Yeah, I don’t think the difference in the 2 values is due to those killmails being missing from zkillboard. That is too large of a gap to be sensible.
So there must be another explanation and I’d love to see the code Squizz is using to generate the count when you search between 2 dates.
I’m an active PvP player and I neither upload my kills nor have my killboard linked. Some people just don’t care about zkillboard. I’d prefer CCP get rid of external kill links altogether as they are an overpowered source of intel.
It’s not that they are unable to, they just use different formulas to do so over larger time periods, so when prices move quickly in game values are deliberately slow to catch up. I think this is because back when it moved quickly people exploited it to make ISK using mechanics that paid out based on item values.
I think the code for zkillboard is available on the site somewhere.
Yeah, you are right. It’s on Guthub here:
Interesting thing just running effectively December last year (30 Nov - 30 Dec), the search result on zkill is now showing 40,645 kills, without the ‘pvp’ filter, but then adding other filters returns different result, so it isn’t clear exactly what ‘pvp’ filters for (since awox, gank and padding are also measured in the MER, but the ‘pvp’ filter seems to disregard them):
I’m just going to dig into the code a bit and see why the different results, but I suspect that the source of the discrepency is in the zkillboard code and it has more kills than it sometimes reports (40K is much less of a discrepency than 22K v 45K in the MER).
You have to select the pvp button, you are showing both pvp and NPC killmails.
Aiko has the “PvP” label selected on hers. It’s not selected on yours.
That’s why I’m going to dig into the code, since awox, ganked and padding are also effectivly pvp, but it isn’t clear what the ‘pvp’ filter applies in terms of the database count at the end.
So you are right, without the pvp filter it’s returning a significantly higher result, but pvp as a filter might not mean what we think it means in terms of how that translates to a database count.
EDIT: It genuinely looks like zkillboard just has a tonne of missing kills from highsec, but it just doesn’t seem logical.
So comparing zkillboard (30 Nov - 30 Dec) to the MER for December they generally align ok, noting the 1 day difference in time. Except for highsec:
| System Class | Systems | MER Losses | zkillboard pvp losses | % on zkillboard |
|---|---|---|---|---|
| Highsec | 1192 | 45,632 | 22,841 | 50% |
| Lowsec | 688 | 146,176 | 137,409 | 94% |
| Nullsec | 3321 | 167,798 | 160,650 | 96% |
| Wormhole | 2604 | 38,926 | 36,179 | 93% |
| Abyss/Pochven | 200 | 17,264 | 12,688 | 73% |
So now I’m pulling all of the killmails that zkillboard recorded last month and then I’ll pull the list from the MER of what it has extra for highsec and post back the results. Just seems really weird that so many pvp kills are missing from zkillboard.
Yeah thanks. Got it also.
The unfiltered 5K discrepancy between MER and zKill would mean MER fairly good data coverage, statistically speaking: zKill has around 90% of all kills. I suspect an overwhelming majority of the missing 10% are NPC related from folks that don’t know about zKill in the first place (rookies, perma-alone pilots, etc). Which implies zKill would have >90% of all PVP related kills, which puts bounds on derived statistics.
They don’t show in the MER.
Ah I see, that’s generally true. It oddly does have killmails involving NPCs though, like Forward Operating Bases.
If the high sec statistic is too hard to look at, we could look at their lowsec one instead and see if there was indeed the trend of people temporarily stopping lowsec activity once Uprising was was announced (but before Uprising release), since you’ve shown good coverage between zKill and MER.
Because the category “structure” contains both NPC (fob, diamond EC) and non-NPC structures . Guess it’s a bug on CCP side.
It also had some drifter kills.
got it. The structures KILLS also count into PVP activity.
So those Tendhyes | Losses | System | zKillboard are also considered as PVP by the MER. (there are three losses in the MER for those filters)
Excess kills between Zkb and MER on 2022-12 region Solitude for HS systems
| type | zk | mer |
|---|---|---|
| Caldari Shuttle | 0 | 1 |
| Atron | 0 | 1 |
| Paladin | 0 | 1 |
| Succubus | 0 | 1 |
| Raven Navy Issue | 0 | 1 |
| Ibis | 0 | 1 |
| Merlin | 0 | 1 |
| Guristas Forward Operating Base | 0 | 37 |
| Astero | 0 | 1 |
| Heron | 0 | 1 |
| Capsule | 0 | 1 |
same for domain.
Note : the number are minimum values, not exact value. if kills a,b,c,d appear during a month for a type and a system, and mer shows a,b,c while zk shows b,c,d the diff will be 0 while the reality is +1,+1
| type | zk | mer |
|---|---|---|
| Badger | 0 | 2 |
| Enyo | 0 | 1 |
| Hulk | 0 | 1 |
| Iteron Mark V | 0 | 1 |
| Bestower | 0 | 1 |
| Barghest | 0 | 1 |
| Blood Raider Forward Operating Base | 0 | 223 |
| Capsule | 0 | 233 |
| Sigil | 0 | 3 |
| Tengu | 0 | 1 |
| Exequror Navy Issue | 0 | 1 |
| Caldari Shuttle | 0 | 27 |
| Boobook | 0 | 1 |
| ‘Packrat’ Mobile Tractor Unit | 0 | 1 |
| Helios | 0 | 1 |
| Sleipnir | 0 | 1 |
| Gnosis | 0 | 1 |
| Cheetah | 0 | 1 |
| Federation Navy Comet | 0 | 1 |
| Dominix Navy Issue | 0 | 1 |
| Anathema | 0 | 2 |
| Typhoon Fleet Issue | 0 | 1 |
| Astero | 0 | 3 |
| Magnate | 0 | 8 |
| Mobile Depot | 0 | 39 |
| Mobile Tractor Unit | 0 | 22 |
| Bantam | 0 | 8 |
| Condor | 0 | 3 |
| Heron Navy Issue | 0 | 1 |
| Slasher | 0 | 3 |
| Machariel | 0 | 2 |
| Rifter | 0 | 6 |
| Reaper | 0 | 10 |
| Thunderchild | 0 | 1 |
| Executioner | 0 | 28 |
| Crucifier Navy Issue | 0 | 1 |
| Inquisitor | 0 | 1 |
| Tormentor | 0 | 18 |
| Impel | 0 | 1 |
| Naga | 0 | 2 |
| Impairor | 0 | 8445 |
| Punisher | 0 | 82 |
| Burst | 0 | 1 |
| Ibis | 0 | 38 |
| Kestrel | 0 | 10 |
| Merlin | 0 | 3 |
| Heron | 0 | 5 |
| Velator | 0 | 75 |
| Imicus | 0 | 1 |
| Atron | 0 | 9 |
| Algos | 0 | 3 |
| Dragoon | 0 | 1 |
| Praxis | 0 | 2 |
| Coercer | 0 | 3 |
| Corax | 0 | 1 |
| Cormorant | 0 | 1 |
| Talwar | 0 | 1 |
| Stabber | 0 | 1 |
| Moa | 0 | 1 |
| Maller | 0 | 1 |
| Venture | 0 | 15 |
| Catalyst | 0 | 3 |
| Crucifier | 0 | 45 |
| Thrasher | 0 | 2 |
| Vexor | 1 | 0 |
| Paladin | 0 | 2 |
| Kronos | 0 | 1 |
| Rupture | 0 | 4 |
| Harbinger | 0 | 1 |
| Gallente Shuttle | 0 | 9 |
| Drake | 0 | 1 |
| Athanor | 0 | 1 |
| Minmatar Shuttle | 0 | 29 |
| Rattlesnake | 0 | 1 |
| Raven | 0 | 1 |
| Amarr Shuttle | 0 | 180 |
The missing vexor correspond to that page Sarum Prime | Losses | System | zKillboard (one present in the MER, two in kz)
The huge impairor gape is Deepari | Losses | System | zKillboard : 8417 in MER, 13 in zk.
kills in the MER where attacker or defender is in an NPC corp (excluding starter, default and FW corps)
attacking corp
1000173 14
1000156 1
1000251 2
1000137 2
1000274 249297
defending corp
1000173 26
1000297 1
1000137 1
1000233 12
1000134 31263
1000132 1346
1000099 3
1000001 338
1000224 9
1000127 15582
1000223 3
1000251 8
1000121 20
1000216 13
1000215 4
1000148 1628
1000048 2
attacker is mostly Vigilant Tyrannos | Corporation | zKillboard
defenders are mostly Blood Raiders | Corporation | zKillboard and Guristas | Corporation | zKillboard so FOBs
Notable other are InterBus | Corporation | zKillboard and Doomheim | Corporation | zKillboard
Those NPC kills only amount to 2% of the kills though
WTF is a Boobook?
