Greetings Fellow Audiosurfers,
For some time now, I've thought it would be a fine quest to explore the intricacies of Pointman strategy by analyzing cold, hard, data. Sure -- intuition has been a fine teacher for many -- but mere intuition won't do to better Humankind. It has at last become possible to do
Science, but I will need YOU to help!
A little background:My good friend and fellow AS enthusiast Uupis, after incubating the idea in his mind for
years decades, has forged a slick program that sniffs packets and snaps up the "extended stats" data from each run, logging the data in a handy comma-delimited text file. I don't pretend to know much about the inner workings of this app, but for our purposes we can just trust the magic to work. Those interested in the programming craft might be interested in contacting the author (he does have source code available).
The Project:The project is to use the science of statistics to measure the impact of various simple and derived variables of Pointman Elite play on the outcome of score. More precisely, I will be testing the effects of such variables as the length of track, traffic, points per block of each color (derived), average size match for each color (derived), the extent of priming (derived), the natural separation of the track (derived). Rather than use "skill rating", I take the raw score divided by the Gold score -- this way all runs are useful for data, regardless of whether bonuses like Clean Finish or Match 21 were obtained.
What's in it for Participants:Games! Prizes! Fun!
No really, you get to play Audiosurf like normal (well, if you play Pointman Elite, even casually), and then send me a text file after a few weeks. If interested, I will send all those who participate the aggregated data and analysis in a spreadsheet so that you can play around with it and continue to add your own data. I myself will be using this to assess how I can refine my technique further. I'd gladly make some comments or observations for anyone else who's interested in that, too.
Attached are Uupis' program and a flatted (no formulae, etc.) sample of some preliminary analysis. To use the app, extract the files to the folder of your choice, run the app and AS, and data for your runs will be logged to the file "AudiosurfData.txt". There is also console output with brief run stats shown for amusement. There are some infrequent crashes, supposedly, but I haven't had a single issue so far.
You will need the .NET 4.0 framework installed for this app to work properly. You will also need this: http://www.winpcap.org/install/default.htm Contact me here for questions on the project. Regarding the data collecting app, Uupis has indicated future support will be limited or non-existent, but it might be amusing to pester him with feature requests anyway.
