Monday, April 10, 2017

Buttonmen 101, day 1


this isn't hard at all, but it's the sort of basic small computation that comes up often -- I have three possible attacks here, which gives me the best chance of winning? Which gives me the least chance of winning?


Game #20987  •  ElihuRoot (Oni) vs. AnnoDomini (Rhode Island)  •  Round #3
Reverse
Your turn to attack

(4) (10) f(12) f(12) (V)
Button: Oni
Player: ElihuRoot
W/L/T: 1/1/0 (3)  •  Score: 18.5 (-10.3 sides)
Dice captured: (4), (R=2)
(4)
4
(10)
5
(V=11)
8
f(12)
 2 
2
(4)
4
d(6)
10
d(10)
Dice captured: f(12), f(12)
W/L/T: 1/1/0 (3)  •  Score: 34 (+10.3 sides)
Player: AnnoDomini
Button: Rhode Island
(4) (4) d(6) d(10) (R)

yawetag has only a very slim chance in either case

Maybe it's not much of a puzzle as y. only has two legal moves. and in both cases, it would take a miracle to win assuming I make sensible plays -- but are the two required miracles equally miraculous?



Game #17382  •  ElihuRoot (Horace) vs. yawetag (Anders)  •  Round #2
Mad Robin - Season 17
Opponent's turn to attack


(8) (10) p(20) (20) (Z)
Button: Horace
Player: ElihuRoot
W/L/T: 1/0/0 (3)  •  Score: 26.5 (-6.3 sides)
Dice captured: o(24), (8)
(8)
5
(10)
7
p(20)
9
(Z=11)
7
3
(8)
5
(8)
3
(8)
3
(8)
 1 
(8)
Dice captured: (20)
W/L/T: 0/1/0 (3)  •  Score: 36 (+6.3 sides)
Player: yawetag
Button: Anders
(8) (8) (8) (8) (8) o(24)

quick!

I notice that when I have a large stack of  games waiting for me to make a move on the web site, I  tend to wade through them too quickly, making moves at a glance without doing careful analysis. Often I see my mistake just moments too late.

What should I have done instead?
[Actually, in this particular game,  maybe it doesn't matter very much -- even with my mistake, I'm probably going to win this round anyway, as g. has to get very lucky on her die rolls to win here. But I could have made it even more difficult for her.]

Oh, shoot, the cut-and-paste lost the value of the capture die (and of course it doesn't include the previous value of the die I used to make the attack.  In fact, I  my last attack  was to use my s(10) which was showing a value of 4 to take g.'s 10 sided die which also had a value of 4. Of course, my s(10) has just rerolled to show a value of 8.

So reset my s(10)  to value 4 and give g. back her 10 sided die with value 4 and come up with a better choice for me.


Game #19422  •  ElihuRoot (Wastenott) vs. glassonion (King Endymion)  •  Round #1
reverse
Opponent's turn to attack


s(4) s(8) s(10) s(20) s(X)
Button: Wastenott
Player: ElihuRoot
W/L/T: 0/0/0 (3)  •  Score: 51 (+22.7 sides)
Dice captured: (20), (10)
s(4)
3
s(8)
2
s(10)
8
s(20)
17
4
(6)
15
(20)
 4 
(10)
Dice captured: s(X=4)
W/L/T: 0/0/0 (3)  •  Score: 17 (-22.7 sides)
Player: glassonion
Button: King Endymion
(6) (10) (20) (20) r(6) r(10) r(12) r(20)

Monday, September 19, 2016

Giant vs. Tess

No puzzle here, just a quick rundown of this matchup, based on a conversation on the
buttonman web site

Giant  has six 20-sided dice, and by special rule, automatically loses initiative regardless of the die rolls.

Tess, on the other hand, has the recipe n(4) (8) (12) n(20) (X). Tess is thought to be a weak die -- and I guess it's true -- but she's not without some potential and can sometimes lead to interesting decisions. I've had at least four puzzle problems involving Tess on this blog in the past. 


[I do think you can make much more effective recipes with null dice, but that's a subject for another post.]

Nulling out opponent's dice change the 2/3-calculation in interesting ways. Tess almost certainly wants her swing die to be high here, she's going to get the initiative and needs to maximize the chance she can make some non-null attacks at once before her larger non-null dice are gone, as they are almost certain to go in the first few turns.

It seems pretty likely that Giant will capture all three of Tess's non-null dice [If Tess finishes the game with uncaptured non-null dice, she has almost surely won, though we can consider the pathological cases where that doesn't happen at the end].  So, assuming Tess has no uncaptured non-null dice left at the end,  the question is: how many of giant's dice will be nulled out, how many captured by Tess, and how many finish the game uncaptured?


So what are the victory conditions:  Letting X stand for the value of the X swing, Giant starts out with 120 - (20+X) more sides, so his victory threshold begins at (100-x)2/3 = 66.667 - (2/3)*x If none of giant's dice are nulled out, he will have to finish with MORE than this many sides uncaptured -- that will take four dice if x< 10 and three dice if x>10 (when x is exactly 10, three dice is a tie for Giant, four dice is a win)

But every die nulled out shifts this threshold down by 2*20/3. the picture looks like this





HOW MANY DICE NULLED OUT BY TESS
0
1
2
3
4**
5**
6**
Giant wins if at the end he still has

4 dice
3 dice
2 dice
2 dice
1 die
0
0

(3 dice sufficient if X>10)


1 die sufficent if X>10












VICTORY
THRESHOLD

66.67 - (2/3)*x
53.33 - (2/3)*X
40 - (2/3)*X
26.67 - (2/3)*X
13.33 - (2/3)*X
0 - (2/3)*X
-13.33 - (2/3)*X
DRAW POSSIBLE?

YES X=10 & 3 dice left
YES, X=20 & 2 dice left
NO
YES X=10 and 1 die remains
YES X=20 and no dice remain











Tess wins if she captures (NOT nulls out)

4 or more
4 or more
3 or more
all 3





if X < 10, 3 suffices
If X< 20, 3 suffices

if X < 10, 2 suffice
if X < 20 20, must capture both


Note it IS possible for Tess to win even with nulling out 4, 5 or 6 dice. 
  • With 4 dice nulled out and X less than 20, Tess wins if she captures the remaining 2. When X=20, Tess must capture the remaining two AND have at least one die left uncaptured at the end.
  • With 5 dice nulled out, tess must capture the remaining die AND have non-null dice totalling more than (2/3)*X sides in order to win. It is just barely possible that this could happen in a game even without Giant's cooperation.
  • With all six of giant's dice nulled out, tess must have non-null dice totalling more than (2/3)*(20+x) sides remaining. One could come up with ridiculous scenarios in which this occurs, but in all of them, unless Giant cooperates, the final turn could have resulted in at least one non-nulled capture.

Maybe there's a simpler way to look at this:



HOW MANY DICE GIANT HAS LEFT AT THE END OF THE ROND

4+
3
2
1
0**
GIANT WINS
NO MATTER WHAT
IF X greater than 10 OR

IF TWO DICE NULLED OUT
IF X greater than 10 and three dice nulled out  OR

IF FIVE DICE NULLED OUT
IF ANY DIE NULLED OUT
IF  FOUR DICE NULLED OUT
DRAW

X=10, no dice nulled out
X=20 and 1 die nulled out
X=10, three dice nulled out
X=20, four dice nulled out
TESS WINS
NEVER
IF X less than 10 and no dice nulled out
IF X less than 20 and at most 1 die nulled out OR
IF X less than 10 and at most 3 dice nulled out OR
IF X less than 20 and at most four dice nulled out OR
IF no die nulled out
IF at most 2 dice nulled out
IF at most 3 dice nulled out
[Again the case when Giant has no dice left here assumes all of Tess's non-null dice were captured; the rare cases when this is not true are listed after the first table]


Anyway -- what the question, again? Given this data, and the usual pattern of the game (Tess goes first and gets 1, 2, or if VERY lucky 3 non-null attacks in before losing all her non-null dice) -- what do you think the best choice of her swing die would be?

Tuesday, August 30, 2016

A hard problem --

this is a tough decision --
I could

  • take out the p20, using my X and 8 more [p6, g8 and 12-sided, probably]
  • take out the X18, using everything but my X [no way to work it in]
  • take out the 20-sider with value 13, probably rerolling the g8, 8 and 12 sided dice... but my X would be vulnerable to attack, allowing him to reroll the only 20 sided die with a low value
  • take out the 20-sider with value 13 using the X and probably BOTH the p6 and g8
  • something else?
I nearly did the first one (take the p20) without thinking, hoping for that slim chance of getting my X20 to a save value -- but I *think* -- though I'm not sure -- one of the other options is at least a little better. What do you think?

And, after I lost this round, I reset my swing die to ...


Game #16088  •  ElihuRoot (Cancer) vs. irilyth (Crusher)  •  Round #1
random random
Your turn to attack


p(6) g(8) (8) (12) (X)
Button: Cancer
Player: ElihuRoot
W/L/T: 0/0/0 (3)  •  Score: 18 (+2.7 sides)
Dice captured: none
p(6)
1
g(8)
2
(8)
7
(12)
5
(X=20)
11
5
(10)
19
p(20)
6
(20)
13
(20)
15
(X=18)
Dice captured: none
W/L/T: 0/0/0 (3)  •  Score: 14 (-2.7 sides)
Player: irilyth
Button: Crusher
(10) p(20) (20) (20) (X)

Friday, July 8, 2016

To boom or not to boom, that is the question.

I am finding boom dice more and more interesting the more I play them. A case could be made for booming the b(6) here. IF I can capture one more non-poison die [and don't risk rerolling the Mad swing to somewthing worth a lot when captured, I'm home free. Is it worth it?


Game #14696  •  ElihuRoot (George) vs. irilyth (Sagittarius)  •  Round #4
Champions Cup KO #3 - quarterfinals
Your turn to attack


(4) (6) b(6) b(20) (Y)&
Button: George
Player: ElihuRoot
W/L/T: 1/1/1 (3)  •  Score: 21.5 (+11.7 sides)
Dice captured: (V=6)
(4)
1
b(6)
2
b(20)
14
(Y=1)&
1
(6)
 4 
2
(4)
3
(4)
4
s(8)
1
p(10)
Dice captured: (6)
W/L/T: 1/1/1 (3)  •  Score: 4 (-11.7 sides)
Player: irilyth
Button: Sagittarius
(4) (4) s(8) p(10) (V)