Just found these step by step instructions: https://www.elitetrader.com/et/threads/jack-hershey-method-since-2008.241615/page-4#post-3531664 Some more googling the different terms would yield more specifics.
YVW, the pdf is based on the Simplemind’s mind mapping software. Each of the nodes are hyperlinked. I believe I can export the mindmap directly so that’s it’s functional if one has their own copy of Simpleminds. @NeedToLearn, I’ll respond more directly when I have some time. In short if they have determination and some programming skills PVT is easily replicatable. If not, IBD’s many offerings would be more accessible. Their hot lists have the RS and EPS ranking as a standard parameter. My PVT software was running based on yahoo data, it broke when yahoo changed their api. The code was simple datascraping and UI with vba excel (for Mac) and filtering and sorting lists with sql. I can offer it up as a structural framework for anyone inclined. I haven’t gotten around to updating it yet since my interests have been primarily on SCT and RDBMS.
Decided to do a simple volatility study the way Makosgu laid it out years ago. The study is based on last year end of May to end of June. I purposely omitted some of the high volume bars from first bar and last bar of the day (RTH).
Another one for recent month. Above 20,000ish volume appears to be much more random in bar lengths. Did not omit first or last bar for this one.
I believe that this is one of those situations in which data seems to show something significant, but actually it doesn't; neither if restricted to a single day, nor when spanned over a period of time. If "volume leads price", you can't expect to have volume information reflected exclusively in its corresponding price bars' length.
I've been trying to wrap my head around this. Can you elaborate a bit on "volume leads price"? Are we anticipating a certain length of a segment with volume? If not, is it a Jokari window (i.e. continue vs change) reference? The above exercise was an attempt to gauge a bar's potential volatility using PRV. I believe @tiddlywinks posted one similar at one point and mentioned one can do well with it. Makosgu's explanation linked here as well https://www.elitetrader.com/et/threads/question-for-grob-hershey.54894/page-228#post-1034264
Yes, this is correct... My calculations are inspired by Maks original Vol vs Vola tables, with one big difference... I use percentile (or decile) to determine probable/likely volatility (bar price-range) given volume of X. In essence volume is the bucket (volume leads price), not the precise price volatility. Maks table uses volume ranges to calculate a simple frequency distribution of price ranges. Percentile is significantly more usable IMO, especially for real-time anticipations/expectations. The volume ranges in Maks table is an error in the analysis. IMO. I have (sort of) reverse-engineered the last table you posted. It is not exact due to the volume range buckets, as opposed to the precise/actual price-bar volume and the corresponding volatility. For usage, you can see for example, volume of 7437 has (based on the sample used) produced likely price volatility of 2.5 points. Just as volume of 17384 is likely to produce 6.75 price range volatility. Again, this reverse engineer is NOT precise due to the volume ranges used for the buckets in the Mak-style basis. HTH
If you can use that data for trading, than use it! Good for you! In my experience, the PRV doesn't reliably predict the volume bar's length (it only helps in noticing the jump in momentum), and the volume bar's length doesn't predict the corresponding price bar's length (especially at turns, when it maters most). "Volume leads price" refers to volume events indicating what the price is likely to do on the next bars, both as direction and as momentum. So, a volume bar contains usable information regarding future price bars. It is all part of of the P, V, A/D sequencing (scoring). Obviously, this is how I see it. You're welcome to disagree. I thought that my input might help others in not wasting their time with irrelevant data. We all read Jack's words and understood them differently, this applying to Spydertrader, Makosgu, and other brighter pupils of the method too. There is a saying "seeing is believing". Unfortunately, more often, it works the other way around: "believing is seeing" (or, in other words: you see what you believe to be).
Something else worth mentioning regarding the Vol vs Vola table... Since volume buckets give the "pace bands", re-calculation can allow one to "see" (and utilize) the more macro-changes in "pace". Along these lines, it can be useful to use an external trigger for re-calculation as opposed to time. For instance, triggering a re-calculation based on a daily % change in the VIX might yield very timely changes versus a recalculation after say, X number of days or some other time-based criteria. In all cases, use what works for you. Know that you know... Be, do, have.
That makes sense. Recent volatility changed drastically up and down and using days with similar vix levels to calculate vol vs vola would make this much easier to compare. Why did I not think of that? Thanks!