Ok, everyone:

@dr_frost_dk: "the 1320mAh battery can in no way supply over 24W they are not built for high power delivery"

You are correct, I said ideal battery on purpose back there, even mentioned resistance.

The proposition was, booting the battery decreased charge by 10% as shown by the indicator. I said it can't be, because bootup is 30 seconds and if you killed 10% in 30 seconds that would be many-a-watt, so it can't happen. For many reasons, one being battery can't deliver, another being the phone would fry. Nobody is claiming it can happen. Au contraire.

@mece
That took about 10 minutes to understand, but yes, you are correct. Roughly 30A, or ~100W, in 15 seconds, that's 180W for 10 seconds.

@shadowjk
My measurements disagree with your last part, but I have a custom kernel. I'm fairly sure that my phone doesn't charge battery at all if left connected after end-of-charge, because the consumption slope is identical when connected and not connected.

Will dig further.

@Estel:
"A candle that burns twice as bright burns half as long". We can all agree to that. 5Wh is 5Wh and nobody is changing that (we wish).

5Wh = 5W in 1h
5 Wh = 10 W in 30 minutes

You said:
"10% of 5Wh battery is damn 0.5Wh, and using that 10% of capacity during 15 seconds is using damn 0.5W = 500mW in 15 seconds."

You again mix hours and seconds like they are the same.

10% of 5Wh battery is damn 0.5Wh(1), and using that 10% of capacity during 15 seconds is using damn 0.5W(2)(see what you did here?) = 500mW in 15 seconds.

Herein lies your problem. You said that 10% of 5Wh is 0.5W. It's not. It's 0.5 Wh. Since 10% of a kilometer is not 0.1 oranges. You should keep the measurements to the end of the equation.

You used 5Wh (at 1), and said that 10% is 0.5W (2). It's actually 0.5Wh, which is 0.5W FOR AN HOUR. Not 15 seconds. 0.5W for an hour, but delivered in 10/15 seconds.

If you still have issues with this, I suggest you take the simplistic approach to physics. Go with metric basic measurements. They are volt, ampere and second. I'll use the extended Watt as V*A.

A 5Wh battery delivers 5W for one hour, that's 5W for 3600 seconds. The measurements is watt hour, NOT watt per hour, mind you, it's not a division. Watt per hour implies flux of energy, watt hour is capacity.

A faucet delivers water at 1 liter per hour, meaning that left alone for an hour, it produces a liter. That allows for a bucket that gets filled in one hour to be one liter (1l/h*1h simplifies to liter). So, Wh is capacity, and Wh per time (hour) is W.

We simplify to metric, so the equation becomes:

5W * 1h = 5W * 3600 seconds = 18000 Ws

Note that when you do math with measurements, measurements keep the operands. So, two apples times two oranges makes four apple*orange. We write that as apple orange.

When you multiply watts and seconds, the result is watt second, or Ws. A grand total of 18,000 Ws. That's it. 18000 watts second. If you take 18000 seconds (5 hours) that's 1W. Makes sense. Also, if you take one second, it's 18000W. Whole battery.

10% is 1800Ws. If it boots in 10 seconds, it's 180W dissipated power, vast majority in heat. It'll glow like a light bulb (well ... -ish).

My original post read:
"A battery is typically over 1000 mAh at 4.2 volts, that's 4.2 watts. N900 is 5000 mWh, and 10% of that is 500 mWh. That's 500 mW for 3600s, that's 1800Ws, and on a 10 second powerup you'd need to eat 180W (my math may be off it's really late)."

Considering the hour, I was surprisingly accurate.