Hi, Nenad (and others interested on this subject);
In the HP42S you need to explicitly create a N×2 matrix, N equals the number of pairs of data available in the sample, column #1 for X-data, column #2 for Y-data. You should store it in a variable so it can be edited later, if needed. The name of the variable does not need to be an specific, reserved name, but I suggest not using REGS for data safety.
This matrix must be recalled to the X-register prior to execution of [Ê+], being Ê the Sigma greek symbol; all matrix contents will be read, line by line, and statistical data - Êx, Êx², Êy, Êy², etc - is generated automatically. If you need to edit Ê-data matrix and use new data set, you need to clear previous data - [CLÊ] - prior to enter new data.
In the HP28C/S and HP48/49, ÊDAT is already an N×M matrix. In these calculators, your Ê-data base is allowed to have more than two columns (Ê-variables) and you choose which variables you want to wrok with as for related statistics with two variables (curve fitting, forecasting, tests, etc.). Also, editing ÊDAT in these calculators means edit data base directly, what means you should not clear statistical data because this means cleareing ÊDAT matrix.
A lot more exists about this, many other resources. I hope this helps as brief notes. Do you use any of these calculators? If you have an HP41C/CV with X-Functions or an HP41CX, I have (somewhere in here...) a program to deal with statistics data stored in a matrix in the X-memory. If full space is available, standard X-registers may hold about 60 pairs of statistic sample data. It is a bit slow to read data and generate statistical data - Êx, Êx², etc.- but allows handling of lots of pairs without "fear"... If you need the program, tell me. I just have to find it...
Best regards, my friend.
Luiz C. Vieira - Brazil