Recent content by maswerte

  1. M

    Combining probabilistic forecasts

    Wrong again. The theory as I described it works and the above formula is -functionally- correct. To better the results of the standard theories the problem seems to be the choice of the form of the q' function. * I 'm sorry for writing some random things in my last post, but it's because I was...
  2. M

    Combining probabilistic forecasts

    No, the difficulty here is to approximate q, for the realistic situations. q' = 0.5 + L . (q - 0.5) is wrong - good only for two predictions. Here is a paper that says something: http://www.math.canterbury.ac.nz/research/ucdms2003n6.pdf
  3. M

    Combining probabilistic forecasts

    The solution of the problem appears to be as I described it for the 2x2 case. One of course has to take measurements to see what departure there is from the ideal case when the two predictions are not 100% independent. But what if we have more than 2 predictions and more than 2 predictors ? If...
  4. M

    What is the Correct Probability in Unorthodox Weather Prediction Methods?

    I believe you start with the ideal formula (independet p-q) and apply maximum likelihood estimators.
  5. M

    What is the Correct Probability in Unorthodox Weather Prediction Methods?

    We established the probability formula for the case of two independent probability forecasts: f(p,q) = p*q / (p*q + (1-p)*(1-q)) In practice the two forecasts are never independent. If p >= q > 0.5 then the "true" q is somewhere in between. That is we have to replace q in the formula...
  6. M

    What is the Correct Probability in Unorthodox Weather Prediction Methods?

    The argument used by mathpages.com to derive the formula is hellishly tortuous. Let me try to do this differently: Each of the two forecasters, weatherman and Indian, follows his own set of methods. We can't tell to what extent those methods are the same. The ultimate case of independence...
  7. M

    What is the Correct Probability in Unorthodox Weather Prediction Methods?

    I have some clues and some lines from you and I 'm trying to put it together. There are some extremes in the problem: The bad extremes: - First predictor makes reasonably good predictions, the other one just flips a coin. Then p2 = 0.5 and the formula from mathpages (the nature of which I...
  8. M

    What is the Correct Probability in Unorthodox Weather Prediction Methods?

    Let's call the statement "rain", statement A and the statement "no rain", statement B. Prior to the event there are four possible predictive statements A-A, A-B, B-A, and B-B, each with its respective probabilities (p1, 1-p1), (p2,1-p2) etc of positive and negative outcome. So if we measure...
  9. M

    What is the Correct Probability in Unorthodox Weather Prediction Methods?

    My answers are not coherent maybe because I have reached my present limits of knowledge about the problem, which after all I sort of made up in the first place without knowing the answer. I 'll think more about it in the light of the day. I did have some examples with figures showing a certain...
  10. M

    What is the Correct Probability in Unorthodox Weather Prediction Methods?

    It adds something, like the 75% that became 90% a few posts above. But some other times the weatherman is unfairly penalized. The way I see it the Indian is useless if he says random things. Above this extreme it should be true that one opinion adds to the other.
  11. M

    What is the Correct Probability in Unorthodox Weather Prediction Methods?

    If you have M guessers and N states of matter they are guessing, it becomes difficult. The system (sun-rain, weatherman-indian) is only 2x2 so it looks easy. Among the N states of matter, or types of weather if you like, some are likely to be pretty infrequent, so how long do we say to our...
  12. M

    What is the Correct Probability in Unorthodox Weather Prediction Methods?

    I see. You break the problem into components and say "correct weather station + correct indian = x%" and so on. And if Indian is infact worthless then it should turn out x = p. Perhaps I was thinking along the lines of probability systems with lots of such components + observers, where the...
  13. M

    What is the Correct Probability in Unorthodox Weather Prediction Methods?

    The scientific method is the first formula ? I don't believe so. It is scientific but under the conditions it specifies. Here is an obvious case of failure: Predictor A is a learned journalist, predictor B is Paul the Octopus (& don't spoil by saying you believed the octopus !). In my...
  14. M

    What is the Correct Probability in Unorthodox Weather Prediction Methods?

    Does it make any sense to do that: f(p,q) = p^a*q^b/( p^a*q^b + (1-p)^a*(1-q) ^b) where a, b are experimental constants ? The sum of f(p,q) over the various outcomes is not 1 but I could normalize f(p,q) to make it equal to 1. It's kind of heuristic. If a=1, b=0 then f(p,q) = p and the...
  15. M

    What is the Correct Probability in Unorthodox Weather Prediction Methods?

    It does n't look outrageous to me if f(p,q) = 0.9, if the two pieces of evidence are uncorrelated. Suppose police receive a UFO call. They say "it's another loonie this time of the night'. But when they receive a second call, they jump to attention. Don't they ? That's because they believe the...
Back
Top