These forums have been archived and are now read-only.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

 
  • Topic is locked indefinitely.
 

All Killmails Public, as suggested FF 2014, 3rd Party Dev Tool Update

First post
Author
Jinn Aideron
#1 - 2014-05-31 14:53:16 UTC
Session's video: CCP Games - Fanfest 2014 - 3rd Party Dev Tool Update
Time index: 22:10 - 23:32

In light of a suggested change making all killmails publicly available, I wanted to, very briefly, log two concerns as requested feedback with the development team. (Securing rights on future "I told you so"s.)

A sample from December was presented on stage citing 96.4% of all killmails with both attacker and defender being players are already available via killboards today. Why not roll out the rest.

Here is why. Please correct me where I'm wrong.


  1. The remaining 3.6% (in this month) are fights in which all involved parties unanimously, albeit separately decided NOT to publicize this data. If only one had, it'd be out there.
  2. Consequently, CCP additionally publicizing exactly these particular fights will in 100% of all cases act against everyone involved in it; while the people positively subscribing to the idea of killboards are never involved in one of these additional fights in the first place.

    This change would exclusively, and very precisely, hit the 3.6% self-selected sample of people unanimously having decided against publication.


  3. You would create a perfect bullet. Take away all uncertainty.

  4. Today, a kill not showing up on killboards is usually no strong evidence for it not having happened. With each and every last fight being published, it will be. You can trace everyone perfectly, without fault, without deception, without effort. I'll be the first to jump on this perfect data mining train, but I do not believe this to be a good game design decision.

    In other places exact formulae aren't published by CCP, events and occurrences are pseudo-random, pseudo-organic. Being able to tell with perfect knowledge that any one(!) event in all history of New Eden DIDN'T happen is damning.

Stealth deletes are bad.

CCP FoxFour
C C P
C C P Alliance
#2 - 2014-05-31 18:27:17 UTC
No need, at this time it is not happening. Some other people convinced me not to. Killmails relating to wars however will still be made available.

@CCP_FoxFour // Technical Designer // Team Tech Co

Third-party developer? Check out the official developers site for dev blogs, resources, and more.

Sven Viko VIkolander
In space we are briefly free
#3 - 2014-05-31 20:37:57 UTC
CCP FoxFour wrote:
No need, at this time it is not happening. Some other people convinced me not to. Killmails relating to wars however will still be made available.


I was really looking forward to this data being released. Mind briefly explaining the reasons which persuaded you?

The reasons cited by the OP are simply bad. First, there's no evidence that "all 3.6%" remaining (from a month like dec e.g.) are from pilots who actually decided to keep the information private. I don't know how it would be possible to even get evidence for this claim, but very likely some of the kills are from new players, not off-the-killboard-grid-by-choice players who only fight other off-grid players. The only guaranteed way to stay off killboards is to never undock.

The second reason cited is negated by the fact that 96.4% of all kills are recorded for a month like dec. Hate to say it but the "uncertainly" of not being able to google players and get corp histories, kill histories, etc. disappeared years ago. No significantly new data mining ability would open up just from the remaining data points being public.
Arronicus
State War Academy
Caldari State
#4 - 2014-06-01 18:09:26 UTC  |  Edited by: Arronicus
Jinn Aideron wrote:

Here is why. Please correct me where I'm wrong.


  1. The remaining 3.6% (in this month) are fights in which all involved parties unanimously, albeit separately decided NOT to publicize this data. If only one had, it'd be out there.
  2. Consequently, CCP additionally publicizing exactly these particular fights will in 100% of all cases act against everyone involved in it;


I have to disagree on this point here. When there is a fight between more than two parties, with more than one attacker present, a killmail is ONLY generated for the attacker that gets the final blow, and a lossmail for the defender. All other attackers and participants in the fight do not actually have any control over posting the kill/lossmails. It is often the case however in these mixed fight situations (common in lowsec, and some highsec fights) that some of those who are on the mail would want it posted, but have absolutely no control over having the mail posted, speaking from experience here, so to say that ALL involved parties don't want the killmails available is simply incorrect.

For those of us who value killmails, who pride ourselves in the kills we get, and being able to show them off, this denies us that bragging right, that prize, simply because whoever got the final blow isn't in a corp that auto-posts mails, and the victim doesn't bother.

Furthermore, as was mentioned above, not posting the killmails is not explicit proof that they chose not to post the mails, but rather they may not have known how to, may not have cared enough to, may not have remembered to, may not have even realised they got the kill, or even known what a killmail was. I would be HIGHLY surprised if over 20% of your 3.6 figure of unposted mails is made up of cases where the aggressor consciously chose not to post the mail.
Jinn Aideron
#5 - 2014-06-02 09:01:54 UTC
CCP FoxFour wrote:
...
Thanks for the heads-up, appreciated!

Stealth deletes are bad.

Jinn Aideron
#6 - 2014-06-02 09:53:50 UTC
There's not much sense in going deeper into this after Foxfour's reply. But since you went through the trouble to post, I'll briefly reply as well.

Sven Viko VIkolander wrote:
...
Simply bad.

  1. If you wanted to go into this amount of detail, namely 'Alarm! Players are acting against their held convictions, so we can't measure that!', then please account for this error on the 96.4% side. Which would occur about 27x as frequent.
  2. This is uncertainty. It doesn't mean unlikely. It means you cannot, in your machinations, in good conscience discount room for error. With this change upon us, you could. xkcd, 882, Significant.


Arronicus wrote:
I have to disagree on this point here. When there is a fight between more than two parties, with more than one attacker present, a killmail is ONLY generated for the attacker that gets the final blow, and a lossmail for the defender. All other attackers and participants in the fight do not actually have any control over posting the kill/lossmails. It is often the case however in these mixed fight situations (common in lowsec, and some highsec fights) that some of those who are on the mail would want it posted, but have absolutely no control over having the mail posted, speaking from experience here, so to say that ALL involved parties don't want the killmails available is simply incorrect.

Now this makes a hell of a lot more sense than Sven, and is a good point. Personally, I suspect this corrected for would make the largest difference.

Arronicus wrote:
For those of us who value killmails, who pride ourselves in the kills we get, and being able to show them off, this denies us that bragging right, that prize, simply because whoever got the final blow isn't in a corp that auto-posts mails, and the victim doesn't bother.

"For those of us who don't want to be tracked publicly every step of the way, we are already subjected to other people's wishes being valued higher 96.4% of the time." This just said towards your bringing forth competing interests. However, that indeed every participant, including logi, SHOULD generate this mail, there is hardly an argument to be had between us.

Arronicus wrote:
Furthermore, as was mentioned above, not posting the killmails is not explicit proof that they chose not to post the mails, but rather they may not have known how to, may not have cared enough to, may not have remembered to, may not have even realised they got the kill, or even known what a killmail was.

"Posting killmails is not explicit proof that they chose to post the mails, but rather they may forget about their auto-post, their corp posts w/o consent, or consultation, may not have cared enough to toggle it off."

You are right that it is inexact w/o sitting behind every player's chair and checking. You say you'd be surprised if the actual number wasn't much less, I say I'd be surprise if the actual number wouldn't go up a lot, if we could correct for these shortcomings. But these are the numbers from CCP and zKillboard for December, have you better?


We can argue around the edges a lot, but it remains that if you publish the now unpublished, you are targeting exactly who detest it the most, for little or no additional benefit to the rest.

Stealth deletes are bad.

Rain6637
GoonWaffe
Goonswarm Federation
#7 - 2014-06-13 17:11:31 UTC
the decision to not publish all killmails surprises me. it's one of the tangible proofs of activity in EVE.

...as long as it's not retroactive. (omg my PVE losses)
Karbowiak
State War Academy
Caldari State
#8 - 2014-06-16 09:55:45 UTC
CCP FoxFour wrote:
No need, at this time it is not happening. Some other people convinced me not to. Killmails relating to wars however will still be made available.


wait what..

i.. Evil