Tom, I think you will have to figure those things out without me. My analytical skills are seriously diminished most of the time. I only discovered IB's SLB historical graph within the last week or three. Before that, I could only find current data and I randomly checked it and put it in a spreadsheet. My spreadsheet has many gaps and inconsistent periods.
One thing in my spreadsheet that it does not show on the graph is a statement from the IB current-SLB window that states "There are at least X shares available to sell short." I recorded many of those in my spreadsheet when I recorded the fee rate, date, time, and price per share. I used to pay attention mostly to that, but it turns out it is the fee rate that reflects supply and demand far better than the minimum number available. The only correlation I have seen between fee rate and shares available is that when the fee rate goes above 100%, the shares available goes to zero and stays there. The shares available have gone to zero at lower fee rates, though, but didn't stay there long. Many things can affect how many shares are available, and when, based around which accounts are buying and selling, and whether their holdings are available to be lent. So shares available going to zero during lower fee rates may not be due to shorting, but due to shares that were available being traded to an account that does not lend, or into limbo until the official transaction settlement occurs, etc.
I think the only thing I think I have figured out on my own is an apparent correlation to fee rates rising rapidly to well over 100% and the shorting of every share available. That has happened two times, and I strongly suspect both of them were criminal Short & Distort attacks.
The lowest SLB rate I have in my spreadsheet is 37.04% from 23Sep2016. Right now it is at 38.17%. It varied significantly without apparently correlation to the price except when it went well over 100%, and it may be important that it went from relatively low to very high very quickly. The highest I have in my spreadsheet was 175% on 11Nov2015. I am only intending to cite 100% as an area, not a hard threshold.
I wish I could find, download, and share the raw data behind their graph, but if it is available, I have not found it.
I don't think the SLB rate can predict when normal shorting might occur, which should be whenever a consensus forms that a peak has been reached.
I think it's best significance may be when the rate skyrockets prior to a scheduled news event, before anyone knows whether the news is good or bad. That indicates maximum shorting when the forthcoming news might be great, so the only reason I can think of for deep pockets to do that is if they know a negative media campaign is going to be launched after the news comes out, even if the news is good. That, coupled with fast-algo downward market manipulation, seems to be a classic Short & Distort.
AI
