The most accurate decisions come from spreading the data out more. Level 2 counts already spread it out about twice as much by having count tags of 2. But dividing by the number of half decks remaining compresses data rather than speed it out. A level 1 count would spread out data as much as a level 2 count if it divided by the number of double decks expressed to half deck accuracy. This would narrow the bell curve SD for the actual compared to the average in each integer TC bin. So if 6 half decks remained you would divide by 1.5 double decks.
To illustrate the point about integer TC accuracy consider the following with all divisors to half deck accuracy with TC floored:
(RC, number of half decks remaining: TC divided by number of half decks, number of full decks, and number of double decks):
8, 10 half decks remaining: 0, 1, 3
8, 9 half decks remaining: 0, 1, 3
8, 8 half decks remaining: 1, 2, 4
8, 7 half decks remaining: 1, 2, 4
8, 6 half decks remaining: 1, 2, 5
8, 5 half decks remaining: 1, 3, 6
8, 4 half decks remaining: 2, 4, 8
8, 3 half decks remaining: 2, 5, 10
8, 2 half decks remaining: 4, 8, 16
Between 1 deck and 5 decks remaining:
For the RC of 8 dividing by half decks remaining spreads the data across 3 integer TC bins for a range of 4 TC.
For the RC of 8 dividing by full decks remaining expressed to half deck accuracy spreads the same data across 5 integer TC bins for a range of 7 TC.
For the RC of 8 dividing by double decks remaining expressed to half deck accuracy spreads the data across 7 integer TC bins for a range of 13 TC.
Now a sim can make half deck bins to get more bins in the same range and spread out data across more bins that way to get tighter bell curves but then you have to take the math to an additional decimal place. People tend to think in integers. So doing what you can to keep everything you need to calculate or memorize to integers is easier for most people.
Bookmarks