Precision of the Alegro ACS715 30A hall effect sensor

Hi

I was just about to press the button to buy a few ACS712 +/-20A hall effect boards for a multi-channel ammeter project when I discovered the ACS715 0-30A option : http://www.pololu.com/catalog/product/1186

The 30A appeals for the overcurrent protection and also the design appears to offer low resistance and power loss (1.2 m? compared to the other designs I was looking at which don't list a spec which concerns me some). The project is in a marine situation, measuring battery charge currents from multiple sources. None of these ranges above 10A but both these options will allow for headroom in the design.

I would like to validate some maths before switching though.

The 20A unit would be half wasted on this since the polarity on the lines I'm measuring will never be reversed, so I only have values from 2.5V (0A) to 5V (20A) to read, which if my numbers are right means 511 possible values spanning +20A or steps, or about 0.04A precision.

The ACS715 datasheet says at 0A it reads 500mV, which means 0A would read 103 (say) on the analogue pin, leaving 920 values to map to the 30A, or 0.033A precision. Am I right in this, that despite losing that first 500mV, or 1/10th of the available analogue 'bandwidth' the 30A still will deliver slightly better precision overall?

Thanks in advance for checking my maths, Geoff