Add training workflow, datasets, and runbook

training_data/relevant/text/98bf203ff5dbb2e929ac7e7bd1f4e413bb5aec81fd9de3d17a7c19cf34316997.txt

@@ -0,0 +1,39 @@
+Chapter 39: Volatility Trading Techniques 815 
+ing, he is generally talking about implied volatility. Thus, it makes sense to know where 
+implied volatility is within the range of the past readings of implied volatility. If volatil­
+ity is low with respect to where it usually trades, then we can say the options are cheap. 
+Conversely, if it's high with respect to those past values, then we can say the options are 
+expensive. These conclusions do not draw on historical volatility. 
+The percentile of implied volatility is generally used to describe just where the 
+current implied volatility reading lies with respect to its past values. The "implied 
+volatility" reading that is being used in this case is the composite reading - the one 
+that takes into account all the options on an underlying instrument, weighting them 
+by their distance in- or out-of-the-money (at-the-money gets more weight) and also 
+weighting them by their trading volume. This technique has been referred to many 
+times and was first described in Chapter 28 on mathematical applications. That com­
+posite implied volatility reading can be stored in a database for each underlying 
+instrument every day. Such databases are available for purchase from firms that spe­
+cialize in option data. Also, snapshots of such data are available to members of 
+www.optionstrategist.com. 
+In general, most underlying instruments would have a composite implied 
+volatility reading somewhere near the 50th percentile on any given day. However, it 
+is not uncommon to see some underlyings with percentile readings near zero or 100% 
+on a given day. These are the ones that would interest a volatility trader. Those with 
+readings in the 10th percentile or less, say, would be considered "cheap"; those in the 
+90th percentile or higher would be considered expensive. 
+In reality, the percentile of implied volatility is going to be affected by what the 
+broad market is doing. For example, during a severe market slide, implied volatilities 
+will increase across the board. Then, one may find a large number of stocks whose 
+options are in the 90th percentile or higher. Conversely, there have been other times 
+when overall implied volatility has declined substantially: 1993, for example, and the 
+summer of 2001, for another. At those times, we often find a great number of stocks 
+whose options reside in the 10th percentile of implied volatility or lower. The point 
+is that the distribution of percentile readings is a dynamic thing, not something stat­
+ic like a lognormal distribution. Yes, perhaps over a long period of time and taking 
+into account a great number of cases, we might find that the percentiles of implied 
+volatility are normally distributed, but not on any given day. 
+The trader has some discretion over this percentile calculation. Foremost, he 
+must decide how many days of past history he wants to use in determining the per­
+centiles. There are about 255 trading days in a year. So, if he wanted a two-year his­
+tory, he would record the percentile of today's composite implied volatility with 
+respect to the 510 daily readings over the past two years. This author typically uses