Tuesday, October 06, 2015

Declaration of Victory: King of the Prison

Reached quorum 5 votes to 0 with the Warden in favour, after 12 hours. Enacted by Kevan.

Adminned at 06 Oct 2015 18:07:20 UTC

Everyone else is Injured. As such, I am the Kingpin and have achieved victory.

Assuming this DoV passes, I intend to pass the mantle for the next dynasty to Thrawn.

Comments

Thrawn:

10-06-2015 00:23:12 UTC

for

ais523:

10-06-2015 00:41:18 UTC

For anyone interested, I’ve had an agreement with Thrawn for a little over two weeks now that if Thrawn helped me to win the dynasty, I’d repay the favour by passing the mantle.

BlogNomic is single-victory, and agreements that persist between dynasties are frowned on, which normally makes it very hard to trade something of value for a win. However, the win and the mantle are both things that players value, so it made sense to split this win as one of each, because this was really a team effort. (And besides, I actually got to pull off a timing scam at BlogNomic, which is something that is very hard to achieve nowadays because coordination is so difficult with the IRC channel dead. We were crossing PMs continuously after it had succeeded…)

Once our agreement was first put in place, I didn’t really have any plans as to how to win, and just decided to float along for a bit, seeing what happened, and putting myself in a generically good position (I have 0 SHIV). Thrawn did have plans, having successfully noticed that Sentence was a good thing via nearly all the dynasty’s game mechanics (with Riot being the only exception, and unlikely to be achieved in the form it had at the time), and successfully used them to Injure Aname. However, it would have started becoming less effective as numbers dwindled (at some point, two players would likely have banded together to Attack Thrawn; I was planning to interrupt that if I could, but obviously there’s no guarantee I’d have been fast enough).

Later on, I realised that I could Injure players that weren’t paying attention via Attack defection. Tantusar showed signs of not being online every 24 hours (which, to be fair, is perfectly reasonable and shouldn’t really be required), so I launched an attack on the already-Injured Aname with him. Tantusar didn’t notice, and so was taken out. The proposals that I proposed recently were self-interested, but not in the usual way (I really didn’t care whether they passed or failed, and don’t particularly benefit or lose out as a result of them). They had two purposes: a) to hide the fact that the dynasty was likely to end soon (via making it look like I was thinking about the long game), but mostly b) to push the Attack on Aname below the first screenful of the page, to reduce the chance that Tantusar noticed it. Aname was the target because as the Attack clearly couldn’t hurt anyone but me and Tantusar, it was unlikely to lead to a retaliatory Attack.

Then at the end, it was just a case of “consortium of two players can take out two other players before they can react via synchronized timing”, and there were only two other players left. I picked a time of midnight UTC (it was actually a few minutes later, due to communications delay; we had to ensure that we were both online first) on the basis that Kevan and Purplebeard would likely be asleep and so unable to react (and because the time was convenient for both me and Thrawn).

Finally, Thrawn Injured himself via Attitude change (I was actually surprised to see that that was legal; my brain filtered it out because it’d be so counterproductive in most circumstances) to end the dynasty before anyone could make a proposal to change the rules and deny the victory that way.

I joined the dynasty quite late, so wasn’t paying attention to how the earlier players got Injured. Anyone care to comment on that, so that this DoV thread sums up the entire history of how the victory came about?

ais523:

10-06-2015 00:43:57 UTC

Final GNDT: (where not otherwise specified, Prisoners are in no Gang, have no Grudge, and are Injured):

Darknight: 9 Sentence, 44 SHIV
Kevan: 9 Sentence, 40 SHIV, Grudge against Purplebeard
Tantusar: 21 Sentence, 48 SHIV
Purplebeard: 8 Sentence, 15 SHIV
Aname: 13 Sentence, 73 SHIV
Thrawn: 24 Sentence, 32 SHIV
ais523: 10 Sentence, 0 SHIV, Converted attitude
delcooper11: 15 Sentence, 0 SHIV

The Kitchen Crew were: Kevan, Purplebeard, Thrawn, and delcooper11.

ais523:

10-06-2015 00:51:21 UTC

I should also note that this sort of ruleset is completely rife with timing scams, which is something that should be kept in mind for the future. Pretty much every proposal that tried to fix them just traded one set of timing scams for another. I could thus effectively veto proposals just by pointing the scams out, which I did against any proposal that would have messed up my chances of victory.

Probably the most controversial proposal in this respect was The Exercise Yard at Three. I typed without thinking and pointed out the existence of a timing scam, without first trying to see if I could exploit it. (Then I decided to disclose the scam so that the original scam would continue working.) The “Attacks spaced 24 hours apart” scam is one with which I’d really love to have won the dynasty, but trying to pull that off would have been far too difficult without a coconspirator who’s very reliable in terms of timing scams. (Out of the people I’m familiar with, Kevan’s the only current Prisoner I’d have trusted to be able to do that, and being the Warden it’s unlikely that he’d cooperate in that.) Luckily, there’s a common behavioural pattern in nomics that if an attempt to fix a bug fails once, people don’t try again for a while (if ever), even though it’s not really logical to do so.

ais523:

10-06-2015 01:01:01 UTC

While I’m busy posting thoughts about this dynasty, another bug that was lying around but I didn’t need (but that Kevan intentionally exploited once, perhaps without realising the full implications): you could break the “anyone could betray you” pattern of the dynasty via Attacking, posting the Hash of a cooperation statement, and PMing the statement itself to your Accomplice (along with an explanation). The Accomplice would then have no fear of being betrayed, and would have a straight choice of who to Injure: the Target or the Suspect. In most cases, they’d choose the Target (because the Suspect had shown willingness to work with them, and because the Target is likely to be someone who you had a clear reason to get rid of).

And, of course, this wasn’t really based around the Prisoner’s Dilemma at all, both because high Sentence was good and that completely broke the payoff matrix, and because Cooperate/Cooperate was no better (even during Snitching) for reducing Sentence than not playing the dynastic mechanics at all. Except that it sort-of was at a meta-level (if someone tries to Escape, Cooperating is trying to stop them, Defecting is not trying to stop them; Defecting is better because it puts you at no personal risk and leaves you with more weeklies), except that the payoff matrix is still slightly wrong (that game is Chicken, rather than the Prisoner’s Dilemma). There was certainly a bunch of game theory going on, but more so than that, there were timing scams.

I suspect that if there’s any attempt to reprise this dynasty in the future, the best option would be to make it turn-based somehow.

Thrawn:

10-06-2015 01:46:11 UTC

Like ais523 said, my main worry was about getting rid of the last other prisoner. Trying to raise their sentence before we both get SHIV over 50 would have been difficult, and probably impossible.

Tantusar:

10-06-2015 02:12:54 UTC

for
I’m going to be honest here, these last couple of Dynasties haven’t really jumped out at me, which is why I hadn’t been participating <strike>much</strike> during this one, only taking a few actions to keep from idling. Actually, my lacklustre performance was probably quite a factor in this Victory.

ais523:

10-06-2015 04:04:31 UTC

@Thrawn: One thing you could have done is repeatedly Take a Fall to keep the other Prisoner’s SHIV below 50. That would stave off a loss, but it wouldn’t actually help you win. (Actually, I think the main danger to your plan was that the remaining players would see that they were being picked off and gang up to Injure you. That might not have been a problem if you had someone (either me or someone else) to work with, but it would have been a problem if you were going entirely solo, because you wouldn’t have been able to find someone for the counter-Attack.)

ais523:

10-06-2015 05:24:06 UTC

@Tantusar: We had backup plans, although we almost certainly couldn’t have won today if you’d been participating more actively. Our alternative plan (for if you’d Defected on the Attack on Aname) would have involved Injuring all but one other player, then attempting to Escape; it would have been very risky because the Escape could be stopped by proposal (and BlogNomic players are split on whether that’s fair play or not), but would likely have been worth trying.

Kevan:

10-06-2015 08:17:24 UTC

for Good work. And enjoyable that there’s still room for a surprise “actually, I defect on my promise to pass the mantle” from Ais.

No time to join in the post-game analysis right now, but I’ll come back to it later.

ais523:

10-06-2015 08:20:52 UTC

Right, it’d be theoretically possible to betray Thrawn on that, but being Emperor for one dynasty really isn’t worth nobody trusting your promises ever again.

@Kevan, for when you get back to the post-game analysis: I’m interested in whether you saw something like this coming.

Kevan:

10-06-2015 08:28:55 UTC

That seems a bit of a shame, if mantle-passing is seen as disconnected metagame because it occurs after the end of the dynasty, and we feel we shouldn’t betray players over it. Maybe we should shift it into the DoV, so that you have to name a winner at that point, and can do so in-character.

ais523:

10-06-2015 08:54:22 UTC

Ah no, I mean I have a personal rule of not breaking promises in nomic ever (however, such promises must be freely and explicitly made, and I don’t make them unless I know I can keep them). This is because it’s sound general strategy; game theorists know that the ability to make a binding promise gives you an advantage, and if you break a promise, people are unlikely to trust your ability to keep promises in the future. Thus breaking a promise is one of the worst things you can possibly do in a nomic.

Some places where I play games (e.g. Werewolf forums) have a rule that you have to try your best to win each individual game, rather than optimizing to win future games, and I respect that rule when it’s there. But as far as I can tell, BlogNomic has no such rule (and you would have to ban most of the players if it did). Thus I play to optimize my expected number of future wins, even if it means a temporary disadvantage. (In general, I’m quite a fan of the long game in nomics.)

That said, BlogNomic does have, not a rule, but a convention that alliances and conspiracies shouldn’t persist between dynasties, and I adhere to that. (The DDA lead to some rather bad blood back at the time it existed, and I think BlogNomic players wanted to move on from those days.)

Note that I also don’t feel any particular pressure to prevent other players winning, if it doesn’t affect my own ability to win. (In BlogNomic, I try to prevent other people winning mostly just because it reduces the number of opportunities that I get to win; there’s no other real reason beyond that. That is, when I’m willing to put in the effort to play at all, which I mostly only do when I see an above-average chance to go for a win.)

Purplebeard:

10-06-2015 11:55:41 UTC

for I always knew I’d never survive in prison.

Kevan:

10-06-2015 17:10:41 UTC

I figured that Ais’s tinker proposals were probably deliberate chaff for something, but hadn’t thought through the possibility of one player trusting another to the point of voluntarily injuring themselves. I was thinking in terms of the usual “X players feel they each have less than a (100/X)% chance to win, so team up and assign the crown randomly” thing, and figured that anyone who’d survived this far would have given themselves more than 50% odds of outwitting the others, and would have struggled to negotiate even a 55% coinflip with somebody, let alone a 100% surrender.

I didn’t think reveal-your-hash pre-commitment was a bug for the reasons you describe: if a player is cunning enough to pull the trick against a less cunning victim, their approached accomplice might decide that they’d rather eliminate the more cunning player, and deal with the less cunning one more easily later. There was a problem coming up with Attacks, though, now that a double-defect caused a SHIV boost: launching an Attack and revealing the hash of your defection would give your (probably randomly-chosen) accomplice a straight and unthreatening choice between “defect” and “die”. And we were only two such Attacks away from a Riot, last night.

The line about breaking promises is interesting. I think I probably take the same tack, although I don’t really feel that it persists across games and can’t think of anyone here who I’d consider particularly untrustworthy. Maybe betrayal dynasties would benefit from some kind of silly “each player secretly chooses whether to be a Cretan or a Theban; Cretans are encouraged to lie and break promises if it would gain them an advantage” rule somewhere so that players could comfortably defect in character on anything they liked, right up until the mantle-passing.