Isn't that kind of arbitrary, though? Wouldn't ${\displaystyle \Delta v=I_{sp}\cdot \ln \left({\frac {m_{1}}{m_{2}}}\right)}$ make more sense?