A few observations from the data logging I have done in the past.

Due to physical properties noted above, water does cool the intake charge some, but not near as well as methanol due to the boiling point, etc. Water seems to do much of it's cooling in the combustion chamber by absorbing some of the heat of the combustion process, and slowing the burn retarding it over a greater number of crank degrees.

One item that should be noted is methanol combustion actually cools the combustion process as well due to cooler burns characteristics vs gasoline.

What usually drives the direction I see tuners going with either water and methanol concentrations is the limitations of the application. Water likes dynamic compression, either high static compression or lots of boost. You also need a strong ignition to light the flame front when water to fuel ratios start to rise other wise power will be lost and misfires will occur. Larger amounts of methanol are much more forgiving because it burns.

The main reason, in my opinion, methanol is more popular than water is in the way the modern ECU works(and weak factory ignitions). The driving force in timing tables and fuel dumping to cool the chamber, is knock and IAT. With a higher octane vs pump gas and cooler burning temps methanol helps the knock issue. The biggest benefit seems to come from cooling the inlet charge, enabling the ECU to give maximum timing. This is assuming you can blow it past the IAT sensor. The IAT timing compensation can be a very strong and evil thing, and water doesn't affect it as much as methanol.

Now, if running a stand alone you can throw out the ECU issues, and with a strong ignition run straight water and do miraculous things. But, with stock ECU's and crap gas, it's much easier, and sometimes more productive, to run high concentrations of methanol for maximum performance.