Add training workflow, datasets, and runbook

training_data/curated/text/aa736818e018c21e562b7ff4ae915576194c49a411b43f4be829a55e13311671.txt

@@ -0,0 +1,37 @@
+472 Part IV: Additional Considerations 
+spreads, some naked writes and ratio calendar spreads, fewer straddles and ratio 
+writes, and a few covered call writes. This theory would be somewhat difficult to apply 
+in practice, because of the massive numbers of calculations involved and also because 
+of the accuracy of closing price data. It was mentioned previously that a computer will 
+assume that "bad" closing prices are actually attainable. By a "bad" closing price, it is 
+meant that the option did not trade simultaneously with the stock later in the day, and 
+that the actual market for the option is somewhat different in price than is reflected 
+by the closing price for the option. A daily contract volume "screen" will help allevi­
+ate this problem. For example, one may want to discard any option from his calcula­
+tions if that option did not trade a predetermined, minimum number of contracts 
+during the previous day. Data that give closing bids and offers for each option are 
+more expensive but also more reliable, and would alleviate the problem of "bad" 
+closing prices. In addition to a volume screen, another way of reducing the calcula­
+tions required is to limit oneself to strategies in which one has interest, or which one 
+is reasonably certain will fit in well with his investment objectives. Regardless of the 
+limitations that one places upon the quantity of computations, some computer power 
+is necessary to compute expected return. A sophisticated programmable calculator 
+may be able to provide a real-time calculation, but could never be used to evaluate 
+the entire option universe and come up with a ranking of the preferable situations 
+each day. On-line computer systems are also available that can provide these types of 
+calculations using up-to-the-minute prices. While real-time prices may occasionally 
+be useful, it is not an absolute necessity to have them. 
+One other by-product of the expected return calculation is that it could be used 
+as another model for predicting the theoretical value of an option. All one would have 
+to do is compute the probabilities of the stock being at each successive price above 
+the striking price of the option by expiration, and sum them up. The result would be 
+the theoretical option value. These data are published by some services and general­
+ly give a different theoretical value than would the Black-Scholes model. The reason 
+for the difference most readily lies in the inclusion of the risk-free interest rate in the 
+Black-Scholes model and its omission in the expected return model. 
+APPLYING THE CALCULATIONS TO STRATEGY DECISIONS 
+CALL WRITING 
+One method of ranking covered call writes that was described in Chapter 2 was to 
+rank all the writes that provided at least a minimal acceptable level of return by their 
+probability of not losing money. If one were interested in safety, he might decide to 
+use this approach. Suppose he decided that he would consider any write that provid-