Oscilloscopes have unique uncertainties. Based on the description, I presume you are discussing a digitizer which has some uncertainties the same and some different than analog scopes.
Digitizer: First and most obvious is the trace width. That will affect the measurement just a little, even if you use the cursors. There is also parallax, which in truth is included with the trace width in a digitizer. Trace width is a rectangular distribution.
Calibration of the cursors is a huge help, but you have to take the uncertainty there as a normal distribution.
Applicable to both digital and analog:
Minor division is not a huge influence, but being between minor divisions is as that cannot really be accurately eyeballed. So take the minor division as the value with rectangular distribution, if you experience that.
Also, different input impedance may have an effect. For example, 50 Ohm, 600 Ohm, 1 MOhm.
If you are near the upper end on your bandwidth (BW) then there could be additional losses. For example, if you use a 50 Ohm coax above 400 MHz.
Watch rise time. The formula is .35 over T rise equals BW, or .35 over BW equals T rise. However, the ratio is not always perfect. Check both BW and T rise, then calculate to make sure you get the same results.
Analog: Hard to get any real reading under 2% due to trace, stability, bandwidth, and so forth, so just use an uncertainty value of 0.5% of reading and you should be OK.
Hope this helps.