Estimate of smooth integers doesn't fit the actual graph

44 Views Asked by At

I was studying smooth integers, that is integers that have only small prime factors. Denoting by $P^+(n)$ the largest prime factor of $n$, we define $\Psi(x,B):=|\{ n \le x:\,P^+(n) \le B\}|$ as being the function that counts the amount of integers smaller than $x$ such that every prime factor is $\le B$. Wikipedia has the following estimate for $\Psi(x,B)$: $$\Psi(x,B) \sim \frac{1}{\pi(B)!} \prod_{p\le B} \frac{\log x}{\log p}$$ where $\pi(B)$ denotes the number of primes less than or equal to $B$. I tried to plot an example of it, first plotting $\Psi(x, 5)$ for all positive integers $x < 100$, and then plotting the estimate 1/(3!) * (log x)^3 / (log 2 * log 3 * log 5) on the same interval, but the graphs didn't match at all:enter image description here

So I guess I have made a mistake somewhere, but I have no idea where. Why does the estimate deviate so much from the actual plot?

1

There are 1 best solutions below

2
On BEST ANSWER

The ratio between the two gets better for much larger numbers. Here is the ratio between $\Psi(x,5)$ and the approximation as $x$ goes from 10 to about half a million enter image description here