i wonder why it has an NTC in the handle, it's too slow and inacurate for any kind of offsetting.
maybe it's a safety to detect the handle heating up.
It is exactly there to detect the temp of the back end of the handpiece, because this is directly related to the temperature at the back end of the cartridge. This is where the cold junction is. The "power leads" to the heater, inside that long T12 cartridge between the contacts and the tip, are made of two difference metals. This is the thermocouple; the power leads are the thermocouple. And this is the reason T12 tips are as long as they are... to better thermally isolate the cold junction from the tip.
If heat from the tip raises the temp of the back end of the cartridge, the voltage this thermocouple produces will drop. And the set temp will effectively increase itself. So yes, it is compensating for temperature change at the cold junction, and it doesn't need to be super fast to do that.
A station that puts a thermistor in the base station is just attempting to compensate for the range of ambient temperature in typical soldering environments. I suppose it would be useful if you move the station around between vastly different temperature work environments. But this is not particularly necessary for most people.
A CJC thermistor in the back of the handpiece would be more important for iron with shorter cartridges and/or higher wattage output. T12 seems to work fine in most scenarios, without any CJC. It is where you are pushing a lot of watts into a heavy ground plane where you might run into some temp drift. I have so far never needed to push a T12 iron that hard to significantly heat up the back of the handpiece/cartridge. And as long as it was self-limiting, and not exponential, it wouldn't necessarily hurt. If you find yourself pumping a lot of watts, it's probably ok or even nice if the set temp actually invisibly raised itself... as long as it would come back down on its own when you put the iron back in the stand, and not go into thermal runaway.