Epson 4800: R=58, G=42 and B=149
Canon iPF5000: R=0, G=44 and B=177"[/b]
I'm not sure that's going to provide much real info considering the process to get the values and we're dealing with a device independent color space. The profile used alone could account for the blue shifts on the Epson (or Canon). Blues shifting to magenta or cyan is VERY common with many profiles! It's usually one of several good test for evaluating an output profile.
What you could do is this. Create a pure blue patch (R0/G0/B255 in a defined RGB working space) and output using both printers/profiles. Measure them with a Spectrophotometer to get LAB values. You could convert back to RGB using an Absolute Colorimetric intent to the original RGB working space. This gives you the values from source to printer back again (round trip values) however, once again, a profile could (does) severely tweak the values. It would tells us more about the profile than the potential of the printer I suspect.
Even if you turned off all color management, the "raw" behavior doesn't necessarily tell us about the potential of the printer. I guess what I'd do is use the same target and profile package and build a custom profile for both printers, then send images through them to examine the color as well as taking a peek at the 3d gamut maps. The various driver settings would play a huge role as well. I don't have a clue about the Canon but do know that for example, with NO Color Adjustments in the Epson driver, you get the greatest gamut potential at the expense of a non linear behavior. Using Color Controls might very well provide a better blue at the expense of other colors.
In the end, you could spend a heck of a lot of media and find that printer A produces better blues at the expense of Reds while the other printer does the opposite (which is what I've been told about these two devices). Then what do you do???