I tried to write a program to validate the false positive rate P$($h$)$ of the bloom filter shown in the textbook by investigating P$($h$)$ with various h$\text{'}$s$.$ I generated $2,000$ different keys from $0$ to $10,000$ randomly, and then divided them into two sets evenly. One was used to map to the bloom filter, and the other was used to query. The false positive rate was got by dividing the number of false positives by $1,000.$ The hash function is h$*$h$*$k$,$ h is the number of hash function, and k is key. The false positive rates obtained by different numbers of hash functions were plotted, but only an increasing curve is obtained, unlike the concave function in the textbook. Where did I go wrong?