In, ~50% of cases our reports are run from a wrapper (a report generator utility used at application level which presents users with a list of reports for a specific applications, a list of 'filter' conditions applicable to any of the reports, preview/print options, can run report specific code before and/or after the report, etc.). For those cases it would be easier to implement the solution you suggest - except... we still have to modify every report. Plus, the same reports can be run directly, with a REPORT FORM statement somewhere in a method or in a program, so to implement the 'wrapper' we would also need to change and recompile all those apps.
Documenting the code in the wrapper is a plus, but then it should be an absolute rule that all reports are run trough the wrapper, otherwise it may still be a little puzzling what the myVar represents in the report.
Not as easy as I wished...

In any case, thank you for your contribution - we'll make use of it I'm sure, one way or another.

Now, some thoughts about _pagetotal and performance...
If the only way to find out the correct value is to run the report twice, why is it not done automatically, in the background, and have _pagetotal always available?
Performance reasons!
The same answer I hear about caching data file transactions in various memory places between the application running on a LAN WS, and the file on the server. Today, some store transactions done and visible on a WS in stores could not be seen on other WS using the same app. After at least 15' they miraculosly showed up... I checked the HDD write cache setting and it was, of course on, and it even explains, the reason: to improve performance!! I would call this unsafe driving, and having write cache enabled by default, to me is like having the brakes of the car disabled by default.
Coming back to _pagetotal, most of our reports I'd gues are under 20 pages, and running them twice woul be unnoticeable. For reports that are let's say 500 pages, the performance hit for running them twice is still insignificant compared with the time it takes to print them or to browse them using preview.
So, I don't see any benefit in the 'option' we have with _pagetotal.