Thought I'd add this here: I've done some VERY BASIC calculations based on the number of players available on the transfer market on 13/01/17 00:00-01:00 GMT, just for players 86 or higher. I have the spreadsheet below:
Green cells represent values obtained from the browser webapp, xbox platform, and number of players as available on futhead.com
Orange cells represent predicted values from simple linear regression based on the values obtained in 86+
Purple cells represent values calculated from the predicted values
"avg": The average number of cards available on the transfer market per player in the given rating bracket (i.e. for 81, you would expect on average 414x 81-rated players available on the transfer market)
"sd": standard deviation
"#ofdiffcard": The number of different player cards for that given player rating (i.e. there are 13x 86-rated players; estimated ~304 77-rated players
"total": The total number of cards for that given rating (i.e. there were 3760 87-rated cards in the market at the time of measurement; predict roughly total of 164000 77-rated cards etc)
"cumulative total": The total number of cards in the market equal to or higher than the given rating.
The proportion of cards available on the market SOMEWHAT represent the proportion of pulled cards from packs. (more on this later)
The number of cards includes all special item cards (TOTY, TOTW etc)
The ratings of TOTY and TOTW are ignored (i.e. TOTY Ronaldo still counts as 94 - basically due to how I searched the numbers - by typing Ronaldo on the webapp)
Yes, I am aware that I've totally ignored retention rates, effects of SBC and quick sell rates but I think the numbers are large enough to state that the rate of transfer listing somewhat reflects the rate of card appearances in packs, especially when averaged across all the players within the same rating grouping (e.g. there were 400 ozils and 66 agueros at the time of counting - these contributed to the average for 89-rated players).
So anyway, if we ignore all other cards (silver, bronze, club items, consumables etc), and the rare and non-rareness of the cards, then based on the spreadsheet:
(1): If there was anything like a 86+ rated single player card pack, there is ~0.3% to pack a Ronaldo, 0.5% to pack a Messi
(2): If there was a pack with a single gold-rated player, there is a 0.006% chance to pack Ronaldo, 0.011% chance to pack Messi, 0.7% to pack someone 88+, and 2% chance to pack someone above 86
(3) Assuming that half the players are rare (a generous assumption I might add), the chance of packing in a 100k pack for: either Ronaldo or Messi is 0.83%; For a player 88 or above is 29.8%; and for a player 86 or above is 65.4%
The above numbers seem about right based on experience. Let me know if there are serious, unacknowledged flaws in these set of calculations. I know there are limitations by just using 86+ rated cards to predict the other rating groups but with the limited info I have this is the best I can do.