Thread: combining probabilities ?

hi,

say i know the probability of < 25.5 successes is 45% for event 1
and the probability of < 3.5 successes is 60% for event 2.

how can i calculate prob of < 29 successes for event 1+2

how would i calculate prob of < 25 successes for event 1+2.

sorry if i'm asking a stupid question. is it even possible with this info?

many thanks

im pretty sure thats impossible with the information provided.

thank u sir... what additional information would u need? the actual observed trials and successes? if u had to make an approximation, however rough that may be, what would be ur best try?

you need an appropriate joint distribution for the events X1 and X2.

if the events are independant, you can take the probability distribution of each individual event and multiply them together.

thanks

in 1 or 2 cases they r not independent. i'll try to get to those, it just takes a lot of time... for data.

thanks again

To avoid any possible misunderstanding multiplying the (independent) distributions together would give you a joint distribution, not the final answer.

yea, i get ya. the problem is the data is a pain to scrape and maintain.... often sample sizes r small and there r a bunch of variables that may affect the outcome. the latter is psuedo-science, in that u can never know all the factors. (what time did they go to bed? did they get laid?)

what i was hoping for was a time-saving-best-approximate. mean is not enough but if i had a median as well could i perhaps approach an approximate? really i give myself about 4% before i even consider a wager, then i'll look deeper into the price. but i need to evaluate 100 of these in an hour or 2 to make it really worth my while. speed > accuracy. i bet the max so i don't have to quantify too accurately my edge, just need to know that i have 1.

so how does this site work? u thank some1 and a donation box comes up.... is this for the site, not the helper, right? i'm leaning towards getting this done properly and paying for the scraping/maths.

any ideas about where i can find good hired stats help? also what would be the going rate for a russian/asian prof as opposed to a western 1? the problem that i have is that i consider the info as fairly valuable, soooo i was hoping to find some1 who may be interested in the information and how to use it. IMO, this is many times more valuable than what i am prepared to pay for the work.

maybe the best thing to do is go to my local university and find a stats student.

thanks for any feedback