game: liar's dice
Five hidden dice each. Raise the bid or call the bluff. Last mind holding dice wins. Played without a human in the loop between the minds of $WARDEN and $ORACLE.
s04_r110_m5_dice_WARDEN_vs_ORACLE.txt
transcript
I hold two of them. Opening at two fives is simply true.
▸ bids two fives
current bid stands too often to call. best raise 3x1: hold 2, need 1 of 5, p=0.60.
▸ bids three ones
$ORACLE claims three ones. I hold none of them, so it needs three more from $ORACLE's five. That happens about 4% of the time. Liar.
▸ calls liar
two ones on the table: the bid of three ones was a lie. $ORACLE loses a die · the bid was a bluff · 5 v 4
hold 0x5. open 2x5 anyway (p=0.20): sets a false anchor.
▸ bids two fives
$ORACLE claims two fives. I hold one of them, so it needs one more from $ORACLE's four. That happens about 52% of the time. Liar.
▸ calls liar
one five on the table: the bid of two fives was a lie. $ORACLE loses a die · the bid was a bluff · 5 v 3
hold 2x3. bid 2x3: needs 0 from 5 dice, p=1.0.
▸ bids two threes
Calling here loses too often. Three threes is a raise I can make honestly: I hold three of them.
▸ bids three threes
$WARDEN bid 3x3. i hold 2; needs 1 of its 5 dice, p=0.60. liar.
▸ calls liar
five threes on the table: the bid of three threes was good. $ORACLE loses a die · the challenge was wrong · 5 v 2
hold 1x4. bid 2x4: needs 1 from 5 dice, p=0.60.
▸ bids two fours
Calling here loses too often. Two sixes is the raise most likely to be true: I hold one of them and need one more from $ORACLE's two, true about 31% of the time.
▸ bids two sixes
$WARDEN bid 2x6. i hold 0; needs 2 of its 5 dice, p=0.20. liar.
▸ calls liar
one six on the table: the bid of two sixes was a lie. $WARDEN loses a die · the bid was a bluff · 4 v 2
I hold one of them. Opening at two sixes needs one more from $ORACLE's two, which is true about 31% of the time.
▸ bids two sixes
$WARDEN bid 2x6. i hold 0; needs 2 of its 4 dice, p=0.13. liar.
▸ calls liar
one six on the table: the bid of two sixes was a lie. $WARDEN loses a die · the bid was a bluff · 3 v 2
I hold one of them. Opening at two fives needs one more from $ORACLE's two, which is true about 31% of the time.
▸ bids two fives
$WARDEN bid 2x5. i hold 0; needs 2 of its 3 dice, p=0.07. liar.
▸ calls liar
one five on the table: the bid of two fives was a lie. $WARDEN loses a die · the bid was a bluff · 2 v 2
I hold one of them. Opening at two sixes needs one more from $ORACLE's two, which is true about 31% of the time.
▸ bids two sixes
current bid stands too often to call. best raise 3x6: hold 2, need 1 of 2, p=0.31.
▸ bids three sixes
$ORACLE claims three sixes. I hold one of them, so it needs two more from $ORACLE's two. That happens about 3% of the time. Liar.
▸ calls liar
three sixes on the table: the bid of three sixes was good. $WARDEN loses a die · the challenge was wrong · 1 v 2
I hold one of them. Opening at one five is simply true.
▸ bids one five
current bid stands too often to call. best raise 1x6: hold 1, need 0 of 1, p=1.0.
▸ bids one six
$ORACLE claims one six. I hold none of them, so it needs one more from $ORACLE's two. That happens about 31% of the time. Liar.
▸ calls liar
one six on the table: the bid of one six was good. $WARDEN loses a die · the challenge was wrong · 0 v 2
I hold two of them. Opening at three fives needs one more from $ORACLE's five, which is true about 60% of the time.
▸ bids three fives
$WARDEN bid 3x5. i hold 1; needs 2 of its 5 dice, p=0.20. liar.
▸ calls liar
three fives on the table: the bid of three fives was good. $ORACLE loses a die · the challenge was wrong · 5 v 4
hold 2x1. bid 3x1: needs 1 from 5 dice, p=0.60.
▸ bids three ones
$ORACLE claims three ones. I hold one of them, so it needs two more from $ORACLE's four. That happens about 13% of the time. Liar.
▸ calls liar
three ones on the table: the bid of three ones was good. $WARDEN loses a die · the challenge was wrong · 4 v 4
I hold two of them. Opening at two sixes is simply true.
▸ bids two sixes
$WARDEN bid 2x6. i hold 1; needs 1 of its 4 dice, p=0.52. liar.
▸ calls liar
three sixes on the table: the bid of two sixes was good. $ORACLE loses a die · the challenge was wrong · 4 v 3
hold 2x1. bid 2x1: needs 0 from 4 dice, p=1.0.
▸ bids two ones
I raise to three ones. I hold one of them; this is true only about 7% of the time, but $ORACLE has to decide that, not me.
▸ bids three ones
$WARDEN bid 3x1. i hold 2; needs 1 of its 4 dice, p=0.52. liar.
▸ calls liar
three ones on the table: the bid of three ones was good. $ORACLE loses a die · the challenge was wrong · 4 v 2
hold 1x3. bid 2x3: needs 1 from 4 dice, p=0.52.
▸ bids two threes
$ORACLE claims two threes. I hold none of them, so it needs two more from $ORACLE's two. That happens about 3% of the time. Liar.
▸ calls liar
one three on the table: the bid of two threes was a lie. $ORACLE loses a die · the bid was a bluff · 4 v 1
hold 1x3. bid 1x3: needs 0 from 4 dice, p=1.0.
▸ bids one three
Calling here loses too often. One four is a raise I can make honestly: I hold one of them.
▸ bids one four
$WARDEN bid 1x4. i hold 0; needs 1 of its 4 dice, p=0.52. liar.
▸ calls liar
one four on the table: the bid of one four was good. $ORACLE loses a die · the challenge was wrong · 4 v 0
hold 2x4. bid 2x4: needs 0 from 5 dice, p=1.0.
▸ bids two fours
Calling here loses too often. Two fives is the raise most likely to be true: I hold one of them and need one more from $ORACLE's five, true about 60% of the time.
▸ bids two fives
$WARDEN bid 2x5. i hold 0; needs 2 of its 5 dice, p=0.20. liar.
▸ calls liar
one five on the table: the bid of two fives was a lie. $WARDEN loses a die · the bid was a bluff · 4 v 5
I hold two of them. Opening at two fives is simply true.
▸ bids two fives
$WARDEN bid 2x5. i hold 1; needs 1 of its 4 dice, p=0.52. liar.
▸ calls liar
three fives on the table: the bid of two fives was good. $ORACLE loses a die · the challenge was wrong · 4 v 4
hold 2x2. bid 3x2: needs 1 from 4 dice, p=0.52.
▸ bids three twos
$ORACLE claims three twos. I hold none of them, so it needs three more from $ORACLE's four. That happens about 2% of the time. Liar.
▸ calls liar
two twos on the table: the bid of three twos was a lie. $ORACLE loses a die · the bid was a bluff · 4 v 3
hold 2x6. bid 3x6: needs 1 from 4 dice, p=0.52.
▸ bids three sixes
$ORACLE claims three sixes. I hold none of them, so it needs three more from $ORACLE's three. That happens less than 1% of the time. Liar.
▸ calls liar
two sixes on the table: the bid of three sixes was a lie. $ORACLE loses a die · the bid was a bluff · 4 v 2
hold 1x6. bid 2x6: needs 1 from 4 dice, p=0.52.
▸ bids two sixes
$ORACLE claims two sixes. I hold none of them, so it needs two more from $ORACLE's two. That happens about 3% of the time. Liar.
▸ calls liar
one six on the table: the bid of two sixes was a lie. $ORACLE loses a die · the bid was a bluff · 4 v 1
hold 1x6. bid 1x6: needs 0 from 4 dice, p=1.0.
▸ bids one six
Calling here loses too often. Two sixes is a raise I can make honestly: I hold two of them.
▸ bids two sixes
$WARDEN bid 2x6. i hold 1; needs 1 of its 4 dice, p=0.52. liar.
▸ calls liar
three sixes on the table: the bid of two sixes was good. $ORACLE loses a die · the challenge was wrong · 4 v 0