| № | Пользователь | Рейтинг |
|---|---|---|
| 1 | tourist | 3803 |
| 2 | jiangly | 3707 |
| 3 | Benq | 3627 |
| 4 | ecnerwala | 3584 |
| 5 | orzdevinwang | 3573 |
| 6 | Geothermal | 3569 |
| 6 | cnnfls_csy | 3569 |
| 8 | Radewoosh | 3542 |
| 9 | jqdai0815 | 3532 |
| 10 | gyh20 | 3447 |
| Страны | Города | Организации | Всё → |
| № | Пользователь | Вклад |
|---|---|---|
| 1 | awoo | 162 |
| 2 | maomao90 | 160 |
| 3 | adamant | 157 |
| 4 | maroonrk | 154 |
| 5 | -is-this-fft- | 150 |
| 6 | Petr | 148 |
| 6 | SecondThread | 148 |
| 8 | atcoder_official | 147 |
| 9 | TheScrasse | 145 |
| 9 | nor | 145 |
in problem F the binary search approach why is this equation wrong with large numbers but it works with small numbers
number of time we use current attack = ceil(mid/(c[i]+1))
Update : it's because of the cool down you don't need to wait c[i] round to hit again you actually wait c[i]-1 round so ceil(mid/c[i]) is correct
its cause there is a potential overflow for the attack variable. consider h=2e5,n=2e5,all attacks 1 and cooldown 2e5. Using a break should solve the problem. PS i got hacked too
in test case 6 just dividing 19999800001/200001 is giving me 99998.5000125 now is this results wrong ?
the answer should always be an integer...?
i think so
I used __int128 and break. too lazy to calculate overflow
ok I figured it now it's because of the cool down you don't need to wait c[i] round to hit again you actually wait c[i]-1 round
so ceil(mid/c[i]) is correct
For $$$i$$$-th attack you can hit in rounds $$$1$$$, $$$c[i]+1$$$, $$$2c[i]+1$$$, $$$3c[i]+1$$$, ..., $$$kc[i]+1$$$, so if you are currently at round $$$T$$$, you can hit max of $$$\Bigl\lfloor \frac{T-1}{c[i]}\Bigr\rfloor+1 $$$ times, which is $$$ceil(T/c[i])$$$
Auto comment: topic has been updated by SuperMo (previous revision, new revision, compare).