It is a reasonable assumption that when the dealer goes further into the shoe, the data that comes as a result of the way we keep track of the cards could mean more. If you are on the 50th hand, for instance, you're obviously going to get more of the true picture than if you are on the 21st hand. The percentages you come up with will carry a lot more weight.
Blackjack card counters are very familiar with this principle, which is connected with the idea of "deck penetration." So for a blackjack player having a "running count" of +2 when one deck has been dealt from a shoe is is quite obviously not as meaningful as a running count of +2 when five decks have been dealt. This should be able to work for you in baccarat as well.
Baccarat is played with either six or eight decks of cards, and as card counters do if they are trying to keep two different numbers - a running count and a true count (which is the running count divided by the number of decks remaining) - you are always in a situation where you have to keep track of how many decks out of a shoe have been played, which in the true count formula is a figure that is subtracted from the total number of decks in a shoe.
Example: it's an eight-deck game. If you can count three decks that have been played in the discard rack, you know it's 8-3 = 5 decks remaining.
Okay, let's say that you want to explore some of this. Well, it might help you to now that mastering this kind of thing is usually just a matter of practice and repetition. What you want to do is go out and get eight decks of cards. Create a makeshift "discard rack," perhaps by cutting an old shoe box in half. Make marks on that shoe box which constitute where every deck or half-deck would be. Then just grab random clumps of cards and start to do your estimations. You'll find yourself getting better in no time at all.
You may be a little concerned about the six decks as opposed to eight decks.
One interesting thing to keep in mind is how little the percentages change when you go from eight decks to six decks in a baccarat game. As we know, in an eight-deck game the Banker hand will win 45.86% of the time, while the Player bet will win 44.62% of the time.
If you reduced to six decks, the Banker will actually win 45.87% of the time, with the Player winning 44.63%. The difference is only one one-hundredth of a percentage point, which is negligible. If you went to four decks, the Player would win 44.63% while the Banker would win 45.88%. Once again, there is virtually no difference at all.
So what in the world could we possibly do when armed with this information? Well, we may have an idea along those lines.