All Open Requests
All Closed Requests
Best out-of-sample test possible at P123
Whether for R2G or private ports, this is the best out-of-sample testing possible at P123--and it is doable.
I propose making it possible to set a port on automatic and updating the price at the end of the day-as is already done with R2G ports. The addition would be to do this with 5 weekly ports run for each day of the week. These ports would then be compiled into a single book. This book of the 5 ports would be the best predictor of future results of this port over a long time period: whether running one port or 5 ports.
If you try this manually you will be absolutely amazed at the range and variation you find in the 5 ports.
The book of the 5 ports gives an assesment of the value of the port much, much quicker than is possible by testing done just one port on Monday. Unless you think there is a day-of-the-week effect, all the ports should be statistically equal. Of course, if you think there is a day-of-the-week effect you should be screaming/praying for this feature--this is the only way you will find it.
This is something I would consider paying more for as far as membership level if this is resource intensive. However, I would recommend it for all R2G ports--it will truly help in the early evaluation of any port--based on my experience.
I admit out front that this is not necessary for those ports based purely on Compustat Snapshot data where the sim will match the port over the same time period. For those sims (possible ports) we have 16 years of data which is enough.
If anyone can think of a better prospective out-of-sample test please let me know.
, # of Votes:
You will be able to view details once you have voted.