Mathematician/Statistician

=================================== introduction =================================== We are searching for a Mathematician/Statistician for a trading system we are currently developing. The alpha model is already build (the hearth of the system that generates signals). =================================== facts/issue's =================================== our trading model trades only 1% of the time on the most "extreme" signals. the are 2 main issue's: 1) there are relatively few signals to use for training in that range (1%) 2) More data/signals from the past means more risk of inaccuracy We want this issue solved by try to normalize the signals so that we can use more "old data". also we want to use the current data as some sort of parameter to the normalized data. the apply the current situation to our prediction. =================================== work =================================== The following work needs to be done -- the signals are partly normally distributed, this must be fit to a better distribution(cut-off normal distribution or some sort of Gamma model) -- we need to get some sort of confidence range on the above model. -- if the signals are fit to a distribution we need to use some sort of Maximum like hood to help predicting the next signals. -- we need a function that takes as input: costs,risk,profit and frequency and outputs the ideal risks reward ratio (some sort of kelly formula) -- we need a function to measure the performance of the signals and possible correct it. -- and a lot more ! The above is just a guideline, if there are better/easier models or ways to do it than that is also possible. You are not required on our location, we can use email to communicate. data can be send.