My CSM election analysis post got record hits and even references from blogs what would never link me. Yet, I was never satisfied with the numerical model used for the analysis, but it was better than nothing. It had many flaws, mostly that it ignored candidate order, treated a first-second cross-vote equally to a fourth-fifth, but the fifth-sixth was cut off.
Finally I came up with a much better model and redid the analysis. I used the published CSM8 vote exhaustion data. It told us that those who voted for only one candidate had 51.9% of their vote exhausted. This means that a first-spot vote has 0.482 weight. A two items long vote had 17.4% exhaustion, meaning that the second spot vote has 1-0.174-0.482 = 0.344 weight, and so on. I smoothed the later spots with a fitted curve and assigned weight to every spot. The exact weights can be seen on the left side of the result of a test vote with 100000 votes, all cast the same way: A, B, C, D...
On the right side, you can see the new cross-vote table. The lines are divided by the self-vote (the diagonal). The first line mean "who did the voters of A voted for", the first column means "whose voters voted for A?". 72% of A voters supported B, because the first spot A votes has 0.482 weight, the second spot B votes had 0.344 weight and their ratio is 0.72. The B voters fully supported A, since they placed A before B on the ballot. As you can see, the third-fifth votes have sub-10% weight, the rest are very small.
Let's run this calculation on the real CSM9 votes to get a ranking of candidates according to the model. You can see them below. You can also see that all the STV winners (blue background) got ranked before losers, so the model is very good.
Now to the point, the cross-vote table:
Let's start with Podli, Psychobitch, corebloodbrothers and DNSBLACK. They are mavericks. Their columns are red, meaning that voters of other candidates did not support them. The lines of the first three are red, meaning that their voters supported nobody else significantly. DNSBLACK voters supported the "good block". Only corebloodbrothers got elected from them, having Providence block behind him. They also had the largest exhausted votes, worst begin corebloodbrothers voters, having 0.68 waste for every vote cast on him. These non-winning candidates should reconsider running again. They all should seriously consider getting political agreements and cross-support, both to get in with votes coming from elsewhere, and to give their spillover votes to friendly candidates instead of letting them waste.
Psychotic Monk, and Funkybeacon should be good example to them: they had mediocre cross-support, and DJ Funkybeacon had cross-support with Gorski Car and Sugar Kyle. He got in in.
The next group to discuss is the wormholers. They were 5 but Karen Galeo isn't on the table for being irrelevant. While the rest 4 cross-supported each other in the green, the numbers are between 20 and 40. Compare it to the Sion-Mynnna 60+! Wormholers did not follow the official "support all WH candidates" line, wasting votes and getting only Corbexx in. More on them at Foo's.
The next block contains 2 PL candidates (Gorski Car, Major JSilva), progodlegend of Nulli and Aram Kachaturian of Pasta and but it's not a real block, as the candidates cross-supported only the ones next to them on the list. The two on the middle, who had two pairs won, the two on the sides with only one buddy did not.
Next is Matias Otero all alone (change from the original post). He has no real cross-supporter, but he has grey coloumn and line, meaning that practically everyone supported him a very little. Lot * small = enough.
Next is the "good ones" block, all elected: Steve Ronuken, Sugar Kyle, Mike Azariah, Ali Aras. Green cross support, though weaker than among wormholers. They also got support support from many candidates, but they - except Ali - only gave very little outside support: Sugar to Funkybeacon and Mike to Xander Phoena.
Finally, the CFC block: Ali, Mangala, Xander, Jayne Fillon (not elected), Angry Mustage (not elected) and the hard core: mynnna and Sion. The openly CFC candidates got minimal outside support, only Xander had a small cross-support with Mike.
From the last election the CFC knew that they have enough voters for 2 seats, but they wanted more. They got in Xander because the voting participation was low and CFC was disciplined and voted. They had two tricks, one failed, the other worked, even fooling me. The one that failed is putting Mynnna behind. The plan was that Mynnna is widely acknowledged to be a skilled and hard working CSM8. Even Ripard - the central figure of the oppossing "non-aligned block" of CSM8 - suggested people to support him and expressed surprise for no outside votes. He got a weak support from the N3PL candidates, probably for putting progodlegend and Major JSilva to the irrelevant positions of the CFC ballot. Oh, and a little from Psychotic Monk, but nothing else. Why? Because Mynnna is a Goon and as such, strongly rejected outside of his own block.
The Goon trick that worked well was planting 3 Trojan Horses: Ali, Mangala and Jayne. They didn't campaign with a Goon arm patch, actually they campaigned with neutral and non-aligned platform. Jayne failed because he was exposed by Seraph IX Besarab as a spy and meta-gamer, and soon the rumor got out "he infiltrated Bombers Bar for Goons". From his #5 CFC ballot position, it's likely true, but that doesn't matter. Even the rumor of Goon affiliation was enough to finish him, he got no support from outside of his block.
There is a huge surprise in the community for Mangala losing 2/3 of his #1 votes, half if we consider the overall voter drop. I shamelessly take credit for that: I spammed every possible channel with "RvB is Goon slave" and as you could see it with Mynnna, being associated with the Goons isn't the best way to get outside votes. While he got in, he did so only with block-votes, he only get low support from outside of his blue rectangle. However it means that my original conclusion was wrong: he is isolated from the rest of EVE and supported by CFC. For better or worse, the fate of RvB is bound to CFC.
Ali Aras was complete success. Her column is red only for maverick voters. Practically everyone else loved Ali, catapulting her into the permanent seat. Despite there were signs like posting on the Goon propaganda site, she could maintain a neutral campaign and got strong outside support. She didn't need the votes coming from the #3 position of the CFC ballot, she would have been elected even if her name was erased from votes containing Sion. I have to admit the genius work of Goons here: getting a Goon CSM in with the votes of those who hate Goons. The only mistake they made is not trusting her enough. If they didn't put her to the CFC ballot, she would still get in without being exposed as a Goon and could be re-used next year. (I mean re-used as a Trojan horse, of course she can get in next year, but with CFC votes only).
What should we all learn from these results: politics matter. You can't just read the programs and support nice candidates. You must look for cross-votes, endorsements, pacts and such. I mean the other candidates and opinion leaders should, and voters should listen to them. I should have listened to those who suggested Steve and Mike to support. I don't have to like them. Pick your #1 after your personal preference, but the rest are hard politics. Learn from my fail! The Goons are not playing nice (surprise!), and we must not be naive. We must be careful and vote for only those who are proven to be not affiliated with those who want to destroy our game.
Finally I came up with a much better model and redid the analysis. I used the published CSM8 vote exhaustion data. It told us that those who voted for only one candidate had 51.9% of their vote exhausted. This means that a first-spot vote has 0.482 weight. A two items long vote had 17.4% exhaustion, meaning that the second spot vote has 1-0.174-0.482 = 0.344 weight, and so on. I smoothed the later spots with a fitted curve and assigned weight to every spot. The exact weights can be seen on the left side of the result of a test vote with 100000 votes, all cast the same way: A, B, C, D...
On the right side, you can see the new cross-vote table. The lines are divided by the self-vote (the diagonal). The first line mean "who did the voters of A voted for", the first column means "whose voters voted for A?". 72% of A voters supported B, because the first spot A votes has 0.482 weight, the second spot B votes had 0.344 weight and their ratio is 0.72. The B voters fully supported A, since they placed A before B on the ballot. As you can see, the third-fifth votes have sub-10% weight, the rest are very small.
Let's run this calculation on the real CSM9 votes to get a ranking of candidates according to the model. You can see them below. You can also see that all the STV winners (blue background) got ranked before losers, so the model is very good.
Now to the point, the cross-vote table:
Let's start with Podli, Psychobitch, corebloodbrothers and DNSBLACK. They are mavericks. Their columns are red, meaning that voters of other candidates did not support them. The lines of the first three are red, meaning that their voters supported nobody else significantly. DNSBLACK voters supported the "good block". Only corebloodbrothers got elected from them, having Providence block behind him. They also had the largest exhausted votes, worst begin corebloodbrothers voters, having 0.68 waste for every vote cast on him. These non-winning candidates should reconsider running again. They all should seriously consider getting political agreements and cross-support, both to get in with votes coming from elsewhere, and to give their spillover votes to friendly candidates instead of letting them waste.
Psychotic Monk, and Funkybeacon should be good example to them: they had mediocre cross-support, and DJ Funkybeacon had cross-support with Gorski Car and Sugar Kyle. He got in in.
The next group to discuss is the wormholers. They were 5 but Karen Galeo isn't on the table for being irrelevant. While the rest 4 cross-supported each other in the green, the numbers are between 20 and 40. Compare it to the Sion-Mynnna 60+! Wormholers did not follow the official "support all WH candidates" line, wasting votes and getting only Corbexx in. More on them at Foo's.
The next block contains 2 PL candidates (Gorski Car, Major JSilva), progodlegend of Nulli and Aram Kachaturian of Pasta and but it's not a real block, as the candidates cross-supported only the ones next to them on the list. The two on the middle, who had two pairs won, the two on the sides with only one buddy did not.
Next is Matias Otero all alone (change from the original post). He has no real cross-supporter, but he has grey coloumn and line, meaning that practically everyone supported him a very little. Lot * small = enough.
Next is the "good ones" block, all elected: Steve Ronuken, Sugar Kyle, Mike Azariah, Ali Aras. Green cross support, though weaker than among wormholers. They also got support support from many candidates, but they - except Ali - only gave very little outside support: Sugar to Funkybeacon and Mike to Xander Phoena.
Finally, the CFC block: Ali, Mangala, Xander, Jayne Fillon (not elected), Angry Mustage (not elected) and the hard core: mynnna and Sion. The openly CFC candidates got minimal outside support, only Xander had a small cross-support with Mike.
From the last election the CFC knew that they have enough voters for 2 seats, but they wanted more. They got in Xander because the voting participation was low and CFC was disciplined and voted. They had two tricks, one failed, the other worked, even fooling me. The one that failed is putting Mynnna behind. The plan was that Mynnna is widely acknowledged to be a skilled and hard working CSM8. Even Ripard - the central figure of the oppossing "non-aligned block" of CSM8 - suggested people to support him and expressed surprise for no outside votes. He got a weak support from the N3PL candidates, probably for putting progodlegend and Major JSilva to the irrelevant positions of the CFC ballot. Oh, and a little from Psychotic Monk, but nothing else. Why? Because Mynnna is a Goon and as such, strongly rejected outside of his own block.
The Goon trick that worked well was planting 3 Trojan Horses: Ali, Mangala and Jayne. They didn't campaign with a Goon arm patch, actually they campaigned with neutral and non-aligned platform. Jayne failed because he was exposed by Seraph IX Besarab as a spy and meta-gamer, and soon the rumor got out "he infiltrated Bombers Bar for Goons". From his #5 CFC ballot position, it's likely true, but that doesn't matter. Even the rumor of Goon affiliation was enough to finish him, he got no support from outside of his block.
There is a huge surprise in the community for Mangala losing 2/3 of his #1 votes, half if we consider the overall voter drop. I shamelessly take credit for that: I spammed every possible channel with "RvB is Goon slave" and as you could see it with Mynnna, being associated with the Goons isn't the best way to get outside votes. While he got in, he did so only with block-votes, he only get low support from outside of his blue rectangle. However it means that my original conclusion was wrong: he is isolated from the rest of EVE and supported by CFC. For better or worse, the fate of RvB is bound to CFC.
Ali Aras was complete success. Her column is red only for maverick voters. Practically everyone else loved Ali, catapulting her into the permanent seat. Despite there were signs like posting on the Goon propaganda site, she could maintain a neutral campaign and got strong outside support. She didn't need the votes coming from the #3 position of the CFC ballot, she would have been elected even if her name was erased from votes containing Sion. I have to admit the genius work of Goons here: getting a Goon CSM in with the votes of those who hate Goons. The only mistake they made is not trusting her enough. If they didn't put her to the CFC ballot, she would still get in without being exposed as a Goon and could be re-used next year. (I mean re-used as a Trojan horse, of course she can get in next year, but with CFC votes only).
What should we all learn from these results: politics matter. You can't just read the programs and support nice candidates. You must look for cross-votes, endorsements, pacts and such. I mean the other candidates and opinion leaders should, and voters should listen to them. I should have listened to those who suggested Steve and Mike to support. I don't have to like them. Pick your #1 after your personal preference, but the rest are hard politics. Learn from my fail! The Goons are not playing nice (surprise!), and we must not be naive. We must be careful and vote for only those who are proven to be not affiliated with those who want to destroy our game.
6 comments:
"From the last election the CFC knew that they have enough voters for 2 seats, but they wanted more."
I'd rather have two CFC and twelve unaligned, because getting everyone in unison is more convincing to CCP that way than if I had five or six or eight CFC.
Don't believe me?
Two words: "Drone Assist"
But you can only convince them if enough of these "unaligned" are Goon spies like Mangala and Ali.
Or, when CCP failed hard with the original design. Even I would have removed drone assist.
Again, was there any candidate whose agenda was to improve
a.) hisec game mechanics
b.) PVE/missionrunning?
c.) improving the game mechanics outside pvp so that cooperation will be rewarded and not punished?
Where can I read a statistic about how many "people", not alt accounts, reside in hi-, lo- and nulsec (wh-space)?
CSM election campaigns are a farce as they try to give lobotomized, braindead players the impression that CSM members actually cared for their voter's interests and would therefore work towards improving the game and helping the interests of their voter's base.
Nothing could be further from the truth!
As EVE is irrelevant for the RL of almost every player, they couldn't care less voting or not!
No matter who is on a CSM seat, CCP will not begin doing the right changes....
So far, all the CSM "winners" are niche players wit hvery specific interests. Therefore only their clients voted for them as they don't have much support within the majority of the player base and that ist the hiseccarebearpubbieshitlordbaby-noobassholenolifer40yoldvirgin!
cont...
Most of these players know that they are treated by all the other players and especially by CCP as scum, shit and dirt. Irrelevant to the game and blablabla....
Only trouble is, neither mittens or gevlon can finance this game alone. All these "braindead losers" are needed, which ccp knows and therefore alle the 0.0 boosts and hisec nerfs are just so that pve/missionrunning isn'T completely unbearable, but just so that one can levelup one's Raven (or maybe even RNO, Golem or SNI) to run constantly Lv4 missions...
I've flewn tengus, ravens, scorpions and Machas... LV4 grind to no end... I fly tanked ships and have never lost a ship to gankers, although flying regularly for SoE and the epicarc empire faction corps...
I only lost a Macha to gankers during an incursion fleet, which really bothere me, not because of the loss, I couldn't care less...
It took me almost an hour to rebuy my stuff and what bothere me most wasthat the a-type invu dropped!!!!!!
I do pvp from time to time in small fleets (3-40) in low, nul or wh space... mostly it is boring as in low everyone docks once the local blob shows up. Even sending tacklers 2-3 systems ahead..
Nul is the same, either empta or docked or hotdropped, no real suspense.
WH!!! on the other hand is amazing, one never knows... and then flying with some RL buddis in bombers and other cloaky ships is just priceless...
cont...
Which brings me back to my PoV of EVE.
The main most activity is PVE missionrunningand this activity is the least cared for by CCP!
No new missions, no new classes of missions, a braindead activity like mining...
Exploration, which used to be fun, was executed and brutally murderen by CCP when most sites where accessible via hitting Alt+D and forbidding tech3 cruisers entering the interesting sites.
The minigame is not a game as such, it is like PI a boring clickfest
Either you have enough virus strength (skill to LV5!!!) or not! Nothing chance based or with finesse... but the worst was the introduction of these MF.. ing minicontainers, which no one can ever pickup all alone. Basically, these sites now take more time and give much less loot. And most loot is just crap! But LO, every newbie can do it, it is fun, it is easy it is unrewarding...
Like MTU's they practically eliminated the need for skilling a noctis! What a wonderful was to make some money for a beginner, accompany a pve veteran and clear the sites after him, share the loot and salvage and you earn more than you ever could doing these things by yourself.
Now the veteran has several MTU in his cargo, deloys them and doesn't need or want the beginner flying with him as with the introduction of MTU's droprates were nerfed because ccp anticipated that everyone would use them...
There is no PVE content for small fleets (2-5) and none can be created by players so FUCK THE SANDBOX!!! EVE is not a sandbox, it is just a game with strict mechanics that can and will be exploited for a certain time until CCP sees it fit to change it, either because more and more people apply this exploit and the CCP friends who found out about it first have made enough iskies with it or because a certain ship/fit/strategy is so commonly used that it becomes a standard (Drake fleets/drone assist, tengu, hulk) which no longer allows for flying other ships...
Every player corp which calls itselv beginner-friendly is insta-tagged as desirable WT! Although this could be totally avoided by not joining a player corp or starting one-man corps with a mutual lightblu network... Is that cooperative multplaying? Biulding your own one-man corp with no corp assets, so one can one minute after being wardecced leave and create the next corp?
How ridiculous is that?
And CSM never tackeld these problems, NEVER EVER!!!!!
If anything some of the CSM strongly oppose the idea that anyone should be able to leave a wardecced corp
So why the hell should dedicated highsec players care about those lobbyists that want to get a seat in CSM???
Why should PVE carebears waste only one millisecond thinking about CSM candidates?
cont...
Oh and by the way:
Another proof of failed game mechanics. Large weaposn have the most difficulty hitting small targets or doing lot of damage (BS vs frig) but small guns are being able to penetrate large and capital shields and armor?
A small tech2 blaster doing damage on a 1600mm armor plate? R U FUCKING SERIOUS???
Post a Comment