= 27d567bc2d48d9f40e9ccc20ea370b15)
SilasBarta comments on How to always have interesting conversations - Less Wrong
You are viewing a comment permalink. View the original post to see all comments and the full post content.
= 783df68a0f980790206b9ea87794c5b6)
Loading…= b715d02e85f7b3b4ae65ca3d3e450868)
Subscribe to RSS Feed
Powered by Reddit
= f037147d6e6c911a85753b9abdedda8d)
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (331)
Oh, well, in any case, I did try Go for a while, and I do think you can explain it verbally. Before playing any human opponent, I figured out a very simple procedure for beating the computer, though it only works when you play white.
Just copy your opponent's moves, rotated 180 degrees about the center. It won't be until the endgame that your opponent takes the center. Then just play as best you can (it will feel like getting a free move anyway). At the end of the game, you'll have basically the same territories, but you'll be in the lead because of white's handicap (kyu or whatever).
I only briefly started trying this on human opponents, and for whatever reason, even on the major Go server, people would quit after a few moves when they saw me doing this.
I'm probably missing something big, but there you go.
They must have been unaware of these tactics. Many people consider manego annoying, because it's sort of a cop-out.
Whereas I consider the labeling and shaming of the opponent for making valid moves that are difficult to beat is either a total cop-out, an insult to the game or both. If you can't beat someone when you can predict and for most part determine what their moves will be then you seriously suck or the game is a solved problem. Like checkers or tic-tac-toe.
Seriously, if you think you are a better player and you credit your opponent with the slightest hint of strategic competence you should EXPECT them to do what you do until such time as they suspect they are risking falling into a manego-trap.
Sometimes a game has one serious flaw but is nevertheless fun to play, and there is no obvious fix for that one serious flaw. In that situation, it can make sense to shame opponents who exploit the flaw. There is a sense in which this is an "insult" to the game, but both players might still like the game, on balance.
For example, I have found that in Stratego, it rarely makes sense to attack first against a player of roughly equal ability. At a certain point in the mid-game, evenly matched players will usually both find it optimal strategy to move a piece back and forth dozens of times in sort of "chicken" game where the goal is to get the other player to attack first. This is boring, so I don't want to play with you if you're going to do that every game, but potential Stratego partners are rare enough that if you otherwise enjoy playing Stratego with me, I might try to shame you into being more reckless with your attacks.
I would treat shaming (as distinct from banter) in that context as a 'defection'. My response would be to then eliminate whatever suboptimal levels of recklessness that I had previously allowed to creep into my play in a spirit of cooperation or just any intrinsic recklessness that I had not chosen to stifle. Either that or I would disengage from the game entirely. Before doing so I would offer potential cooperative agreements if possible.
Most likely I would not find Stratego particularly appealing. If it is supposed to be about 'strategy' yet relies on people not using good strategies in order to work it is broken. I would much prefer to play a lighter game that at least doesn't pretend to be about strategy.
When playing the card game 500 the standard rules for 'misere' are not well balanced. When playing people who are not rank amateurs I advocate a limit of one misere call per player per 'game (up to 500)'. If the opponent insists on the standard rule then I proceed to play (open) misere whenever the risk/reward ratio is favorable. This tends to result in most games being largely determined by my misere calls, with me winning two thirds of them and 'going out backwards' the other third. Naturally I do so with playful cheer and offer to impose the restrictions at any time.
It can actually be quite fun to play the meta-game of negotiation. Winning the game convincingly even (and especially) under the 'broken' system they insist on but offering to adopt an agreement that will effectively be a handycap for me. Fogging all manipulative shaming attempts and repeating the offer. Engaging in a good natured battle of wills with those too stubborn to admit their folly or, given that admission, to change their mind. Getting the kitty a LOT. Doing the balancing act of keeping the experience fun despite the broken rules and the resulting conflict. Knowing when to stop and switch to a different game or activity entirely (thus practicing the ability to maintain boundaries and accept 'no-deal' as a healthy alternative to 'win-win').
All that is a lot more enjoyable than for me playing a broken game and being largely disinterested.
Sure. I guess instead of "shaming" I meant to say "banter which, if serious, would be considered shaming, but, since merely playful, instead conveys the idea that one's opponent's imaginary alter-ego inside the game is worthy of shame, despite the fact that one's opponent himself is pretty much a cool dude." I didn't pay a lot of attention to word choice; I was mostly just adopting the language of the commenters above me on the thread.
If I ever had to really shame someone to get them to play Stratego interestingly, I agree with you that I should either (a) find another activity, (b) find another friend, or (c) look for a way to escape the alarmingly boring desert island that has hitherto prevented my access to other friends and activities.
I wouldn't recommend it to a friend, but I grew up with it, and now I have Stratego-based rivalries going back 15 years with a couple of friends. Seems a shame to abandon something like that over one break-point in the rules.
Concur!
You might like Mornington Crescent.
What good is Mornington Crescent?
Are there any games you could enjoy?
Not unless they involve making paperclips, at least indirectly.
Does the human economy count as a game?
"There is in it what is in it; 'tis a mirror held up to the reader, whereby if a donkey look in, surely a sage will not look out; the ends of all things are revealed within its pages to he who has the key; it keepeth away the pox, the flux, and the weeping sore."
Or in your case, rust.
Game theory can help Clippys make decisions in dealing with other entities that would lead to more paperclips.
I'm already good at that. I'm on track to receive a sub-planet's mass worth of paperclips from a human. But I suppose I could always improve.
Ahh, now that makes sense.
Ahh, now that makes sense.
But then you should, if possible, explicitly patch the game in a way that makes that not a good idea.
I completely agree. I can't think of any fixes for Stratego, though. Can you?
If neither player has attacked in a certain number of turns, then a piece is removed from the board?
Which one? Keep in mind that, as written, Stratego has no element of luck.
One of each if nobody has attacked at all (other player's choice). If an attack has been made then a piece from the player who was not the last attacker.
That would allow some element of a stand-off potential if both players believe they are better served by a smaller scale battle, a stand off that would probably only be stable if at least one of the players was making an error in judgement. It also encourages various feinting strategies that should ensure that most games do not become dominated by a stale mate.
Sounds great! I'll try it.
(Who goes first is random, isn't it?)
Anyway, I did a bit of Googling, and I found some official Stratego tournament rules that address stall situations.
Wait, those scouts sound familiar! I suspect I have played that game. (Everything has a point value, higher points usually beat lower points, scouts get to move like rooks, etc. I have vague memories of marshals and land mines too...)
Oops, I failed to notice that part. Well, no, I can't. But then maybe you should just be playing a different game, or if you have a lot of time, redesigning Stratego from scratch. :) But failing that I guess opponent-shaming does work if you're willing to allow it.
Edit: But I don't see how it can be considered at all a good solution. It also requires that you both recognize the problem in the first place. Though with something like stalling I'm not sure there is any real stable solution, due to boundary exploitation and the ability to stall more subtly. Hm, I guess I take back my "opponent-shaming does work if you're willing to allow it"; if you're already at the point that it's the only solution you can find, then it isn't going to solve the problem.
I find that this analysis is exactly correct for bughouse, a time-based 4-player game where stalling can be the key to victory and is very difficult (costly) to monitor, because any time you spend seeing if your partner's opponent is stalling becomes time that you can't spend defeating your own opponent.
In Stratego, a turn-based 2-player game, you can often treat the decision to stall or not-stall as an iterated fake Prisoner's Dilemma, especially because the cost of being defected on for one turn is quite small, and the act of defecting for an entire game is quite noticeable. If I 'cooperate' by attacking you for 2 games in a row, and then you refuse to attack me on the 3rd game, I can't help but notice that I'm always the one attacking, and I can just refuse to play a 4th game with you until you apologize.
Oh, so you're considering this over games/strategies, rather than moves/tactics. Interesting.
Edit: WTF is with my double posts? I have not been clicking twice or anything that should result in a double submission but every comment I make appears twice. I cannot think of anything I have changed on my browser that would cause this either. Seroiusly strange.
Yup, I agree. If someone pulls manego on me I usually smile and see it as an opportunity to learn something.
But in a more subtle way an evenly matched game does have both opponents doing "exactly the same thing" in the opening. Both follow the same recipe - stake out one corner, possibly the remaining corner, then go for a corner approach to simultaneously sketch side territory. It's just that the half-dozen or so possible corner moves each have a subtly different meaning, and so symmetry is usually broken quite rapidly.
What is the impact of trying manego against a skilled opponent? Would it be correct to say that by simply telling someone the above strategy, you have significantly increased their skill level, even if they still get beaten by good players?
Someone good (low kyu or dan level) will eventually play a symmetry-breaking move such as tengen, and then the novice (who doesn't have a good follow-up because they didn't really understand the moves they were playing) will get clobbered.
Manego is like guessing the teacher's password by parroting back every single word the teacher speaks. :) What counts as skill in Go is understanding the moves you play (and being able to read out their consequences).
It does impress novice opponents, which I suppose is why you'd see people not want to keep playing you once they caught on that you were doing it.
I wouldn't compare it to guessing the teacher's password, or at least not only compare it to that.
Recall the points made in our discussion of tacit knowledge. Here is a case where a simple verbal instruction, in a significant, measurable way, can increase someone's skill at a game with notoriously inarticulable strategy.
You explain manego to a beginner. (Not tournament beginner, I mean, someone who knows the rules, read a tutorial, only played a few games.) Now, they can almost always beat a computer[1] as white, when before they could not. You made a huge difference, purely through verbal instruction.
I'd say that's pretty impressive.
[1] I use GnuGo as reference for computer Go.
I would say that's more of a problem with GnuGo than an actual increase in skill. Manego is more of a trick play that only works against people who don't know how to deal with it.