Hi, I'm a second year Master's student in Applied Statistics. I am doing a project using average weekly U.S. regular gasoline prices (in cents, per gallon) from an Excel file (from the years 1990- May 2010). I want to find the probability that the average weekly U.S. regular gasoline prices (in the long term) goes over 400 cents a gallon (or $4.00 a gallon). I am using the extRemes program (in R), and I've already uploaded my data, and found the probability using the fitted GEV distribution (I used the pgev function from the evd package). I want to compare the results from the GEV distribution to those of the GPD distribution. I've already fitted the GPD distribution to my data (at a threshold of 400 cents, where there are 52 entries a year (for most years, at least)) and have obtained the supporting graphs. My question is about how to find the probability that the average weekly U.S. regular gasoline prices (in cents, per gallon) goes over 400 cents a gallon from the Maximum Likelihood Estimates I found when I fit the GPD distribution. MLE's MLE Std. Err. Scale (sigma): 62.63489 1.999964e-06 Shape (xi): -11.59905 3.948669e-04 Negative log-likelihood: -12.2703516139708 The results did not give me a location (mu) or a beta value. I'm wondering which function I can use to find the probability (like when I used the pgev). I've already tried to use the pgpd function from the evd library, but it kept giving me an error message (I don't remember exactly what), and I've tried to use the ppareto and pgenpareto function from the actuar package, but they did not work either (I think because the shape parameter has to be positive, which mine is not). Since the shape paremeter is less than 0, I believe that the GPD is of the beta type. If any of you could help me answer this quandry (I'm a bit of a beginner in R) it would help me a lot. Thank you. ---------------------------------------------------------------- This message was sent using Illinois State University RedbirdMail