Cosmic Ray Detector - Data Center
About fits and chi-squared's
We see that the H=12000 curve goes through the data points, more or less.
We can also give a quantitative measure of how well the curve fits the data. One measure is called the chi-squared of the fit. To compute the chi-squared, you evaluate the following formula:
where yi is the i'th measured value, y(curve) is the value of the curve at that point, and si is the uncertainty of the i'th measurement.
If you divide this number by the "number of degrees of freedom", in this case the number of data points minus the number of free parameters, then for a randomly fluctuating data set, you should get something close to 1 for the chi-squared. If it is smaller than one, the curve fits "better than one might have expected", and if the value is larger than 1 then the curve fits worse.