EVE General Discussion

That's a great question.

The data is there for us as players to be able to answer that, but it's a huge task. Something CCP Quant could do more easily (but still a lot of data to analyse).

Getting the data is why it's a big task for us as players.

Let me get some data and the post the size of the problem back here. Will update in about 2 hours after pulling some of the data.

For an NPC corp CAS is easily the most structured and active with a lot of initiatives from the "natives"... for an NPC corp. While that's good in its own way what it does advocate is "don't join a player corp because scary things happen if you do, you're just better off not joining an actual corp". Not all of them are like that obviously, lots of recruitment happens from the trips they organise. There are however quite a few who explain it that way.

So it's commendable that they do put in a lot of effort but at the same time they sortof send the wrong message.

The vast majority that I've heard say to hold off on joining a player corp until you feel that the NPC corp is holding you back, and even then, don't just rush in blindly to the first player corp who promises you the stars (Nightcrawler 85's guide is linked often in that case). While there a few people who say to avoid joining player corps entirely, they're generally poo-poo'ed pretty quickly by others.

I hear more people say to avoid joining player corps in the Rookie and Help chat than I do in CAS, which, honest, is bloody terrifying. I'd love to be able to report those folks to get them muted (or even banned) from those two channels, because it really is a terrible disservice to make that claim.

Anyway, that's just my own thoughts on it based on what I've heard during my own time in CAS.

e: I'll edit to add that I don't think that the "wait until you feel you're held back" applies to all of the NPC starter corps. Just CAS, since they have a pretty good support system in place. Every other starter corp I've been in has been either a ghost town or has been nothing but bad advice and massive carebears. It would be interesting to try and get similar systems set up in them, but I haven't the RL leadership abilities to try, and I'll be damned before I start saying "someone should go do this" when I can't do it myself.

No problems and great that you'll check tomorrow.

This is turning into a all day project.

To obtain data that we can use as players to see if their is any statistical significance in retention differences between the starter Corps, there are a few questions to be answered:

1. How many characters have started in each starter Corp?
2. What percentage of those are still active if still in the starter corp?
3. What number left their starter corp and remained active for a period of the next 3 months?

We could check much later after a character leaves a starter corp, but reasons for staying and/or leaving at that point are probably not influenced much by their starter corp (that's an assumption I'm making to limit the amount of data required).

There is another underlying assumption - that other reasons for staying/leaving (eg. RL financial changes, etc.) are the same across all starter Corps, so the influence of those factors can be ignored in the final result.

Unfortunately as players, we don't have direct access to the status of each character to see if it is currently active or not, so we only have circumstantial data available to see whether a character is active.

These include indicators such as:

1. killmail and/or lossmal activity
2. change of corporation
3. contracts data (I'm still checking if this contains character information)
4. market data (also checking at the moment to see if we can obtain character data from this)

The amount of data is massive and requires accessing data from a number of sites, but by placing some constraints on the amount to be collected, I'm hoping we can obtain a measure that is valid (though that is still to be determined).

That creates another issue - what constraints to put on so that the amount of data is still valid. Obviously I can't 'reasonably' pull data for every character from every currently active corporation, so aside from the starter corporations and other NPC corporations (ie. corps you go to if you leave a player corp), I'm checking the top 50 Corporations as a starting point.

Just the starter corps alone is a huge task because of the total number of characters, most of which are not active at all:

Current number of characters in starter corps

So currently I'm working on filtering the number of characters to check within this number of characters (while trying to be kind to Squizz Caphinator by limiting the number of calls I make to zkillboard and evewho).

The biggest hurdle is to request the data while still being considerate of the servers and my own ISP. Will post an update in another couple of hours.

That bit is not difficult. You can just call the Eve api for the corporation sheet.

The code for that bit is here:

https://gist.github.com/shaetads/a774f39e015ac20db5d961a8994fd3d7

Getting the player Corps is also not that difficult as Squizz Caphinator has a great summary on evewho:

https://evewho.com/corp/

That provides a good start, to identify active corporations (both in and not in an Alliance) to then look at their members to see if there is a statistically significant difference in the starter corporation (though I'm still not sure if that is a valid approach, and will figure that out through some testing).

It's also relatively easy to pull the list of Corporations in Alliances from the Eve api, but then finding the top 50 is a bit slower as all Corporation Sheet for each one needs to be requested and the member count obtained individually. That's about 4500 Corporation sheets.

In total there are over 350,000 Corps (both active and closed), so I'm not going to pull the data for all of them from the eve api.

After that, to obtain the details of activity, there is are several sources (eg. eve api, zkillboard api, evewho api, eve-icsc alliances xml, etc.). For example, to pull kill/loss data for a Corp from zkillboard, the api call is:

https://zkillboard.com/api/corporationID/98388312/ (for Pandemic Horde Inc.)

zkillboard has a response limit of 200 records, so to pull additional data '/page/X/' is requred (where X is the page number).

From each set of 200 records, it's easy to obtain the characterID of the victim and each of the attackers.

So by requesting a substantial amount of data over time (I have emailed Squizz to let him know), it's easy to build up a list of characters from each Corp that is currently active as an attacker and or victim, which provides an indication of activity based on pvp.

Obviously pvp activity doesn't capture all activity. So from there it's on to pulling data based on Corporation changes of membership and anything else that shows a signs of activity by a character.

It's a matter of digging for data from several places that indicate activity of characters and then comparing that back to their starter corp, then looking to see if there is a statistically significant different between the starter corps.

In 2014, there was an average of 420,000 unique players active per month and there is an average of 1.5 accounts per player. The number of logins is lower now, but even a rough calculation suggests that the amount of data for even the last month of activity across different activity types is going to be large.

The big concern is being polite about the number of requests required to pull a representative sample of that data; and ensuring that the requests and responses meet the requirements of the server owner (eg. requesting gzipped data from zkillboard). It's going to take a while., but writing the code is easy.

I highly doubt you've been in all 12 starter corp chat channels so how can you confirm they are mostly ghost towns?

I have plenty of ground to stand on since I belong to RUN starter corp and I know exactly how it is in corp chat.

Why did you create this thread anyway, was it to incite rage or to stealth promote the removal of NPC starter corps?


DMC

Some of us have, or had in my case, alts everywhere and for various reasons. That means we get to experience many/most of the NPC corps. Two of them stand out for being quite active (for an NPC corp), CAS and SWA and generally that's a positive thing. Most others are either quiet, filled with Russians or the usual memes and insecure jokes nonsense.

Post 2 of 2 because the full text has too many quotes.


I have no clue how you were running them, but CAS still have a lot of success with doing similar ops. There are occasional troubles, nobody will deny that. A part of EVE is teaching people how to prepare for them and how to react to them once they show up. If your fleets rolled over and died and "became a memory" it sounds like you were too focused on efficiency.



I, for one, have absolutely no desire to see NPC corps go. I would love to see a bit more structure available (the "social corps" that have been discussed in the past) as a transitional half-way point between NPC corps and Player corps. Channels and mailing lists don't always cut it.

CAS has a lot of good going on, including a variety of sub-groups dedicated towards specific areas of play (nullsec, mining, wormholes, etc). If that can be replicated across the other starter corps, it can only be good for the game. Unfortunately it will end up coming down to needing people able to take the lead to set up things in those groups. I said before that I would love to do it myself but don't have the (RL) leadership abilities to do it.

Nobody in this thread mentioned anything about getting rid of NPC corps except you.

Just a slight correction:
The reason why a lot of people stay in CAS, it's because it's active.
Events are a consequence of the activity.

Another reason is the environment you just don't want to leave ...
... or you'll plant an alt there just so you can stay.

Thank you for your input. :)