These forums have been archived and are now read-only.

The new forums are live and can be found at https://forums.eveonline.com/

EVE General Discussion

 
  • Topic is locked indefinitely.
 

Statistical questions that can be asked of EVE
Author
Previous Topic Next Topic
Pixel Piracy
Native Freshfood
Minmatar Republic
#1 - 2015-10-08 02:47:14 UTC
Hi all,

I am currently doing a crash course through statistics and probability. I want to try and apply what I am learning in EVE as I think I would be much more engaged with it. It would also serve as a functional tool to keep my skills in practice.

So my question is this: What sort of statistical questions could I ask of EVE? I am looking to collect data to calculate averages and derive other statistical results from.

I've considered taking note of the damage of my shots (slow arties) and calculating the average damage over the course of a fight(s) or mission. Also thought about calculating average ISK per/h from various sources.

Can anyone else propose some questions that I could apply to the game in this manner? It need not necessarily be useful information, just collectible data that can be processed statistically.

Any thoughts would be greatly appreciated as I believe I can turn EVE into a useful tool for my study.
Teckos Pech
Hogyoku
Goonswarm Federation
#2 - 2015-10-08 04:11:43 UTC
Pixel Piracy wrote:
Hi all,

I am currently doing a crash course through statistics and probability. I want to try and apply what I am learning in EVE as I think I would be much more engaged with it. It would also serve as a functional tool to keep my skills in practice.

So my question is this: What sort of statistical questions could I ask of EVE? I am looking to collect data to calculate averages and derive other statistical results from.

I've considered taking note of the damage of my shots (slow arties) and calculating the average damage over the course of a fight(s) or mission. Also thought about calculating average ISK per/h from various sources.

Can anyone else propose some questions that I could apply to the game in this manner? It need not necessarily be useful information, just collectible data that can be processed statistically.

Any thoughts would be greatly appreciated as I believe I can turn EVE into a useful tool for my study.


Well, provided you can get the data there are all sorts of things you can do. I was curious to see if there were any sort of seasonal effect in PLEX prices. So I downloaded some data and put it into SAS and did a simple trend regression with monthly dummies and even weekly dummies...nothing. Well, the trend variable was significant, but no monthly dummy was.

With the PLEX data and the recent PLEX price change, depending on how sophisticated you want to get you could do quite a bit there. For example a two staged least squares model could be estimated to determine the demand curve for PLEX prices, possibly even the price elasticity of demand (which is getting into economics).

As for simple averages, your wallet contains a considerable amount of data you should be able to download as a csv file you can bring into Excel and use as well. There you can calculate averages, variances, standard deviations, correlation coefficients, and all kinds of stuff.

Another thing you could do is try invention, if you have the skills, and keep track of your successes and failures and see if how far off you are from what one would expect from invention results.

"The curious task of economics is to demonstrate to men how little they really know about what they imagine they can design."--Friedrich August von Hayek

8 Golden Rules for EVE Online
Xearal
Dead's Prostitutes
The Initiative.
#3 - 2015-10-08 04:26:41 UTC
You could try your hand at exploration and see if you can figure out the chances of various bits and bobs found in
sites, type of sites they are most found in and ofcourse average value per given type of site.
Then expand this to various racial types of sites and figure out which racial type of site is the most profitable.

Does railgun ammunition come in Hollow Point?
Otso Bakarti
Doomheim
#4 - 2015-10-08 04:35:01 UTC
Do something useful. Give us the stats on the actual customer decline. How much and at what rate...I doubt you'd have time to test this "seasonal" excuse - It's summer. Some interesting data is here.

There just isn't anything that can be said!
Trevor Dalech
Nobody in Local
Of Sound Mind
#5 - 2015-10-08 05:36:06 UTC
Some ideas I have been playing with, note that these go a bit beyond "calculate averages" and are more in the category of "derive other statistical results".

1) Do an ARIMA time-series modelling of the number of players logged in, or of the Jita price of (commodity of choice).

2) Find the statistical distribution of the price of (commodity of choice) and use either max. likelihood or moments methods to estimate the various parameters of this distribution.

3) Use multiple variable regression to figure out the relationship between the price of (commodity of choice), sec status, distance to trade hubs, and any other significant variables.

4) Figure out the price relationship between the various ores/minerals and how variable this is.

5) Use logistic regression to calculate how the odds of getting blown up on a gate depend on sec status (origin and destination system) distance to trade hubs, average population of systems, average number of people traversing the gates, etc... Identify any outliers (relatively safe/dangerous gates.)
Aoife Fraoch
The Scope
Gallente Federation
#6 - 2015-10-08 08:42:34 UTC
If you can navigate a package like R there is a lot of fun to be had with the market APIs. As a data source there is a lot available and you can get a reasonable volume of observations to play with.