I was looking at sieves today, mainly sieving for primes and I noticed a pattern type thing. As I crossed out primes in a small set, the number of composites that were crossed out decreased. I haven't tested this for larger sets but I reckon it won't be as quick. So I had the following question:
Given a set of numbers, at what rate do the number of composites decrease after each prime is cancelled out?
So for example, I have the following sieve
$$\begin{array}{ccccccccc} 2 & 3 & 4 & 5 & 6 & 7 & 8 & 9 &10 \\ 11 & 12 & 13 & 14 & 15& 16& 17 & 18 & 19 \\ 20 & 21 & 22 & 23 & 24 & 25 &26 & 27& 28 \\ \end{array}$$
Now when I begin to sieve out the primes, by cancelling out multiples of $2$ $$\begin{array}{ccccccccc} \require{cancel} 2 & 3 & \cancel{4} & 5 & \cancel{6} & 7 & \cancel{8} & 9 & \cancel{10} \\ 11 & \cancel{12} & 13 & \cancel{14} & 15& \cancel{16}& 17 & \cancel{18} & 19 \\ \cancel{20} & 21 & \cancel{22} & 23 & \cancel{24} & 25 &\cancel{26} & 27& \cancel{28} \\ \end{array}$$
So you can see that $46$% of the numbers are already gone (of course) and then we continue with $3$. The number of multiples that are cancelled out are only $4$ and then when we continue to $5$, it is only $1$. So you can see that the number of numbers that are crossed out decreases rapidly after each prime, which makes sense but is there a general rate at which this happens? Like some number or some formula that one can apply to find that rate given a set?
Thanks!
The first prime removes $\frac{1}{2}$ of the composite numbers.
The second prime removes $\frac{1}{3}$ of the remaining composite numbers. In other words, it removes $\frac{1}{2}$*$\frac{1}{3}$ of all the composite numbers.
The third prime removes $\frac{1}{5}$ of the remaining composite numbers. In other words, it removes $\frac{1}{2} \frac{2}{3} \frac{1}{5}$ of the composite numbers.