Files
ollama-model-training-5060ti/training_data/relevant/text/f225092c496be381732d9d3eca65e23f62cbb491793cfe72a77843e96d1494e1.txt

36 lines
1.9 KiB
Plaintext
Raw Permalink Blame History

This file contains invisible Unicode characters
This file contains invisible Unicode characters that are indistinguishable to humans but may be processed differently by a computer. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.
Chapter 38: The Distribution of Stock Prices 801
back over the last 1,000 trading days for XYZ. A 100-day historical volatility can be
computed, using 100 consecutive trading days of data, for 901 of those days (begin­
ning with the 100th day and continuing through the l,000th day, which is presumably
the current trading day). Admittedly, these are not completely unique time periods;
there would only be ten non-overlapping (independent) consecutive 100-day periods
in 1,000 days of data. However, let's assume that the 901 periods are used. One can
then arrive at a distribution of 100-day historical volatilities. Suppose it looks some­
thing like this:
Percentile 100-Day Historical
oth 34%
10th 37%
20th 43%
30th 45%
40th 46%
50th 48%
60th 51%
70th 58%
aoth 67%
90th 75%
1 ooth 81%
In other words, the 901 historical volatilities (100 days in each) are sorted and then
the percentiles are determined. The above table is just a snapshot of where the per­
centiles lie. The range of those 901 volatilities is from 34% on the low side to 81 % on
the high side. Notice also that there is a very flat grouping from about the 20th per­
centile to the 60th percentile: The 100-day historical volatility was between 43% and
51 % over that entire range. The median of the above figures is 48% - the 100-day
volatility at the 50th percentile.
Referring to the early part of this example, the current 100-day historical is
80%, a very high reading in comparison to what the measures were over the past
1,000 days, and certainly much higher than the median of 48%.
One could perform similar analyses on the 1,000 days of historical data to deter­
mine where the 10-day, 20-day, and 50-day historical volatilities were over that time.
Those, too, could be sorted and arranged in percentile format, using the 50% per­
centile (median) as a good estimate of volatility. After such computations, the trader
might then have this information: