-
In a dice game, six fair dice are rolled. Depending on their
faces or combinations of faces, points can be scored. I am trying
to find the probability of a successful roll that can yield points.
See Appendix A for a table of all the probabilities. Successful
Rolls include:
One(s) or five(s) are present
Three, four, five, or six of a kind
Full house (3 pair / four of a kind and a pair)
Long straight (1-6 are present)
Short straight (1-5 are present / 2-6 are present)
Six Dice / Full Roll Here is a Venn-diagram for what I believe
contains the correct probabilities. The colored regions are
represented by the equation the arrow points to. If I sum all the
probabilities together, the successful roll probability becomes 211
in 216 chance or 97.6851851851% chance.
One(s) or Five(s) are Present
Concept A successful roll may have one or more ones or fives. To
find this, I calculate the probability of a roll with no ones or
fives and find the complement.
Math There is a 4/6 chance of a not rolling a one or five. For
all six dice to not roll a one or a five, the probability would be
4/6 to the sixth power.
: (4
6)
6
Then subtract that number from one to finish the equation.
: 1 (4
6)
6
-
Statistical Reasoning Each die roll is independent of the each
other, and each face is always a potential outcome. Under these
circumstances, the probability of failure is p^n with p = 4/6 and n
= 6. To find the complement, or ones or fives probability, I
subtract that from one.
Three, Four, Five, or Six of a Kind
Concept I dont want to find the probability of all three, four,
five, or six of a kinds. The three, four, five, or six of a kind
with the ones and fives are already calculated in the one(s) or
five(s) are present section. Also, the rolls that have a three,
four, or five of a kind with the additional dice containing a one
or five are already calculated in the one(s) or five(s) are present
section. I only want to calculate the probability of the rolls that
contain twos, threes, fours, and sixes. To do this, I take the
probability of a three of a kind, four of a kind, five of a kind,
and six of a kind for one face value, and multiply it by a
combinational factor to account for the four possible face values.
However, when finding the probability of three of a kind through
this method, there is a chance that two three of a kinds occur.
This means that each three of a kind is counted twice. This needs
to be accounted for and removed.
Math The probability of the successful face is 1/6. For the
unsuccessful faces (one, five, or the successful face), the
probability is 3/6. For three of a kind, the probability is 1/6
(desired) raised to the third power and 3/6 (undesired) raised to
the third power. All is multiplied by a combinational factor.
: 63 (1
6)
3
(3
6)
3
For a four of a kind, the successful face is raised to the
fourth, and the unsuccessful faces are raised to the second. All is
still multiplied by a combinational factor.
: 64 (1
6)
4
(3
6)
2
For a five of a kind, the successful face is raised to the
fifth, and the unsuccessful faces are raised to the first. All is
still multiplied by a combinational factor.
: 65 (1
6)
5
(3
6)
1
For a six of a kind, 1/6 only the successful face is raised to
the sixth power (but the pattern still fits).
-
: 66 (1
6)
6
(3
6)
0
I sum all these together and multiply the sum by an additional
combinational factor. , , , :
41 ( 63 (1
6)
3
(3
6)
3
+ 64 (1
6)
4
(3
6)
2
+ 65 (1
6)
5
(3
6)
1
+ 66 (1
6)
6
(3
6)
0
)
For the two, simultaneous, three of a kind overlap probability,
the unsuccessful faces are only one face with probability 1/6. And
then that is multiplied by another combinational factor.
: 42 ( 63 (1
6)
3
(1
6)
3
)
I take this value and subtract it from the overall sum to reach
the final probability. , , , :
41 ( 63 (1
6)
3
(3
6)
3
+ 64 (1
6)
4
(3
6)
2
+ 65 (1
6)
5
(3
6)
1
+ 66 (1
6)
6
(3
6)
0
) 42 ( 63 (1
6)
3
(1
6)
3
)
Statistical Reasoning To find the three, four, five, or six of a
kinds probability, I need to consider that each die roll is
independent of the each other, and each face is always a potential
outcome. Because of this, the probabilities of each die roll needs
to be multiplied together. A certain amount of rolls must be of the
same die face value (probability p), and the rest of the rolls must
not be the same die face value, a one, or a five (probability q).
Total rolls (power of p + power of q) must equal six. Also, because
order doesnt matter, and there are various combinations of dice
that yield the same result, a combinational factor needs to be
included. For a three of a kind: p^3 * q^3, combinational factor
six dice, choose three to be the same face value. For a four of a
kind: p^4 *q^2, combinational factor six dice, choose four to be
the same face value. For five of a kind: p^5 * q^1, combinational
factor six dice, choose five to be the same face value. For a six
of a kind: p^6 * q^0, combinational factor six dice, choose six to
be the same face value. p = 1/6 and q = 3/6. There are four
potential face values that I care about (not ones or fives) to roll
a three, four, five, or six of a kind, and the math is based around
calculating one face value. This shows the need and determines the
values of the additional combinational factor four faces, choose
one of them. This was all done assuming the events were mutually
exclusive. When events are mutually exclusive, the events
probabilities can be summed together as was done. However, three of
a kinds are not mutually exclusive: Two, simultaneous three of a
kinds can occur. By doing the above process while including the
three of a kinds, the overlap (two, simultaneous three of a kinds)
needs to be removed.
-
To find the probability of the overlap, I need to apply a
similar method to finding the three of a kind probability. The
difference is that the unsuccessful faces (probability q) must all
be the same face, bringing down its value to 1/6. The combinational
factor, six dice choose three to be the same face value, is still
present. There are still four potential face values that I care
about (not ones or fives) to roll a three of a kind, and the math
is based around calculating two face values. This shows the need
and determines the values of the additional combinational factor
four faces, choose two of them. After finding the overlap
probability, I can subtract it from the previous three, four, five,
of six of a kind probability, thus accounting for the overlap.
Full House
Concept I dont want to find the total probability of all full
houses. I only want to calculate the probability of the rolls that
contain twos, threes, fours, and sixes. The ones that contain ones
and fives are already calculated in the one(s) or five(s) present
section. In addition, the full houses that contain a four of a kind
and a pair are calculated in the three, four, five, or six of a
kind section. To find this limited full house probability, I break
down the probabilities of three two of a kinds while multiplying by
various combinational factors.
Math Using the same pattern from the three, four, five, or six
of a kind section, the two of a kind probability can be found.
: 62 (1
6)
2
(3
6)
4
However, the remaining four unsuccessful face rolls need to be
broken down into another two of a kind probability with the
unsuccessful faces not including ones, fives, old two of a kind
faces, or the new two of a kind faces. This gets multiplied by
another combinational factor. This yields the two simultaneous two
of a kinds probability.
: 62 (1
6)
2
42 (1
6)
2
(2
6)
2
Because I am calculating a full house, the last unsuccessful
faces need to be the same, bringing its probability down to 1/6.
All of this is finally multiplied by another combinational
factor.
: 43 ( 62 (1
6)
2
42 (1
6)
2
(1
6)
2
)
-
Statistical Reasoning To find the full houses probability, I
need to consider that each die roll is independent of the each
other, and each face is always a potential outcome. Because of
this, the probabilities of each die roll needs to be multiplied
together. I use the same pattern from the three, four, five, or six
of a kind section to find a two of a kind probability. For a two of
a kind: p^2 * q^4, combinational factor six dice, choose two to be
the same face value. p = 1/6 and q = 3/6. I breakdown the last four
rolls represented by q^4 into another two of a kind probability
with the unsuccessful faces not including ones, fives, old two of a
kind faces, or the new two of a kind faces. For two simultaneous
two of a kinds: p1^2 * p2^2 *q^2, combinational factors six dice,
choose two to be the same face value, and four dice, choose two to
be the same face value. p1 = 1/6, p2 = 1/6, and q = 2/6. To find
the probability of a full house, the unsuccessful faces
(probability q) of the two simultaneous two of a kinds must all be
the same face, bringing the down its value to 1/6. The
combinational factors are remain the same. There are four potential
face values that I care about (not ones or fives) to roll a full
house, and the math is based around calculating three face values.
This shows the need and determines the values of the additional
combinational factor four faces, choose three of them.
Long/Short straight Because the long and both short straights
must contain a five, its probability is contained within the one(s)
or five(s) present section.
Five Dice Remain Now I want to check the probabilities if there
are five dice that remain. With five dice, it is impossible to get
a six of a kind, a full house, or a long straight. If I sum all of
the probabilities together, the successful roll probability becomes
299 in 324 chance or 92.283950617% chance.
One(s) or Five(s) are Present
Concept (See previous section.) The only difference is that
there are five dice instead of six.
-
Math (See previous section.) The only difference is that there
are five dice instead of six.
: (4
6)
5
: 1 (4
6)
5
Statistical Reasoning (See previous section.) The only
difference is that there are five dice instead of six (n = 5).
Three, Four, or Five of a Kind
Concept (See previous section.) Other than the inability to roll
a six of a kind, the only difference is that it is impossible to
roll two simultaneous three of a kinds. This means that the three,
four, and five of a kinds are mutually exclusive.
Math (See previous section.) The differences are that there is
no six of a kind section, and there is no overlap.
: 53 (1
6)
3
(3
6)
2
: 54 (1
6)
4
(3
6)
1
: 55 (1
6)
5
(3
6)
0
, , : 41 ( 53 (1
6)
3
(3
6)
2
+ 54 (1
6)
4
(3
6)
1
+ 55 (1
6)
5
(3
6)
0
)
-
Statistical Reasoning (See previous section.) Other than the
inability to roll a six of a kind, the difference is that it is
impossible to roll two simultaneous three of a kinds. This means
that the three, four, and five of a kinds are mutually exclusive.
Also, the power of p + power of q must equal five. In addition, the
intermediate combinational factors change. For a three of a kind:
p^3 * q^2, combinational factor five dice, choose three to be the
same face value. For a four of a kind: p^4 *q^1, combinational
factor five dice, choose four to be the same face value. For five
of a kind: p^5 * q^0, combinational factor five dice, choose five
to be the same face value. p = 1/6 and q = 3/6.
Short straight Because the short straights must contain a five,
its probability is contained within the one(s) or five(s) present
section.
Four Dice Remain Now I want to check the probabilities if there
are four dice that remain. With four dice, it is now impossible to
get a five of a kind or a short straight. If I sum all of the
probabilities together, the successful roll probability becomes 91
in 108 chance or 84.259259259% chance.
One(s) or Five(s) are Present
Concept (See previous section.) The only difference is that
there are four dice instead of five.
Math (See previous section.) The only difference is that there
are four dice instead of five.
: (4
6)
4
: 1 (4
6)
4
-
Statistical Reasoning (See previous section.) The only
difference is that there are four dice instead of five (n = 4).
Three or Four of a Kind
Concept (See previous section.) The only difference is that it
is impossible to roll a five of a kind.
Math (See previous section.) The only difference is that there
is no five of a kind section.
: 43 (1
6)
3
(3
6)
1
: 44 (1
6)
4
(3
6)
0
: 41 ( 43 (1
6)
3
(3
6)
1
+ 44 (1
6)
4
(3
6)
0
)
Statistical Reasoning (See previous section.) The difference is
that it is impossible to roll a five of a kind. Also, the power of
p + power of q must equal four. In addition, the intermediate
combinational factors change. For a three of a kind: p^3 * q^1,
combinational factor four dice, choose three to be the same face
value. For a four of a kind: p^4 *q^0, combinational factor four
dice, choose four to be the same face value. p = 1/6 and q =
3/6.
Three Dice Remain Now I want to check the probabilities if there
are three dice that remain. With three dice, it is now impossible
to get a four of a kind. However, the rules state that if there is
a three, four, or five of a kind rolled with dice remaining, then
that three, four, or five of a kinds face behaves as if it were a
one or a five (though no points are awarded). If I sum all of the
probabilities together, the successful roll probability becomes 13
in 18 chance or 72.22222% chance.
-
If a three of a kind was previously rolled and its face was not
a one or a five, then the successful roll probability becomes 8 in
9 chance or 88.88888% chance.
One(s), Five(s), or Three of a Kind Face(s) are Present
Concept (See previous section.) The difference is that there are
three dice instead of four. Also, when accounting for the potential
extra, three of a kind face, the successful dice probability
changes.
Math (See previous section.) The difference is that there are
three dice instead of four. In addition, if a three of a kind (not
of a one or five) was previously rolled, there is a 3/6 chance of
not rolling a one, five, or the three of a kinds face.
: (4
6)
3
, , : (3
6)
3
: 1 (4
6)
3
, , : 1 (3
6)
3
Statistical Reasoning (See previous section.) The difference is
that there are three dice instead of four (n = 3). Also, if a three
of a kind (not of a one or five) was previously rolled, p =
3/6.
Three of a Kind
Concept (See previous section.) The difference is that it is
impossible to roll a four of a kind. Also, if a three of a kind was
rolled (not of a one or five), then I dont want to find the
probability of all three of a kinds. The three of a kinds with the
ones, fives, and three of a kinds face are already calculated in
the one(s), five(s), or three of a kind faces are present section.
I only want to calculate the probability of the rolls that contain
values that arent ones, fives, or the value of the three of a kinds
face.
-
Math (See previous section.) The difference is that there is no
four of a kind section. Also, if a three of a kind was rolled (not
of a one or five), the probability of the unsuccessful faces (one,
five, three of a kind value, or the successful face) is 2/6. (This
number is actually unimportant. It has no statistical use. It only
exists for continuing the patterns.) Also, this causes
combinational factors changes.
: 33 (1
6)
3
(3
6)
0
( ) :
33 (1
6)
3
(2
6)
0
: 41 ( 33 (1
6)
3
(3
6)
0
)
( ) : 31 ( 33 (1
6)
3
(2
6)
0
)
Statistical Reasoning (See previous section.) The difference is
that it is impossible to roll a four of a kind. Also, the power of
p + power of q must equal three. In addition, the intermediate
combinational factors change. Also, if a three of a kind was rolled
(not of a one or five), the probabilities and the outer
combinational factor changes. There is one less possible face to be
considered unsuccessful (probability q), and thus becomes 2/6. For
a three of a kind: p^3 * q^0, combinational factor three dice,
choose three to be the same face value. p = 1/6 and q = 3/6. For a
three of a kind with a previous (not one or five) three of a kind:
p^3 * q^0, combinational factor three dice, choose three to be the
same face value. p = 1/6 and q = 2/6 If a three of a kind was
rolled (not of a one or five), there becomes only three potential
face values that I care about (not ones, fives, or the value of the
three of a kinds face) to roll a three of a kind, and the math is
based around calculating one face value. This continues to show the
need and determines the new values of the additional combinational
factor three faces, choose one of them.
Two Dice Remain Now I want to check the probabilities if there
are two dice that remain. With two dice, it is now impossible to
get a three of a kind. If I sum all of the probabilities together,
the successful roll probability becomes 5 in 9 chance or 55.55555%
chance.
-
If a three or four of a kind was previously rolled and its face
was not a one or a five, then the successful roll probability
becomes 3 in 4 chance or 75% chance.
One(s), Five(s), Three, or Four of a Kind Face(s) are
Present
Concept (See previous section.) The difference is that there are
two dice instead of three. Also, when accounting for the four of a
kind face, it acts no differently than if a previous three of a
kind was rolled.
Math (See previous section.) The difference is that there are
two dice instead of three.
: (4
6)
2
, , , : (3
6)
2
: 1 (4
6)
2
, , , : 1 (3
6)
2
Statistical Reasoning (See previous section.) The difference is
that there are two dice instead of three (n = 2).
One Die Remains Now I want to check the probabilities if there
is one die that remains. If I sum all of the probabilities
together, the successful roll probability becomes 1 in 3 chance or
33.33333% chance. If a three, four, or five of a kind was
previously rolled and its face was not a one or a five, then the
successful roll probability becomes 1 in 2 chance or 50%
chance.
-
One(s), Five(s), Three, Four, or Five of a Kind Face(s) are
Present
Concept (See previous section.) The difference is that there is
one die instead of two dice. Also, when accounting for the five of
a kind face, it acts no differently than if a previous three of a
kind was rolled.
Math (See previous section.) The difference is that there is one
die instead of two dice.
: (4
6)
1
, , , : (3
6)
1
: 1 (4
6)
1
, , , : 1 (3
6)
1
Statistical Reasoning (See previous section.) The difference is
that there is one die instead of two (n = 1).
-
Appendix A Tables of Probabilities Six Dice / Full Roll
Successful Roll Failing Roll 97.6851851851% 2.314814814%
Five Dice Remain Successful Roll Failing Roll 92.283950617%
7.716049382%
Four Dice Remain
Successful Roll Failing Roll 84.259259259% 15.740740740%
Three Dice Remain Successful Roll Failing Roll 72.22222%
27.77777% 88.88888%* 11.11111%*
Two Dice Remain Successful Roll Failing Roll 55.55555% 44.44444%
75%* 25%*
One Die Remains Successful Roll Failing Roll 33.33333% 66.6666%
50%* 50%*
* A previous roll contains a three, four, or five of a kind (not
of a one or a five).