# producing a stock market game based on variable data

I recently launched a stock market game (http://linkdaq.net) based on the top 50,000 sites ordered by the amount of links, it works reasonably well but I'm not 100% happy with the maths behind it.

Let me first be upfront and say I've no experience in the stock market, this was just a little weekend project that has taken a life of it own.

Each morning I get a list of sites, what rank they are (1-50000) the amount of links they have today and the amount of links they had yesterday.

I currently use this as my formula

(((50000 - rank ) + number_of_links) * 0.0010)

This gives me a dollar amount that is reasonable

I also work under the assumption that a site will gain 0.03% of its links per day, if it does more than this, it gets a daily bonus of 0.05% if it does less it loses 0.05% due to lack of growth.

So my question is two fold really,

1) How should I go about improving this, all the current stock market formulas I've found are based on the number of people investing in a stock and I'm not ready to use that model yet. 2) My next task is to add other elements such as twitter shares and facebook likes but I'm a little uneasy about giving them a dollar value the difference is just massive 999736 vs 4 for some urls, do you have any reading suggestions on how I can combat this?

Thanks a lot

-
I do not totally understand your approach, but it seems like it may have hidden flaws. For example, use Google Chrome Collusion and see the number of site links, cache entries and cookies that visiting a single site creates and how much data is shared by visiting just one site. Why wouldn't that data count (if this is even measurable) in the quality of the site as a basis for ranking? – Amzoti Oct 6 '12 at 12:48
In order to know how to tell yo how to improve, it would be nice to knwo how to measure quality, i.e. what to optimize. – Hagen von Eitzen Oct 6 '12 at 13:39