These forums have been archived and are now read-only.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

Seeking Providence Player with long term logs.
Author
Previous Topic Next Topic
Satchel Darkmatter
Hedion University
Amarr Empire
#1 - 2017-02-17 20:07:50 UTC
Ideal situation is I am looking for a player who's active in Providence, who's in the main intel channel and who spends an obscene amount of time logged in creating logs/data, one of those running 10 account always mining or ratting type guys whos got at least 6 months worth of intel channel logs.

I know there are a handfull of you out there, so if your willing all id need it you to zip your log's and mail them or upload them drop box or what ever we can arrange the finer details..

WHY I'm working on an Intel Map, like near2 and vintel but with extra features that I wanted personally, one of the things I'm working on now is what I like to think of as an EvE weather forecast, basically checking the overall threat level of a given region and then using past data to predict hot or cold periods and then presenting that information to the tool user in a weather forecast type info glyph.

I'm not 100% sure if this will even worth the way I plan but I cant really test it without a decent amount of data at my disposal.

anyway if you need more info or have questions either mail me or get me in game in the EmberIntel channel.
DeMichael Crimson
Republic University
Minmatar Republic
#2 - 2017-02-18 01:10:40 UTC
Shocked

Sounds cool, definitely interesting to say the least.

Hope you have good luck and much success with your project.



DMC

'The Plan' | California Eve Players | Proposal - The Endless Battle
Soel Reit
The Scope
Gallente Federation
#3 - 2017-02-18 02:23:47 UTC
SPAI SPOTTED
Neuntausend
Republic Military School
Minmatar Republic
#4 - 2017-02-18 02:29:52 UTC
It's not like spying on providence is particularly difficult.
Salvos Rhoska
#5 - 2017-02-18 08:49:08 UTC  |  Edited by: Salvos Rhoska
I dont think your idea will work.

Paradoxically, the more historical data you have incorporated into this project, the more inaccurate your predictions will become.

The reason for this being, that though it is true that the past is the matrix that defines the future, that past has also informed/resulted in each past future, infinitely.

You are attempting to build a bridge over a river, from which to observe the river flowing towards it, under it, and past it.
But the problem is that that river's flow is never the same, twice, ever.

The more data of the past you include, in an attempt to form a predictive model for the future, the further it becomes skewed towards predicting the past accurately, rather than the future.

This notion is explored in Asimov's fictional/theoretical concept of Psychohistory.

PvE v PvP <> Old School Exploration <> CODE Licenses <> CODE Special Agent
Satchel Darkmatter
Hedion University
Amarr Empire
#6 - 2017-02-18 09:08:04 UTC
Salvos Rhoska wrote:
I dont think your idea will work.

Paradoxically, the more historical data you have incorporated into this project, the more inaccurate your predictions will become.

The reason for this being, that though it is true that the past is the foundation/structure on which the future is built, that past has also informed/resulted in each past future, infinitely.

You are attempting to build a bridge over a river, from which to observe water flowing towards it, under it, and past it.
But the problem is that that river's flow is never the same, twice, ever.


Hence the term weather forecast ;) . The basic idea behind it is that in any given year the busy periods will mostly be the same for most regions, big random political war's aside the average day to day activity levels should for the most part be similar, so for example with this weeks data from last year for providence I could see that mon-fri would be fairly quiet but this time last year at the weekend its crazy busy.

I have already tested my logging over a few days and written some code to present it as a graph and its actually interesting to see the peeks and troffs through the day as waves of reds come and go.

at the vert least this is interesting to see.
Salvos Rhoska
#7 - 2017-02-18 09:13:58 UTC  |  Edited by: Salvos Rhoska
Satchel Darkmatter wrote:
Salvos Rhoska wrote:
I dont think your idea will work.

Paradoxically, the more historical data you have incorporated into this project, the more inaccurate your predictions will become.

The reason for this being, that though it is true that the past is the foundation/structure on which the future is built, that past has also informed/resulted in each past future, infinitely.

You are attempting to build a bridge over a river, from which to observe water flowing towards it, under it, and past it.
But the problem is that that river's flow is never the same, twice, ever.


Hence the term weather forecast ;) . The basic idea behind it is that in any given year the busy periods will mostly be the same for most regions, big random political war's aside the average day to day activity levels should for the most part be similar, so for example with this weeks data from last year for providence I could see that mon-fri would be fairly quiet but this time last year at the weekend its crazy busy.

I have already tested my logging over a few days and written some code to present it as a graph and its actually interesting to see the peeks and troffs through the day as waves of reds come and go.

at the vert least this is interesting to see.


You have some very interesting hypothesis there that your model would indeed test.

Such as whether behavior is cyclical, seasonal or periodic.

However, as I said, the more historical data you include, the more accurate your model will be for predicting the past futures (with 20/20 hindsight), but the less accurate it will be to predicting the current future. The weight of past data, will pull your predictions into the matrices of the past.

I hope you manage to build this model. I will be very interested in the results.

PS: The choice of Providence, according to its literal definition, is certainly an auspicious and fitting choice for this project!

PvE v PvP <> Old School Exploration <> CODE Licenses <> CODE Special Agent
Satchel Darkmatter
Hedion University
Amarr Empire
#8 - 2017-02-18 10:30:34 UTC
it most certainly is cyclical and seasonal and periodic, the reasons behind this I suspect will be heavily related to national holidays or other seasonal and annual events, for example in a region where the attacking force is mostly American then the region under attack should see a reasonable uptic in red activity simply because more of them are off work or school and have more time to put into EvE, This behaviour should be reflected in the results, should be annual and should be a good predictive indicator for future uptics in red activity.

Given that EvE is multi national I would be willing to bet that there are event's throughout the year for many countries that the rest of the world are clueless about, I bet most Americans don't know when the UK has long weekend's / Bank Holidays just like we don't know when they have theirs.

This type of stuff is for the most part obscured, but I suspect that tracking the red activity in any region will begin to highlight these periods, and given their annual nature , yes they should be fairly accurate in predicting the future.

Will it be 100% accurate, NO!, nothing ever will, what it should be is Interesting, and accurate enough to tell the Rorqual pilots that mining on a day or few days when their hostiles have a holiday and have spare time to attack is probably not such a good idea.
Salvos Rhoska
#9 - 2017-02-18 11:31:17 UTC
Yes, but the very fact EVE is multinational, means there are players with days off all the time.

Furthermore, how many stay at home to play EVE on the 4th of July in the US, or Christmas in any country that doesnt celebrate it on the 24th of December (such as in Finland we celebrate on the eve of the 23rd, and in Russia weeks later according to their calendar., or those that dont celebrate Christmas at all).

US and European schools/universities typically have weeks of difference in when terms begin/end.

Anyways, good luck!

PS: Out of curiosity, how much are you willing to pay for the data you are requesting?

PvE v PvP <> Old School Exploration <> CODE Licenses <> CODE Special Agent
Kraxalious
Domini Caedis
#10 - 2017-02-18 19:33:32 UTC
This type of model would only have a remote possibility of working in regions with long term stability of the same owning alliances. Once a new alliance rolls in then their habits would take over. Go through and look at the history of many individual systems. Certain systems have never been held by the same alliance for more than a few months at most in the entire history of eve although some transfers are peaceful. This makes predicting the future much more difficult however.
Soel Reit
The Scope
Gallente Federation
#11 - 2017-02-18 19:39:15 UTC
your datas are legit as ****!!!!!!!!!!!

you said that provi would have been busy in the weekend...
well only 3/4 different fleets atm roaming there!

NC, provi, spectre, bjorn

lol