OK, we're all discussing this without the log. Let's see what the log has to say!
http://dom.retrobox.eu/?/20130611/log.5113bc4be4b06719e45f163f.1370945950365.txtTurn 16: All players buy a Province, putting the split at 2/2/2.
Turn 17:
P1 plays Possession and buys a Province to give 3/2/2. He draws 5 cards, and sees that he can't make a Province buy his next turn from his hand. He can, however, upgrade Potions into Duchies if P3 uses Governors.
P1 Possesses P2, and sees he can only get a Gold. He plays the Possession, giving P2 control of P3's next hand.
P2 plays his regular turn (is that the right order?), and buys a Province, making the split 3/3/2.
P2 Possesses P3 (ashersky), uses Governors to get two Provinces and an Estate. He also gets Potion->Duchy from his own hand, and P1 gets 2x Potion -> Duchy. The split is now 3(+2D)/5(+1D)/2.
P3 (ashersky) is able to get a Province, then quits (form 3rd place).
OK, now let's rewind. Pretend P1 had
not played P2's Possession. The sequence is:
P1 plays Possession and buys a Province. The split is 3/2/2, with no chance of P1 getting a Province next round.
P1 Possesses P2, doesn't play the Possession, and buys a Gold.
P2 plays his regular hand, buying a Province. The split is now 3/3/2.
P3 plays his hand, and uses Governors. He gains two Provinces. P1 gains 2 Duchies, and P2 gains 1 Duchy. The split is now 3(+2D)/3(+1D)/4.
----
One last data point: Ashersky, when making this thread, did not have access to the knowledge that P1 wouldn't have been able to get a Province on the next round; nor did P1 know what P2 would be able to do when he played P2's Possession.
----
In case 1, P1 ends up behind P2 by quite a lot. In fact, P2's lead is almost certainly insurmountable from P1's perspective, as he knows he won't get a Province next turn. In case 2, P1 ends up
ahead of P2.
In other words, if P1 was assuming P3's deck was much better than his own, P1's play of P2's Possession was deliberate kingmaking.