Add training workflow, datasets, and runbook

training_data/curated/text/a7323c3726d51400eb09863d294c4a662af74c1c6397995749259886a0d46e6d.txt

@@ -0,0 +1,28 @@
+Calculating Volatility Data
+Accurate data are essential for calculating volatility. Many of the volatility
+data that are readily available are useful, but unfortunately, some are not.
+HV is a value that is easily calculated from publicly accessible past closing
+prices of a stock. It’s rather straightforward. Traders can access HV from
+many sources. Retail traders often have access to HV from their brokerage
+firm. Trading firms or clearinghouses often provide professional traders
+with HV data. There are some excellent online resources for HV as well.
+HV is a calculation with little subjectivity—the numbers add up how they
+add up. IV, however, can be a bit more ambiguous. It can be calculated
+different ways to achieve different desired outcomes; it is user-centric. Most
+of the time, traders consider the theoretical value to be between the bid and
+the ask prices. On occasion, however, a trader will calculate IV for the bid,
+the ask, the last trade price, or, sometimes, another value altogether. There
+may be a valid reason for any of these different methods for calculating IV.
+For example, if a trader is long volatility and aspires to reduce his position,
+calculating the IV for the bid shows him what IV level can be sold to
+liquidate his position.
+Firms, online data providers, and most options-friendly brokers offer IV
+data. Past IV data is usually displayed graphically in what is known as a
+volatility chart or vol chart. Current IV is often displayed along with other
+data right in the option chain. One note of caution: when the current IV is
+displayed, however, it should always be scrutinized carefully. Was the bid
+used in calculating this figure? What about the ask? How long ago was this
+calculation made? There are many questions that determine the accuracy of
+a current IV, and rarely are there any answers to support the number.
+Traders should trust only IV data they knowingly generated themselves
+using a pricing model.