29 июл. 2024 г.

mathstral, Newton-Pepys

 Что вероятней:


...     A. Шесть честных кубиков бросаются независимо друг от друга, и выпадает по крайней мере одна цифра «6».
...     B. Двенадцать честных кубиков бросаются независимо и выпадают по крайней мере две «6».
...     C. Восемнадцать честных кубиков бросаются независимо и выпадают по крайней мере три «6».


1. **Calculate the probability for option A:**
   - Event $A$: At least one six appears when rolling six fair dice.
   - The complementary event $\bar{A}$ is that no six appears on any die.
   - Probability of not getting a six on one die:
     \[
     \mathbb{P}(\text{not six}) = \frac{5}{6}
     \]
   - Since the dice rolls are independent, the probability of not getting a six on all six dice is:
     \[
     \mathbb{P}(\bar{A}) = \left(\frac{5}{6}\right)^6
     \]
   - Therefore, the probability of getting at least one six is:
     \[
     \mathbb{P}(A) = 1 - \mathbb{P}(\bar{A}) = 1 - \left(\frac{5}{6}\right)^6
     \]

2. **Calculate the probability for option B:**
   - Event $B$: At least two sixes appear when rolling twelve fair dice.
   - The complementary event $\bar{B}$ is that fewer than two sixes appear (i.e., zero or one six).
   - Probability of not getting a six on one die:
     \[
     \mathbb{P}(\text{not six}) = \frac{5}{6}
     \]
   - The probability of getting exactly one six in twelve rolls can be calculated as follows:
     \[
     \binom{12}{1} \left(\frac{1}{6}\right) \left(\frac{5}{6}\right)^{11}
     \]
   - Thus, the total probability for fewer than two sixes is:
     \[
     \mathbb{P}(\bar{B}) = \left(\frac{5}{6}\right)^{12} + \binom{12}{1} \left(\frac{1}{6}\right)
\left(\frac{5}{6}\right)^{11}
     \]
   - Therefore, the probability of getting at least two sixes is:
     \[
     \mathbb{P}(B) = 1 - \mathbb{P}(\bar{B}) = 1 - \left[\left(\frac{5}{6}\right)^{12} + 12 \cdot \frac{1}{6}
\cdot \left(\frac{5}{6}\right)^{11}\right]
     \]

3. **Calculate the probability for option C:**
   - Event $C$: At least three sixes appear when rolling eighteen fair dice.
   - The complementary event $\bar{C}$ is that fewer than three sixes appear (i.e., zero, one, or two sixes).
   - Probability of not getting a six on one die:
     \[
     \mathbb{P}(\text{not six}) = \frac{5}{6}
     \]
   - The probability of getting exactly two sixes in eighteen rolls can be calculated as follows:
     \[
     \binom{18}{2} \left(\frac{1}{6}\right)^2 \left(\frac{5}{6}\right)^{16}
     \]
   - Thus, the total probability for fewer than three sixes is:
     \[
     \mathbb{P}(\bar{C}) = \left(\frac{5}{6}\right)^{18} + \binom{18}{1} \left(\frac{1}{6}\right)
\left(\frac{5}{6}\right)^{17} + \binom{18}{2} \left(\frac{1}{6}\right)^2 \left(\frac{5}{6}\right)^{16}
     \]
   - Therefore, the probability of getting at least three sixes is:
     \[
     \mathbb{P}(C) = 1 - \mathbb{P}(\bar{C}) = 1 - \left[\left(\frac{5}{6}\right)^{18} + 18 \cdot \frac{1}{6}
\cdot \left(\frac{5}{6}\right)^{17} + \binom{18}{2} \left(\frac{1}{6}\right)^2
\left(\frac{5}{6}\right)^{16}\right]
     \]

4. **Comparison:**
   - The events can be compared by their complementary probabilities because the function $\mathbb{P}(\bar{A})$
decreases as the number of dice increases (since the probability that none of the dice show a specific value
decreases with more rolls).
   - Hence, we have:
     \[
     \mathbb{P}(\bar{A}) > \mathbb{P}(\bar{B}) > \mathbb{P}(\bar{C})
     \]
   - Consequently, the probability that event $C$ (getting at least three sixes with eighteen dice) is the highest
among the given options.

Conclusion:
\[
\boxed{\text{C}}
\]

 ==========

 все верно, кроме ответа. вычисляет плохо, как все. поэтому нужно просить написать код. пишет с плохими отступами, но пишет верно, ну трудно написать неверно, а codestral неправильно пишет формулы.... такие дела - медленней и неправильно, но вычисляет верно.

chat gpt 4o -     решил верно, но ошибся в вычислениях.

mistral2 (?) - не уверен, что такой стоит - не стал вычислять 

(мудро) - а сказал, что кажется мне, что А из общих соображений..

попросил вычислить - тут же привел пример программу на питоне  и все верно вычислил...,

вы хотите песен - их есть у меня.  разумный.


не знаю что такое будет gpt 5 ...   но открытый код мистраля...   делает свое дело.

просто разработчики бедней и умней. llama 3.1(фейсбук) хуже, так же как продукт 

google (ge..), не говорю о 

microsoft (какими были такими и остались)

llama 3.1 (8,70, 403) может в чем-то и лучше (не знаю в чем) - но глупей ггг.

=====================================

модели меняются, сегодня (и месяца не прошло, после рождения) они уже другие, процесс пошел, так сказать. Теорему Пифагора мистраль уже доказал. mathstral еще нет, но ..  стал быстрей и все менее и менее напоминает усердного, тупого школьника, mathstral поумнел, учитывая его небольшой размер и способность работать на лаптопе...процесс побежал ... пока не очень быстро, но побежал.


Комментариев нет: