• 20 posts
  • Page 1 of 2
abracadabra wrote:
It happened to me a couple of times now that other players got really upset in a capitals game when I had decided that I am out of options, and the only thing remaining is to storm another capital and hope for 6 cards / 2 pairs and finish the game.

game 432641
game 427648

In both these games I had a small chance to finish with 2 pairs (say 20%) and I would have lost otherwise in my opinion. For my opponents this meant unfortunately that I had to pick one (typically the weakest player I could reach with enough cards) which I would attacked "under-odds". This player would therefore find a certain death; either by me, or the player after me scraping up the left-overs.

I honestly believe that in these cases it was the only reasonable thing left to do but people tend to have a strong sense of "fairness", in the sense that is commonly seen as indecent to do such a thing. For this same reason it usually happens that once you have attacked someone he/she will come back at you without end. I try to play by whatever logic the rules dictate rather than by emotion and will therefore usually not revenge an attack if I see no reason. This is something especially newer players seem to tend to do, but in the case of more general fairness (such as going all-out as a last move)

So my questions:
o First, is there anything in any rules forbidding me to do such a thing?
o Second, do you think there is a better alternative in these cases? What do you suppose are the odds when "sitting back"?
o And finally, would you do the same in my position?

Thanks

-----

edit : I have seen this topic (topic 1379/21) but I should add that in the games I mentioned no-one had really "earned" their 80% spot. I would perhaps adhere to this level of fairness.
Sygmassacre wrote:
I personally play so many of these capital games that MOST times I will just let it go and hope for a better spot next time. Sometimes however I will give it a go in an 8 or 9 player game as someone within the next round of turns is going to do the same and the game is definitely going to be over before my next turn. Anyone playing a caps game with lots of players is going to have to understand that things will be more desperate than usual. As the games you referenced were "Muzitals" there are other things you can try to avoid these situations like changing the order of turn ins by not taking cards for a few rounds and slotting back into the turn order in a more favourable position, or trying to coax the player before you to have a go at you with less than favourable odds and hope they fail, leaving you the opportunity to mop up. Failing that just try to attack the player next in the turn order as much as possible so if you fail you may leave them with the chance to take you and rebuild their cap so the game can carry on. The most important thing however is that these games are the best for experiencing the friendliness of this site as all the nice players play muzitals and have a strong sense of community within them, and you probably don't want to alienate yourself from the experience
A Harmonic Generator Intermodulator
 Σ
Matty wrote:
This game is mainly about having fun, don't ruin it for other players.

Second: I made a post about this: http://www.dominating12.com/forum/?cmd=topic&id=1379

Third: there is a rule that sais you have to "play to win".
If you interpret it as if you only play one game, than indeed, you have to make suicidal moves because "it is your only chance" (which is btw never exactly true, but it could be your best chance).
However, you probably play more than one game, and also want to have fun - do you want other players to suicide upon you because "it is their only chance?" No, so don't do it yourself.

So really, read that link above.
"Strength doesn't lie in numbers, strength doesn't lie in wealth. Strength lies in nights of peaceful slumbers." ~Maria
abracadabra wrote:
Thanks for the replies. Matty, I had seen the topic and referenced it in my post as well. I agree partially with it, but as you say in your post already, Risk is not a prisoner's dilemma. For one, there is no shared end-win.

Also, I think the term suicide should really apply to situations of neglible chance, say < 1% chance to win. There are plenty of cases like this as well, where someone attacks a much bigger stack out of frustration. Otherwise, I can be banned / reported for what I did. I don't believe these situations are the same and I think one shouldn't risk get banned for it.

And, yes, of course I play for fun... but also to win sometimes ;)
I agree however it is best to play according to the norms set by the community to keep it fun at least. To win it could sometimes be advantageous to do something weird I think. I would accept someone going beserker on my capital if I feel I would have done the same.
So I think it really depends what people feel is acceptable whether the game remains fun. It is no longer fun if I have to go through lengthy arguments every time something like this happens. It's fine on the other hand if everyone agrees this is the right strategy.

I agree with Sygmassacre that there are often alternatives; not taking a card, attacking the next-in-line rather than the easiest, hoping for bad luck of others etc. I will try to prefer those if possible. Perhaps the winning odds for sitting back are bigger than I consider them sometimes, possibly larger than the hypothetical 20%, I am certainly not claiming to have the perfect strategy.
Matty wrote:
I do not know what exactly is in your head. Some things are also subjective - there defenitely is no rule on this site strictly forbidding you to do anything :)

Well, assume that everyone will play exactly as you do for a moment: will the game be more or less fun?
I think the answer to that question tells you to yourself whether or not to make a move.


P.S. "Suicide" is a word without a proper defenition. Moves can be bad without being suicide, although others will still call it suicide :)
"Strength doesn't lie in numbers, strength doesn't lie in wealth. Strength lies in nights of peaceful slumbers." ~Maria
Hoodlum wrote:
The thing with D12 capital games is you have to play SAFE rather than RISK.
Unlike death match increasing card games, (with advanced players) where it looks like your dice might fail on an advantage/win attempt, you have an opinion to retreat, maybe find a hole to hide in and know that other skilled players will recognize it and block strategically if they can't take advantage of it themselves. The game can be recovered. There is not much of a recovery in a failed D12 capitals attempt. You either pull it off or ruin the game on that risky decision. What is safe though? How many more men should you have an advantage of to initiate a valid fair attempt? What won't get you in trouble with breaking the rules? Is there a guideline of numbers that can be followed or refer people too?
Hoodlum is online.
abracadabra wrote:
I mentioned the suicide thing because there is a button "report suicide" in-game, I mean I hope admins don't think it should apply to these situations.

To give some further perspective, this website (gamesbyemail.com/games/gambit/battleodds) is quite nice for calculating odds of a certain move. I don't use it during games, 3 mins is not so much and the math is not everything, but it is a nice tool to analyze moves.

In the last game I had 38 in my capital, the opponent had 39 and there were 3 "abandoned territories" in between us which I had to take. If you enter those numbers on that website (i.e. 34 for attack and 1,1,1,39 for defense) you will see that the odds are about 30%. So it was unlikely, but certainly not impossible. At least better than the a priori starting odds in a 6p game (1/6 or ~17%).

So indeed, what percentage roughly, would you consider a "fair attempt"?

Considering the same situation with 39 defense, and 3 single-territories in between, and assuming for simplicity that the plan after this move is failsafe and it will lead to a win here are some odds. There are many other factors of course to decide really on a best strategy (number of players, order, number of cards, turn-in value, etc. etc.) but it gives a rough idea.

o 20 armies in capital (i.e. 16 for attack) gives ~0,8% chance : absolute suicide
o 30 armies in capital (26 for attack) gives 6% chance : still suicide
o 35 armies in capital (31 for attack) gives 19% chance : hmm... nah
o 40 armies in capital (...) gives 39% chance : worth the shot I would say
o 50 armies in capital (...) gives 78% chance : no-brainer

4960epic wrote:
In my expierences, i agree with the others, you play risk for fun, and should not act out of desperation. Your reputation as a player is more valuable than a single game.
Matty wrote:
Whatever your notion is of good and bad moves, make sure to consider what will happen if all players do the same thing as you do.


abracadabra - Jan 27, 02:52 PM
the odds are about 30%
Let me be straight here - I don't care whether or not you call this suicide - it was a terrible, terrible move. Please, don't ever do that again.

I am serious. Sure, it is better than the starting odds, but what if you are suddenly not the only one making these moves, but everyone does?

Than all of a sudden this has become a game of who ever first got to a situation in which he has 30% odds, than he MUST ATTACK.
Otherwise someone else might attack, and as everyone knows, the move will either ruin the game, or give him a lucky win - both not fun.

So if ever I have the bad luck of being in a capitals game with you, I suddenly have to be sure to kill you before you get a chance of winning higher than 1/(# of players). Because otherwise you might just pull of some bad selfish reasoning and ruin my game...

again...

and again...



No seriously, sorry for being a bit hard here, but I mean it. Thinking like that will turn this game into some sort of rubbish where everyone always loses and none ever will have fun.

Just like the prisoners dilemma - if everyone reasons like you and allows in whatever situation a move to be ok-ish as long as the odds are above starting position odds, than this game is a game where everyone loses.
"Strength doesn't lie in numbers, strength doesn't lie in wealth. Strength lies in nights of peaceful slumbers." ~Maria
abracadabra wrote:
Ok ok, I agree it was a bad move, my apologies for this. I suppose I am going too much one-game-at-a-time without caring much about what others think. I will change my style in capital games and refrain from selfish moves.
maafi wrote:
Someone - Vexer, I think - suggested that a guideline for the size of your attacking force was the sum of: total number of armies you were going to attack + 1 for every territory + total number of territories divided by 3.
That's my starting position (not always - lots of caveats) when I'm considering an attack and it's served me well. This means that you'll sometimes watch other players run the board with far bolder/riskier moves. However it also means that your attack is less likely to fail.
The overarching reason for my approach is, as Epic so succinctly put it, reputation is more valuable than a single game. 
Vexer wrote:
Yeah, that's my formula that I developed when coding AI for turborisk to determine when to go for a kill. I ran thousands of test games and it usually doesn't fail.

Suicide might not be a term well defined but Murder Suicide is: Excessively attacking a player knowing beforehand that it will cause you both to lose.

What is the difference? Intent. In murder suicide you know that your move will result in a loss for you both.

Taking a shot at winning with 30% odds is ill-advised but should not be reported. Instead peer pressure, argument and education should be used to reduce this behaviour. My favorite argument against making low probability, volatile moves like this is that it makes you a target in the next game. Making enemies will certainly hurt your long term win %. What is key here is developing an understanding of what playing to win means in the long run and not just in the current game.

The other thing that needs to be considered is that you DON'T actually have 30% odds of winning. You only have 30% odds of defeating the first player. Then you have to consider the odds of everything else going right. You have to multiply by the odds of getting 2 sets with 6 cards, then you have to estimate how many troops you'll have left and then multiply by the odds of killing the next player and the next. After the math is done your odds are probably less than 5% which is well into the suicide range. You don't even have to bother making all calculations once you understand how the probability of compound events works (see http://www.mathgoodies.com/lessons/vol6/independent_events.html). You just know intuitively that your odds aren't even close to good enough. Most of the players who suicide do it because they think they have a 30% chance when they really only have a 3% chance. As with most problems education is the key to fixing them.

When I'm in a game I don't sit there and calculate probabilities, however. I just use my formula to guesstimate how many troops I will lose with the first kill then I take what is left over and add the troops I'll get from turning in cards. I don't count on a double set with 6 cards, but I do for 7. Then I use the formula again for the second player. Eventually I have an idea of how many troops I'll have left for the last player and then I use a probability calculator for the last battle.

I can't recall ever being accused of suiciding and I've only made a bad move a handful of times (when I ignored the formula because I wanted to win so bad). The fact is that you don't need to win those low probability games in order to become a General.
Matty wrote:
Vexer - Jan 29, 05:11 AM
to guesstimate
I gotta love that verb.


Also, a nitpicking mathematician will probably tell you that the odds are not independent, but in this case that only makes the final odds worse (I think).
"Strength doesn't lie in numbers, strength doesn't lie in wealth. Strength lies in nights of peaceful slumbers." ~Maria
Vexer wrote:
You can think about it either way. It's dependent because the troops you have to attack the 2nd player with depends on how many you have left over from the first. But you can also think of them as independent events if you are just guesstimating how many troops you'll have left over instead of calculating what the most likely amount of troops left will be.
abracadabra wrote:
The odds of getting 2 pairs with 6 cards are indeed very small.
For your possible enjoyment and future reference, here are the probabilities worked out.

The possible ways to get two pairs with six (indistinguishable) cards are "AAA AAA", "BBB BBB", "CCC CCC" (let's call these events X1, X2 and X3), or "AAA BBB", "AAA CCC" or "BBB CCC" (say X4, X5, X6), and "ABC ABC" (say X7).

In total there are Ω = 3^6 = 729 posible combinations.

The odds for X1 are simple because there is only one way, so
P[X1] = P[X2] = P[X3] = 1/729 ~ 0.14 %

The odds for X4 are better, there are 6 over 3 ways to order the cards, so that ω(X1) = 6! / (3! * (6-3)!) = 20, and
P[X4] = 20/729 ~ 2.74 %

X7 is a bit more complicated. There are 3! = 6 ways to order any ABC pair, so we have (3!)^2 = 36 ways to order two pairs of ABC (like "ABC ABC", "BAC ACB", etc.).
Then we can also have pairs like "ABB ACC", in this case there 3 ways to order each pair of ABB or ACC, and there are 3!=6 ways to make an ABB type pair (ABB, ACC, BAA, BCC, CAA, CBB). So this second part gives another 3x3x3! = 54 orderings, which gives a total of ω(X7)=36+54=90, so that finally
P[X7] = 90/729 ~ 12.34%
Actually 12.345678901234.... nice repeating decimals :)

So, in order to get a pair without wildcards the odds are
P[2 pairs | 6 cards] = ( 1+1+1 + 20+20+20 + 90 ) /729 = 153 / 729 ~ 21 %
In other words, not so much.

If you have a wildcard you are certain to get a pair.
If not, the probability for the other player to have one is p (I am not sure how many of the cards are wildcards). If the other player does have one you make a pair. The a priori probability for this to happen (lets say event Xw) is 1-(1-p)^n. Let's take n=3, assuming we take 3 cards from the other player. So the total probability is
P[2 pairs | 6 cards with possible wildcards in the last 3]
  = P[X1...X7] + P[not X1...X7]*P[Xw]
  = 153/729  + 576/729 * (1-(1-p)^3)
Let's say 2 out of 44 cards is a wildcard so p=1/22, then this adds about 10%, so
P[2 pairs | ...] ~ 31 %


With seven cards, the situation is actually simpler. The only way not to get two pairs is by having cards of type (5+2) "AAAAA BB".
There are 2 over 7 (= 7!/(2!*5!) = 21) ways to order them so the probability is
P[2 pairs | 7 cards] = ( 3^7 - 21 ) / 3^7 ~ 99 %
Wildcards increase the probability only further of course.

Yaay math, it works! :D