I took a random selection of forumites that I found on posts, ranging from heavy users to people who hadn't posted much (just to let you know, you were the third one I stumbled over, and had the highest rep total I found!). Other profiles included well known names such as The Tramp (only one I could find with over 8,000 posts), Lynn, Tasmanian Devil, DavidJames etc. Then I checked the members list to find lighter users who'd been members for a while to find the weighting for length of time as members.
Interesting profiles included Gus with 3,668 posts and three rep power points (the extra two came from the four years on the forum), Fletch with 2,486 posts and rep total 1,036 and only one rep power point, hers was difficult to fit into the model when compared to David Franklin who had less posts, only slightly more rep total but three more rep power points (two from years on the forum and one from 1,250 rep total).
Then I set up a spread sheet and entered all the data for 54 forumites, set up a calculation and compared the predicted results against the actual ones. To determine the calculation firstly I concentrated on the distinction between one rep power point and two and worked out the length of membership rating, once I'd found that the rest of it started falling into place. Two or three iterations later I had a model that fitted every profile.
That's what you get for many years of statistical analysis and systems experience (makes a pleasant change from trying to work out business flow models ). I feel like a real geek after my long winded explaination!
Last edited by Blueshoes; 4th-October-2006 at 09:04 AM.
Franck.
There's an A.P.P. for that!
Okie. Half each to Blueshoes and fletch then!
Hey, I've not been on the forum for long and am still excited about my reputation. It's different when you're old and experienced like TheTramp here.
There are currently 1 users browsing this thread. (0 members and 1 guests)
Bookmarks