Among $168$ prime numbers in range $1$ to $10^3$, there are $84$ prime numbers $n$ such that: $p^k> n.rad(p^{k+1}−n)$ where $1 \le n<p$ and $k=2,3,4$. There are also $84$ prime numbers $n$ such that: $p^k> (n.rad(p^{k+1}−n))^{1.01}$ where $1 \le n<p$ and $k=2,3,4$. The ratio $\frac{84}{168}=0.5$ is high. **Question 1:** Are there infinitely many primes $p$, positive integers $k, n$ such that $1 \le n < p$ and $p^k > n.rad(p^{k+1}−n)$? **Question 2:** If the answer to question 1 is yes, is there $\varepsilon > 0$ such that there are infinitely many primes $p$, positive integers $k, n$ such that $1 \le n < p$ and $p^k > (n.rad(p^{k+1}−n))^{1+\varepsilon}$? **Question 3:** If the answer of question 2 is yes, then is the ABC conjecture false? if we let $A=p^{k+1}-n$, $B=n$ $A+B=p^{k+1}-n+n=p^{k+1}=p.p^k>p.(n.rad(p^{k+1}-n))^{1+\varepsilon} \ge (p.n.rad(p^{k+1}-n))^{1+\varepsilon_0}=rad(ABC)^{1+\epsilon_0}$ List of $84$ pair $(p, n)$ as follows $p=A(2i-1)$, $n=A(2i)$ for $i=1,2,...,84$ Examples: $13^2-10.rad(13^3-10)=139>0$ $23^2-17.rad(23^3-17)=19>0$ $37^2-28.rad(37^3-28)=949>0$ $A=$ [13 10 23 17 37 28 73 55 107 43 137 89 181 136 191 89 281 41 313 1 337 253 353 89 379 3 383 287 433 17 433 325 467 5 541 406 563 422 631 31 769 577 811 83 863 647 937 703 3 1 5 1 7 1 17 1 19 1 31 1 41 1 53 1 73 46 97 1 107 1 127 1 127 19 131 121 163 1 181 49 193 1 199 1 239 1 241 1 251 1 257 1 271 1 307 1 331 49 337 1 443 1 449 1 487 1 557 1 577 1 593 1 599 193 751 1 797 49 821 561 881 385 907 817 937 289 983 289 3 1 23 2 29 5 43 15 47 7 103 7 163 82 197 17 211 3 229 24 251 1 277 8 281 249 283 43 449 193 487 1 563 11 821 5 827 11 853 5]