141\r\n2021-06-07 2845.185059 2583.995117 261.189941 433.175586<\/pre>\n
#20<\/code> for the date 2021-06-06<\/code>, we have our first DR value where there are enough preceding values to calculate the average using our 20-period lookback window.<\/p>\nModified ADR<\/h2>\n
$ETH-USD<\/code> during the past several days has been around $220USD.<\/p>\n$ETH-USD<\/code>‘s ADR to an asset whose price is between $5USD and $10USD? It’s certainly not going to be taking $220 swings! However, the\u00a0percent change<\/em> in price<\/a> would still be comparable.<\/p>\n\n
![\"modified](\"https:\/\/www.alpharithms.com\/wp-content\/uploads\/2283\/modified-average-daily-range-alpharithms.jpg\" "\\"modified-average-daily-range-alpharithms\\"")
<\/a># Get High\/Low for each period\r\ndaily_ranges = [d[1] \/ d[0] for d in data]\r\n\r\n# Get SMA of daily Ranges\r\nsma = mean(daily_ranges)\r\n\r\n# Calculate\r\nmodified_adr = 100 * (sma - 1)\r\n\r\n>>> 85.0<\/pre>\n
$ETH-USD<\/code> DataFrame<\/code> and employ a once-again-simplified version of our modified ADR:<\/p>\n# Get pricing data\r\nimport yfinance as yf\r\n\r\n# Get the previous 6 months pricing data for ETH\r\nETH = yf.Ticker('ETH-USD').history(period='6mo', interval=\"1d\")[[\"High\", \"Low\"]]\r\n\r\n# Calculate Daily Range for each period and normalize as pct change\r\nETH['dr_pct'] = ETH.apply(lambda x: 100 * (x[\"High\"] \/ x[\"Low\"] - 1), axis=1)\r\n\r\n# Calculate the average daily range over a 20-period interval\r\nETH[\"mod_adr\"] = ETH['dr_pct'].rolling(window=20).mean()\r\n\r\n# View Result\r\nprint(ETH)\r\n\r\n High Low dr_pct mod_adr\r\nDate \r\n2021-05-18 3562.465088 3246.404053 9.735727 NaN\r\n2021-05-19 3437.935791 1952.460205 76.082246 NaN\r\n2021-05-20 2993.145264 2170.229004 37.918407 NaN\r\n2021-05-21 2938.205078 2113.347168 39.030876 NaN\r\n2021-05-22 2483.983154 2168.124268 14.568302 NaN\r\n... ... ... ... ...\r\n2021-11-14 4689.842773 4516.935547 3.827976 5.057539\r\n2021-11-15 4764.636230 4546.599121 4.795609 5.075362\r\n2021-11-16 4891.704590 4144.334961 18.033524 5.507737\r\n2021-11-17 4300.306641 4107.125977 4.703549 5.246916\r\n2021-11-18 4343.557129 3985.766113 8.976719 5.480348\r\n\r\n[185 rows x 4 columns]<\/pre>\n5-5.48<\/code>. This reflects an average daily price range<\/strong> as a percentage<\/em> movement which, once again, requires the first 19 dr_pct<\/code> values to have been calculated before the first ADR value will be calculated.<\/p>\nInterpretation<\/h2>\n
\n
Average Daily Range (ADR) vs. Average True Range (ATR)<\/h2>\n
Final Thoughts<\/h2>\n