Is a simple logit/probit regression not sufficient?

<t>QuoteOriginally posted by: Harry12345Eric,My reasons for not doing a MSc in Finance or Fin Maths are as follows:1) cost - the science masters are cheaper - 5k vs 20kThis is pocket change compared to the NPV of lifetime earnings.Quote2)a) long term plans - i do not think i want to stay in finance ...

<t>QuoteOriginally posted by: KackToodlesQuoteOriginally posted by: mrmisterNobody. Glassdoor relies on the law of averages to get a point estimate on popular job titles. Assuming that everyone is a lying bastard who misquotes his salary, one would be advised to discount the number that he/she sees ...

<t>QuoteOriginally posted by: Finance1987Yah, I don't have as much work experience as you have though..QuoteOriginally posted by: twofish If you totally mess up in a company, then it's not the end of the world. If you totally mess up in academia, it usually is. But, there are many universities, righ...

Are they independent? If so then the cov would seem to be zero..?If they aren't independent, it would seem you need more information to answer the question?

<t>The notation is confusing. Do you mean "d is the number of defective items in the sample"? Are you testing the hypotheses d <= D and d > D? I assume the uniformity is being taken over D? (and thus D is a fixed but unknown parameter) If so, then since d <= D must always be true, the most powerful ...

<r>QuoteOriginally posted by: WinthorpeIn order to make the default probs uniform I was under the impression I could just doF(x) = (x-a)/(b-a) where a<x<bAh. F is the distribution of the data X, not the distribution you want to end up with (uniform in this case). See the probability integral transfo...

<t>QuoteOriginally posted by: WinthorpeHiSimple question I think but just want to check I am not missing something.If I have a time series of probabilites (x) and want to transform them into uniform via marginal distribution (ie. U= F (x) ) then it is simply;each data point/100............correct ?T...

<t>To determine whether your points are uniformly distributed, you could look at their histogram, for example. This assumes that each point is drawn from a single, fixed probability distribution (which may or may not be the uniform distribution). You could then estimate the distribution function to ...

<t>Not sure I understand the question... can you be a bit more precise?You have a time series of probabilities? So you have x_1, x_2, ..., x_N where each x_i is between 0 and 1?More importantly, are the x_i iid with distribution F? If so, then U_i = F(x_i) will all be uniform. Not sure what you mean...

<t>It's clear that prices may be different in a market with traders compared to without, but are prices in the former more accurate estimates of "true" future prices? Or is the net effect of traders only to add volatility? It would be difficult to study, but have attempts to do so been made?Certainl...

<t>QuoteOriginally posted by: Traden4Alpha0.999... = 1 holds for reals, but not for floats.Really? I know what you are getting at, but are those two numbers to be interpreted as floats or as reals to be converted into floats then tested for equality? What I mean is, real "0.999..." refers to a numbe...

3) let the sum given in the question be S.S - aS = a + a^2 + a^3 + ... = a / (1-a)So S = a / (1-a)^2.4) Higher vol implies lower gamma, I would think.5) use the Jacobian.6) too many?

<t>QuoteOriginally posted by: finalguyQuoteOriginally posted by: ArthurDent2 and 20.What does this mean? Hedge fund?"A hedge fund manager will typically receive both a management fee and a performance fee (also known as an incentive fee) from the fund. A typical manager may charge fees of "2 and 20"...

QuoteOriginally posted by: AvovAQuoteOriginally posted by: LandscapeSimple example, let X be r.v. w. cont. distrib. on interval I \subset R, let k belong to IE[ X | X = k ] = kP( X = k ) = 0OK, simple example: Say X is a normal r.v.What is E[X | X = 1 or 2] ?Borel-Kolmogorov paradox