[Olympus] f(Q^2)=const fit to our points

[Alexander Kiselev]

   Hello colleagues,

   thanx to the writers who included the table with data points in the
draft, one can try exotic options to describe our data. Like "no effect
whatsoever" for example. Why not?

   The f(Q^2)=const fit to the main acceptance (a) points (bins 1..19), 
taking into account stat.errors only yields

   chi^2/ndf ~ 15.5/18     -> looks fine

   <R> = 0.9909 +/- 0.0008 -> consistent with the 12-degree measurement
                              of 0.9971 +/- 0.0046 within ~1.5 sigma or so
                              (sys.error taken for this point of course)

   One can do it differently: fit all 20 points at once with f(Q^2)=const 
function (taking stat.errors for the main acceptance points and sys.error 
for the 12-degree point). This yields chi^2/ndf ~ 17.3/19, also fine.

   Looks consistent with the statement "once rad.corrections taken into 
account, no sign of any difference between e--p and e+-p scattering 
cross-sections if MWPC luminosity measurement used". Which is quite an 
interesting observation by itself. And would probably be our final 
conclusion if either (1,3) peak in the right SyMB ADCs spectrum happened 
to be off-scale or Axel would not come up with the MIE analysis idea.

   Since everybody else from JLAB or Novosibirsk can do the same exercise
based on the table II data, would not it make sense to mention it somehow,
at least indirectly?

   Regards,
     Alexander.