Of course, the trouble with tackling Wittgenstein is you can bogged down so easily and distracted into various kinds of detective work (sometimes I wonder if half the attraction the sage has for some of his less critical fans is that he offers his readers the pleasure of puzzle-cracking as one tries to track down the sense of the more gnomic utterances). So, I’m still wrestling with Sec. 108 of the Big Typescript, the first section in the last third of the book which comprises remarks on the foundations of mathematics. But here, to be going on with, are my reconstructive efforts organizing and padding out the remarks of Sec. 108 into more continuous prose: commentary to follow. [For an updated link for the version with commentary, see posting for 1st Oct.]
So far it seems to boil down to moves being good or bad and propositions being true or false. (A good move being one that increases your chances of being in the winning state). Another point is that games are competitive (where one player has the best score) even in the case of a single player (competing against yourself or your own expectations) whereas a problem is not (and if there is more than one person involved, it tends to be collaborative).
I have to say, Wittgenstein is pretty irritating even in the connected prose version. So if even the irritation isn’t lost by such a transformation, then whatever is lost, if anything, must be very obscure.