Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
Aran, I believe my posts are on topic because the poll is a statistical assessment of the audience's intentions, and a discussion on the validity of statistics is fundamental to the topic.
Nach, I have read the two books you link. They are my favorites, especially the first one. I encourage you to read them carefully because they do not state what you claim. For "How to lie with statistics", that can be inferred from the title alone. See this for a good definition of "lying":
Harry Frankfurt wrote:
It is impossible for someone to lie unless he thinks he knows the truth. Producing bullshit requires no such conviction. A person who lies is thereby responding to the truth, and he is to that extent respectful of it.
Therefore, when the author attempts to teach you how to lie with statistics, he implicitly states that you can tell the truth with it. In fact, modern science is founded on the rigorous application of statistics. No details will be provided here because this forum is not the place to clear misconceptions.
In fact, statistics is so important, that when an experimental paper comes to me and does not include a rigorous analysis, I reject it without reading, explicitly telling this to the editor, which is always very thankful to me pointing this out. There's absolutely no negotiation with respect to this. Without statistical analysis of empirical results, the work is simply storytelling with numbers. Even considering this for publication is degrading the work of many who labor many hours to produce rigorous results.
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
Nach wrote:
p4wn3r wrote:
You are saying that you can see one post and use psychology. That's not how it works, because statistics doesn't work for a sample of one.
Statistics don't work for explaining anything at all, because correlation does not indicate causation.
If you thought psychology is about statistics, then you don't understand psychology.
Certainly, you would not mind signing your name under this statement and send it to the committee who condemned a guy for misconduct after concluding he faked data after they grew suspicious of many statistical anomalies and ask them to reconsider the verdict. If statistics does not matter at all, there's no problem if the data is fake in the first place.
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
Nach wrote:
p4wn3r wrote:
You are essentially stating that the podcast from Scott Adams you linked to allows one to understand the psychology of the poster better than the poster himself, which is quite an extraordinary ability that Scott Adams appears to have.
One of the things he discussed was taking a statement from someone else and restating it to expand it into an absurd absolute.
That's proof by contradiction. It's a very useful method.
I also could say that you shouldn't pay attention to this because it's not published in a peer-reviewed channel, not authored by someone with a degree in psychology, and not funded by research agencies that review grants. Although, to be fair, not even these factors can be taken for quality nowadays.
But, seriously, when you see something in psychology, even not very controversial statements like "Men are more aggressive than women.", it only means that someone picked a large sample of men and women, made some efforts to make sure it's not biased (for example, the men are not all schoolboys and the women are not all cagefighters) and observed that more men made more aggressive acts than women.
You are saying that you can see one post and use psychology. That's not how it works, because statistics doesn't work for a sample of one.
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
Nach wrote:
People are often unaware of their subconscious biases. I doubt he was aware he was performing mind reading, but it's clearly the case from the post itself. Please see everything I wrote in this thread about psychology and the links I provided to understand this topic better.
I have difficulty verifying this claim. You are essentially stating that the podcast from Scott Adams you linked to allows one to understand the psychology of the poster better than the poster himself, which is quite an extraordinary ability that Scott Adams appears to have. After hearing his podcast, I asked my subconscious if his description of me is more accurate, but unfortunately my subconscious has not returned my request for comments.
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
The thing that gives some hope is that, for practical purposes, you don't need THE optimal input that completes the game. You only need one that's good enough. So, you might come up with some models and approximations that might generate a good enough TAS much more efficiently than a human.
As an example, Deep Blue, that beat Kasparov in chess, was a supercomputer. Nowadays even smartphones, that process much fewer positions, can play better than it. The difference is that engines today are much better at selecting positions that matter and discarding useless ones, and so are much stronger.
Incidentally, that overlaps with a bit of my work. There are many physical models that don't have analytical solutions and you have to use computers to approximate the solution. Because these theories always involve some constants that you determine experimentally, there's some uncertainty in the theoretical prediction, and it doesn't make much sense to run the simulation at a higher precision than these values are known.
Many people confuse this and believe that a simulation that came out of a supercomputer is always better, because it calculates more. In practice these simulations use very general numerical algorithms (because they have to work for lots of systems) that converge extremely slowly. Sometimes they even have to make ugly approximations to make the method eventually stop, and the value they converge to is not really the theoretical prediction. Depending on the system you are working with, if you choose a model with care, you can get a more accurate value even with pencil and paper than a general numerical package running on a supercomputer.
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
feos wrote:
If this post doesn't even try to talk about actions staff is supposed to do regarding submission voting, then I stand corrected.
Wait, so by "censoring votes" you actually mean that I suggested that the site should censor votes from sockpuppet accounts? Then, by all means, yes, I do suggest that these votes should be censored. Don't you do that already?
feos wrote:
p4wn3r wrote:
feos wrote:
Stop blaming people on a web site. You've been asked for sensible suggestions repeatedly, repeatedly refused to provide them, only to repeat useless whines.
I'm sorry. Now you are simply lying. Let me remind you what actually happened. You asked me for suggestions in this thread, and I immediately replied that I thought a good idea was to abolish anonymity. If I am repeating this, it's simply because you refuse to understand this and keep repeating that I'm wanting you to censor votes, and I really don't know why you are saying this..
This indeed happened. And my detailed reply tells why voting abuse potential is vanishingly small. I addressed this once again in the post you're replying to, but for whatever esoteric reason you don't want to discuss your original idea. Your later posts don't mention anything that can't be done without making the votes public. But they mention enough impossible to implement and controversial things.
Please, stop this. I discussed the idea with you on another post. It's fine if you did not want to, but don't suddenly put the blame on me.
feos wrote:
This is the key to this whole issue. I look from the perspective of a person who wants to resolve these problems officially. Because of that I always try to find pros and cons to any idea, on the long run. If something is inapplicable, it should be figured out in discussion. If something isn't fine-tuned to the point common staff agreement, then it won't work and it won't become a policy. Staff agreement implies user agreement as well, but is not limited to it. If you are not ready to properly discuss your own ideas, don't expect us to do the mental leg work for you.
Really? To me it looks like you're just JAQing off so that you can make me look incompetent and at the same time avoid doing your job. I should remind you that it's not my job to do mental leg work for the staff. When I write something about site policy, I am not in fact making a submission for rule changes, I am simply writing something about the site policy. You can address it or ignore it, I really don't mind. But, please, don't keep coming to me for a solution to your problems. That's consulting, and in the real world, people charge money for that.
Is this a joke? That post gives the definition of an investigation. I am not suggesting anything to be done with the voting system.
feos wrote:
Stop blaming people on a web site. You've been asked for sensible suggestions repeatedly, repeatedly refused to provide them, only to repeat useless whines.
I'm sorry. Now you are simply lying. Let me remind you what actually happened. You asked me for suggestions in this thread, and I immediately replied that I thought a good idea was to abolish anonymity. If I am repeating this, it's simply because you refuse to understand this and keep repeating that I'm wanting you to censor votes, and I really don't know why you are saying this.
Later, I suggested that you could avoid drama by investigating the matter. A sensible reply to this is "I did not think there was enough to warrant an investigation in this particular case". Instead, for some unknown reason, you say that it's not appropriate to request an investigation unless I, of all users of this site, provide you with concrete guidelines on when to start or not start an investigation.
I don't know why you are obsessed with me, a regular user, having to provide advice to you. Are you willing to pay for these services? In that case, I might consider that.
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
feos wrote:
Since you're not suggesting any official improvements, your posts are basically just counter-productive whines.
I am very happy by the recent revelations that my posts have many readers. I didn't know I was so influencial. I find it nice that in such a short time span I've read that my posts are well articulated and now that they are counterproductive whines. Such feedback is what every respectable person dreams of. Thank you.
I don't know why you are saying that I'm wanting the site to censor votes. What I suggested is to stop making them anonymous, but since at this point it's just text interpretation skills, I don't think it's worth talking about it.
For the record, I do whine a lot. Especially at my printer, it does not work very well. But the fact that some people at an internet website have difficulties handling complaints is at most a minor annoyance to me.
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
Dude, there's never a well-defined set of criteria about when to start an investigation. Do you like the idea of doing investigations over whether you should start an investigation? Or maybe doing an investigation over whether you should start an investigation to determine if the case needs an investigation?
In practice, it goes like this: if someone requests the investigation, and you deny it, and eventually another party determines that there was indeed a problem, that makes you automatically an accomplice. So, usually you tend to investigate most things and ask people to not take it personally. Of course, in this forum we are not dealing with lots of money, so this is more of an inconvenience.
But take the case of rotten tomatoes for example. Suppose a studio suspects that a rival is creating facebook bots to review-bomb their movie, and Rotten Tomatoes replies that there's no need to investigate. Then they hire a private investigator who proves this is happening. Now they can sue Rotten Tomatoes for millions of dollars, even if their staff never review-bombed the movie.
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
Zeupar wrote:
I am surprised to see you use this approach to articulate your arguments. I don't think irony is the most clear or elegant way to get your point accross. Then again, my impression is that in recent times your posts have become notably bitterer than they used to be, so it might be related to that. I honestly just hope that your recent exposure to Europe hasn't had anything to do with that. :(
I will not answer remarks like this. If I write anything that's offensive, please point out that I will do my best to understand. However, I don't argue about style.
@feos:
The problem is not what the rules already say. The thing is, that if you don't investigate, you give the impression to people that you are covering up. What the rules say is useless if you never investigate anyone.
feos wrote:
Where should the borderline for this be? Suspecting unfair feedback is a valid concern. But if there's no clear and sensible borderline to unfair feedback that has to be investigated, then requesting such investigations can also become abusive.
I really don't see how a "borderline" would help. Say you make a borderline that if someone makes three offensive posts a week you will investigate for trolling. Then someone starts making two offensive posts every week so that you don't investigate. See the problem?
Besides, I really don't see what is abusive about sending a PM to some people asking "It was alleged that you voted on this submission to degrade the author. Is this true?". It's definitely easier than having long-winded arguments about the poll that somehow failed to stop drama in submission threads for all these years. Even if people admitted voting unfairly, the punishment would be just a warning anyway. I have seen people in threads reply that voted Yes without watching a movie, and were just warned by other forumers. To my knowledge, no one is asking for more rigorous punishing.
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
Memory wrote:
This to me is the definition of a witch hunt:
Merriam Webster's wrote:
: the searching out and deliberate harassment of those (such as political opponents) with unpopular views
Why specifically the "No/Meh" voters, why not the "Yes" voters?
Absolutely not. The difference between an investigation and a witch-hunt is that, in a witch-hunt the crime/violation is imaginary. It's impossible for the accused to be guilty, therefore investigating an imaginary crime is the same as harassment.
However, in this case, suppose you open up the "No" votes and it appears that all come from users with the same IP address. That would be conclusive evidence that manipulation took place. Since it is possible for the violation to exist, it's not a witch-hunt.
When do you start an investigation? When there's sufficient suspicion that a crime/violation happened. As was pointed out by many in the original thread, submissions such as those are usually uncontroversial, and the number of No/Meh was unusually large. There was nothing unusual with the number of Yes votes.
And what happens when an investigation finishes? There are many possible things:
1) You determine the subjects are innocent. In this case, you can decide to dismiss the initial allegation with or without prejudice. Dismissing with prejudice means that the investigator concluded the allegation was filed in bad-faith, so the accuser should be punished. Without prejudice means it could be part of a misunderstanding.
2) It's impossible to determine whether the subject was guilty or not. In this case, no action is taken because of presumption of innocence, and the investigator suggests what changes should be made to determine rigorously in the future.
3) The subject is determined guilty. In this case, the evidence is provided in the final report, which is forwarded to the person who gets to decide the punishment.
As I said, it's perfectly possible that an investigation finishes with the party that accused having problems. There's nothing to do with harassment, and in case anyone get investigated eventually, I really suggest collaborating with them instead of calling it a witch-hunt (as Trump likes to do).
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
Nach wrote:
The problem here is that some people are not voting on the run at all, they're voting on what they liked or disliked about what you had to say regarding SMB and TASing in the past couple of weeks. They'll even make up excuses why they voted the way they did instead of just outright saying "I'm angry at you because I don't like what you've written recently, and penalizing all your SMB submissions because of it." (yet in their own minds not necessarily see it as an excuse, they may 100% believe they don't like the run)
That's exactly what people were saying at the thread at the time, and instead of addressing the issue, all responses were like "yeah, the poll is like that. it's not appropriate of you to complain about that. stop calling out the voters."
The situation is extremely easy to solve. When HappyLee canceled the submission, say something like "it has reached us that the submitter canceled the submission in protest regarding the validity of the No/Meh votes. Given that this submission has received an unusual amount of negative votes, we decided to investigate. Upon consultation with the No voters, we received the response that they would have preferred a more minimalistic approach to the levels than the ones provided here. Although troubling, we did not find any conclusive evidence of voting guideline violation. We ask that the users vote based on the merits of the movie and not on previous events and ask the submitter to reconsider the cancellation in view of these facts.", and that's it, no drama, no bans.
Instead, we get lots of philosophical arguments about art and professionalism, arguing about arguing, and talks about statistics (which, I would let you know, is used everywhere to detect anomalies to spot fake data, in elections and scientific publications) which lead nowhere at all.
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
Actually, what you are saying is simply that one algorithm is better than the other. For one particular problem. You can search all word combinations, or search all dictionary words. In this case, searching all dictionary words is more efficient.
However, if you are dealing with computer programs, to improve on the naive method requires the code to be extremely simplified. We already have an example of this. By the authors' account it consists of around 800 instructions and there is only a main loop without hardware interrupts. That's what makes it possible to simplify everything and perform the search.
With respect to the game this was discussed (SMB), one possible situation that a movement simulation does not cover is some sort of ACE setup that could glitch Mario through all the stages at enormous speed. Ruling out ACE quickly becomes unfeasible as the code grows in complexity. Moreover, for the particular submission, a shorter movie was presented. That alone suffices to prove the simulation in question was not complete.
It's like, when someone says "to make a computer play chess, you have to search millions of positions", and you reply "not really! chess might be one of these games where you can find a winning strategy without searching. While that really is true, there's no reason to believe we are anyway close to finding such strategy.
Extraordinary claims require extraordinary evidence. At some point if some guy comes up with a formula that plays chess perfectly, you have to ask what is more probable. That chess players and AI researchers for the last centuries all missed a key point in the game that someone eventually discovered like pulling a rabbit out of a hat, or that his/her proposed solution contains a mistake. In general, when something seems too good to be true, it usually is.
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
I think the situation is similar to election polling. For someone who understands statistics, there's some noise with respect to the methodology used and, because of that the result should not be taken extremely seriously. However, in the real world people do take it seriously, that's why it is usually a crime to conduct electoral polling without registering with the authorities.
It's simple to make a poll to favor a candidate, even if you don't rig the poll. You can interview a biased sample, that you know is more likely to vote for the person. You can run many polls and only publish the one where the candidate you favor has more answers, etc.
But, seriously, either abolish anonymous voting or simply don't put the poll at all. In the original thread (which had some parts grued with a nonsensical title implying that you cannot detect manipulation using statistics), people drew parallels between TASing and arts. Well, art critics do sign their reviews, and they eventually get sued or fired if they write inappropriate stuff.
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
Zeupar wrote:
I love your posts because they are well articulated in general and usually include interesting points and arguments. That's why I am dumbfounded after reading the quotes above. It's like you read a different topic from the one I linked...
I really appreciate that you like my posts, Zeupar. I hope this discussion does not change that. All I can tell you is that I am completely sincere that I see absolutely no controversy in these topics. Since HappyLee agreed with me, I think my description of his intentions was accurate.
If you're interested in seeing where I come from, I recommend reading this book, about a rather extreme case. It's about a guy who was deliberately faking all his results and most people were simply not considering this possibility because his results were claimed to be revolutionary.
If you read the book in detail, you'll see that his fraud was rather bad, and most people who doubted the results did not accuse him of cheating directly. They first assumed they were true, but difficult to reproduce . Then, they thought the measurement was in error, only when they could not think of any honest error that could produce the results they started suspecting deliberate fraud. Nevertheless the managers were giving bad reviews for people who questioned because they were not "good team players".
Now, is it impolite to say the managers were ignorant when they messed up the whole thing? Perhaps, but this is a very accurate description, because that's exactly what they did! Sometimes, when you're an expert on something, it is your job to contain the hype, even if that goes against most people's expectations.
That's why I strongly recommend that people don't jump to conclusions when they see something like this. All that this thing has generated is several pages of nonsense and unnecessary tension between MrWint and HappyLee.
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
Zeupar wrote:
For a good example of how to defend one's work in a respectful manner, something that you seem to be quite bad at, check MrWint's posts here.
I have just read that thread and unfortunately all I can see is lots of people with severe misconceptions about what bots are and how they are used.
To start off, I heavily object to the name "exhaustive search bot". Just by a simple calculation at the size of the NES RAM and the possibility of inputs, it's pretty obvious that with current technology it's impossible to make a bot that's truly exhaustive. The amount of possibilities is too much, even for a computer to check.
What MrWint did, and actually very well, is to put the game mechanics into a simulation so that he could solve it with algorithms. That's a highly nontrivial achievement and I did not see anyone dispute that.
However, what most people should not confuse this with, is that just because a run uses a simulation it is more optimized than one that doesn't. The result of a simulation can only be trusted if what is assumed to make it is valid in the first place. HappyLee was simply pointing this out in that thread. There were many instances where a bot search was, strictly speaking unnecessary.
This should not be interpreted as belittling the run (and despite your statement, it was MrWint who actually assumed it was). The previous run is a very elegant solution to an artificial intelligence problem, namely how to TAS SMB with algorithms.
Regarding the "No" votes, I can say that any input on low entertainment they could have has already been diluted in the pretty obvious bias against the author that some people who posted demonstrated. It was stated that it was unfair that MrWint is not an author (even when HappyLee offered him an authorship and he declined on that thread you link to), that Lee is a one-trick pony for TASing one game, and that he was arrogant in that thread. Sometimes you come up with new ideas and end up only reinforcing what was already known. That's life, it's dumb for other people to assume otherwise.
To sum up, what this looks like to me is: someone submits a TAS using cool and interesting techniques to a known game. People get overhyped because of "OMG BOTS", then an experience TASer comes and points out the simulation was incomplete, and then a group of people, after the hype dies out, start nitpicking and cast some very dubious "No" votes.
With all due respect to GoddessMaria, mass hysteria and mob rule are very real things. I don't see why the community should be above criticism and find depressing that people cannot point this out without threats of intervention.
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
Warp wrote:
"The requirement to apply for this job is either a degree in electrical engineering or at least two years of work experience in the field."
Again, based on cultural context, this is most certainly an inclusive or. It would be baffling and extraordinarily unusual if the requirement would exclude someone who has both.
Incidentally, that reminded me of a very nice joke.
A logician leaves his profession and decides to become an obstetrician. After his first patient has her baby with no complication, he congratulates the father, which promptly asks him:
"Doctor, is it a boy or a girl?"
And he instantly replies: "Yes."
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
The book I'm referring to, which I wrote in my first post, and the precise edition I have is this.
I don't know if the guy somehow changed things, but on the version of the book I have, on pages 22 and 23, the signs say exactly what I wrote here. I can scan if you want to see...
Anyway, let us see what happens if we change sign V as in the blogpost. What happens is that sign II is now true, which means that room II is indeed empty. The rest of the reasoning is the same and the lady is still on room VII.
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
FWIW, the reason I used some strange wording is that, because in the actual book, there are several puzzles of the lady and the tiger sort, this is the last of them, and in most of the previous puzzles there is only one lady and one tiger, so I explained the rules of the trial with these previous puzzles in mind and did not bother to check that my wording was consistent with this problem.
By the way, it is still not clear to me. You have assumed that there was at least one tiger, not exactly one? The solution to the puzzle (which is provided in the book, btw) really does have at least one tiger, as Fractal pointed out. So, if your reasoning is correct, you should have arrived at room 7, no?
About "either or", I copied it from the book and I always assume the same as Fractal. Anyway, at which step of your solution do you rely on "either or" being exclusive? That's not clear to me.
Sure, I could have said that the king always says the truth and the prisoner is a perfect logician who always acts by logic. What did you assume in your solution?
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
feos wrote:
Do silent votes help with any of this? No way.
Another problem, votes can be manipulated to provide desirable outcome. You just have to simplify the problem in people's eyes, and provide options that are based on that simplified problem. You can preset people's opinions if you're smart (and mean) enough.
We don't want to rely on that kind of things. This is why we don't believe in silent votes officially. We only believe in helpful discussion. Submission votes are a legacy thing, maybe people want it to be there. I don't think removing its has ever been considered. But it's virtually non-existent when it comes to serious mental effort.
tl;dr: don't mind silent votes, treat them as youtube comments. Because that's what judges do.
I agree with that, but I still think the votes help a little bit. It's a fast way to check on the popularity of things. Realistically, not everybody will take their time to post. The votes are a heuristic more than anything. If something gathers lots of No votes, it might not be a reason to give it a lower tier, but at least it's reason to check out what's happening.
There are many ways to manipulate votes and even what can constitute manipulation is debatable. I suspect many Yes votes in popular submissions come from the runner's friends or people who visit the site rarely and only post on their favorite game's threads. Can someone eventually see this as a group of people trying to manipulate the community's opinion? Perhaps, but we must also ask what's the problem with voting on submissions of a game you like or telling your friends "Hey, I made a run. Check it out!".
In any case, all the problems could be solved by simply revealing the names behind the votes (Facebook reveals the name of people who liked something because it's so easy to manipulate). If that information is public, it is the community's problem if they fall for the manipulation.
Also, in the case of a troll No/Meh vote, the thing a troll values most is anonymity. He just wants to click on a button and be done with it. However, if the name is public, he might have some explaining to do later, so I guess people would be inclined to vote No only if they truly believe it's important.
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
What I can tell you is that you are assuming this
Aran Jaeger wrote:
(1) it holds true that any statement (that refers to the system) that can be expressed is either true, or false, or undecidable
I know there are some logics where this holds, but this puzzle was thought in the context of classical logic, where this doesn't happen. The truth of a statement and its decidability are two different things.
If you have a consistent system, you can have statements that are not reachable by inferences from its axioms. These statements can be true or false. If it is possible for us to assign a true or false value to it and keep consistency either way, we say the statement is "independent".
As an example, take geometry without Euclid's fifth axiom. It's still a formal system. But if you want a system that's consistent (in the classical logic sense), you must decide whether the statement of Euclid's fifth axiom is true or false. If you say it's true you have Euclidean geometry. If it is false, you have elliptic or hyperbolic geometry (which one depends on what truth value you assign to other statements).
The fact that you cannot reach Euclid's fifth axiom by the other four means that it's independent of them, not that it is undecidable. Nevertheless, if you are in a consistent system, it is either true or false, there's no third option. Decidability involves statements that are always true in any consistent formal system, but nevertheless cannot be proven from the axioms (and for any sufficiently expressive system, they will always exist, because of Gödel/Tarski theorems).
The tricky thing about the puzzle is that it includes a statement about its own solvability, but in that sense, solvability should be treated in a classical sense. All statements are indeed either true or false, the question is whether their truth can be determined uniquely from the hints. I never thought about the problem when you allow a third truth value.
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
Now that you mention, an idea is to make the user names of all entertainment votes visible to all users in submission threads. It would certainly help mitigate abuses.
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
Essentially, if you spend months making a run and explaining how everything does, you should be open to people bashing it without knowing anything useful.
But if you just spend 20 minutes watching a movie, then you are free to write and vote whatever you want for any reason, because it is not acceptable to say that their criticism is dumb. Youtube users that scream "Cheats!" at every TAS should find this policy very nice!
Experienced Forum User, Published Author, Player
(42)
Joined: 12/27/2008
Posts: 873
Location: Germany
Correct!
Here's a rigorous proof as to why the king said Room VIII is not empty.
Suppose the king said Room VIII is empty. Then, the prisoner solved the problem and deduced that the lady was at room R1. Since the puzzle was unsolvable before, there must be at least one solution with a tiger on room 8 (since it can't have a lady) and the lady at a room R2, different from R1.
Now, sign VIII is:
VIII - This room contains a tiger and room IX is empty.
If VIII contains a tiger, then this sentence is false, which means that room IX is not empty.
Now pick the solution with a tiger at 8 and the lady at R2, and change room 8 so that it's empty, but keep its sentence false. Since room IX is still not empty, and room VIII no longer has a tiger, its sentence is indeed false. Because no other sign refers to the content of room 8, changing it from tiger to empty does not change the truth of other sentences.
So, we have found a solution with the lady at R2 and room 8 empty. However, that contradicts the hypothesis that the puzzle is solvable after the king gives us the information.
Therefore, the king must have replied that the room is not empty.
