What is it? - page 21
You are missing trading opportunities:
- Free trading apps
- Over 8,000 signals for copying
- Economic news for exploring financial markets
Registration
Log in
You agree to website policy and terms of use
If you do not have an account, please register
да, как угодно. Увеличиваешь кол-во бросков увеличивается возможное отклонение. В пределе бесконечность может отклониться бесконечно далеко ;)
Само собой, я не считаю, что при 10 бросках монеты (орел=+1,решка=-1) кумм.сумма уйдет на расстояние больше 10 от начала координат О))))
Uh... That's a relief. ;)
Итак, заново. Создаем новый объект - систему событий (напр. рулетка). Зеро нет. Красное/Черное - 50/50. Сделали 1000 испытаний. Произошло событие A1 (одно событие) при котором Красное выпало 600 раз, Чёрное выпало 400 раз. Соответственно есть крайне малая, но допустимая P(A1) например = 0.0001, т.е. находится в районе третьей сигмы (в нашем случае уже дальше).
Теперь (будь по Вашему) расчитываем вероятности и получаем, что P(A3) ={в следующей серии из 1000 испытаний выпадет не менее 600 на красное} равно P(A4)={в следующей серии из 1000 испытаний выпадет не менее 600 на черное}
I.e. we get equal probabilities that the other theorem works or does not work
II) With a large number of trials n, the number of events A will tend to n*P(A) - Understood and accepted.
This is exactly where the misunderstanding is present. For the described procedure one series of 2000 throws is one trial, but if our argument took place then we would have a large number of trials.
So you are trying to draw conclusions from the results of a single trial that ended with an unlikely outcome.
A craving for knowledge is wonderful. Have you read the book recommended by Mathemat, it doesn't say that? About averaging-integration in Yandex search?
P.S. Understand, unlikely events happen.
Как Вы эту цифру ни назовете, - матожиданием, прогнозом или еще как, - все равно 500+600 будет в центре того, что Вы получите в результате от серии из 2000 испытаний.
"The 'test centre' is looming. There's already a mate's expectation and the average is low. Yeah...
OK, conditional expectation.
read.
Where did you get this from?
II) With a large number of trials n, the number of events A will tend to n*P(A) -- I understand and accept.
There is no such a thing. The number of events A can deviate any far from n*P(A). Look up laws of arcinus. http://polbu.ru/safonov_dealing/ch61_all.html
Yeah. Well, as a variant I took just from your link. I quote:
Actually there is no contradiction here. The law of large numbers is so called because it is only true for an increasing to infinite number of test series. That's when the win rate tends to 1:2.
And this is where you got it <<.... The number of events A can deviate as far as you like from n*P(A)......>> ??? Especially: as far as you like.
....
And, please, let's refer to materials arousing some confidence, at least close to scientific.
For you Mr. Safonov V.S. from site PO-LBU.RU is an authority in TheorWer? All the more so, he is a mess... Again I quote:
Results:
T = 154,126,100,75, 50, 35,20, 9, 2;
P = 0.9, 0.8, 0.7, 0.5, 0.4, 0.3, 0.2, 0.1.
This means, in particular, that with probability 0.9 the luckier player will be in a winning position 211 days a year, i.e. almost 60% of the time. Not bad!
Even the numbers are scattered.
In general, the article looks like chewing gum from the School of Trading in DC. (Or is that the right high school?)
A thirst for knowledge is wonderful. Have you read the book recommended by Mathemat? Have you tried searching for averaging-integration on Yandex?
Look how you read it. DECAN. No less. ))))
Stop referring me to DC Trading Schools, Yandex, etc.
I asked a specific question here. Didn't get it, sorry. I'll restate.... Do you want to be more specific? I will always answer....
If I'm wrong about something, lack of knowledge, and it's been reasonably proven. OK. So I will study and try to understand what is being proposed. It's in my best interest. But you can learn something useful from me too, believe me.
I'm ready for a dialogue, but with adequate people.
So, the question is asked. What is the conclusion?
We need an answer to this simple question.
I would like to ask you to speak in substance. If you have nothing to say, then sit quietly on the sidelines.
I no longer have time for this kind of idle correspondence. By your classification: I'm an ignoramus. And studying takes a lot of time. Don't distract me, I'm asking you nicely.
.....
I would like to hear the opinion of Vinin, KimIV, Prival and many others.
If it turns out that most of them think everything I wrote is nonsense, then I'll take the last word, apologize and leave. I'm not claiming anything here.
Итак, вопрос задан. Какой вывод?
Нужен ответ на этот простой вопрос.
I will try to simulate many Bernoulli series with the above parameters and see what might happen. The script is ready, I just need to remember how to use it. Don't expect a quick answer.
At the same time on purely experimental material and see what fraction of trajectories will end up in the area of your gain.
У меня более нет времени на подобную пустопорожнюю переписку.
A woman with a cart, a mare with a mare
I will try to simulate many Bernoulli series with the specified parameters and see what might happen. The script is ready, I just need to remember how to use it. Don't expect a quick answer.
At the same time, purely on the experimental material and see what fraction of trajectories will end up in the range of your winnings.
Thank you. I want to clarify: Will this experiment be based on example of trading or roulette? i.e. only pose entries will be made by MathRand? Or will the entire CB sequence be generated?
I've been modelling roulette too. )) And it will be very interesting to see your results.