NO source is completely reliable.
http://jeremyreimer.com/uploads/notes-on-sources.txt
He does provide some information on his sources.
When we talk about sales, are we talking about UNITS, or about dollars?
(an important distinction for such as the ZX80 V anything else!)
His curve fitting is flawed. For example, in his 1975 to 1981 graph, he
has points for sales of TRS-80 of ZERO in 1976, and LOTS in 1977. But,
then he just drew a straight line, which erroneously implies a linear
increase, with sales in 1976. THAT derivative should be a vertical line,
not a slope, unless/until you add more resolution for months, instead of
years, representing an extremely steep slope in the end of 1977.
At least his last graph leaves off the "USA Today" fill in graphics, and
shows the individual data points. Which, I'm pleased to say, show C64
outselling ANYTHING ELSE for a couple of years.
But, the vertical scale?? That's 100% of WHAT? If it were percentage of
market share, then it should TOTAL to 100% each year, albeit with a
significant "Other" category. ("Other" is approximately "PC
Share"
inverted from the 50% line?)
Does anybody know of a BETTER source for sales numbers?
Again, "everybody/MOST uses/used" doesn't work.
But, "everybody/MOST who knew what they were doing" does.