Problems with fit in a sub range

[Emanuele94]

Hello everyone,
I tried to perform a fit in a sub interval and to repeat the same procedure by defining a histogram of the sub interval.
The number of bins and the width remain the same.
However, while the coefficients of the central curve are in agreement within the sigma, those of the tails (coefficients of the Chebyshev polynomial) are in disagreement.
What can this be due to?
I show you what I mean in the picture:07.pdf (26.6 KB) example.pdf (34.1 KB)

P.s: For the fit in the sub range I simply used:

 x.setRange("title", x1_min, x1_max);
 model.fitTo(dh, Range("title"));

[moneta]

Hi,

The definition of the coefficents for the Chebychev polynomial is different if you change the range, since the polynomial is rescale and translated to be defined in the [-1,1] range. So, you cannot compare the obtained values.

Lorenzo