• Master Mind from Journeyman project

    From Anton Shepelev@21:1/5 to All on Sun Mar 7 21:02:40 2021
    Hello, all

    I am stuck at the Master-Mind-like puzzle from Journeyman
    Project, where the player must find a sequence of three
    *different* colors out of five available, in five attempts,
    so that there are 60 combinations to choose from. In re-
    sponse to each guess, he receives the number of matching
    items, which have the right color at the right position. To
    complicate matters, the puzzle has to be solved within a
    very short time limit.

    Has anybody an idea how to make it honestly, without the as-
    sistance of a computer program that keeps track of exlcuded
    combinations and suggests a next attempt that efficiently
    narrows-down the remaning possiblities? I can't find a
    strategy simple enough to employ in mind & memory.

    --
    () ascii ribbon campaign -- against html e-mail
    /\ http://preview.tinyurl.com/qcy6mjc [archived]

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From leflynn@21:1/5 to Anton Shepelev on Wed Mar 10 16:57:12 2021
    On Sunday, March 7, 2021 at 1:02:42 PM UTC-5, Anton Shepelev wrote:
    Hello, all

    I am stuck at the Master-Mind-like puzzle from Journeyman
    Project, where the player must find a sequence of three
    *different* colors out of five available, in five attempts,
    so that there are 60 combinations to choose from. In re-
    sponse to each guess, he receives the number of matching
    items, which have the right color at the right position. To
    complicate matters, the puzzle has to be solved within a
    very short time limit.

    Has anybody an idea how to make it honestly, without the as-
    sistance of a computer program that keeps track of exlcuded
    combinations and suggests a next attempt that efficiently
    narrows-down the remaning possiblities? I can't find a
    strategy simple enough to employ in mind & memory.

    --
    () ascii ribbon campaign -- against html e-mail
    /\ http://preview.tinyurl.com/qcy6mjc [archived]

    Anton,
    Let’s start with five colors represented by a, b, c, d and e.
    There are 10 possible choices for the three colors not counting positioning.
    I would list the ten cases and then cross them off when they are eliminated.

    How can we find the colors? (I’ll talk about position strategies in another post.)
    If we give some version of (abc) for Guess 1,
    Case 1: three correct colors - the colors are (abc).
    Case 2: one correct color - the colors are (ade), (bde) or (cde).
    Case 3: two correct colors - the colors are
    (abd), (abe), (acd), (ace), (bcd) or (bce).

    For Case 1, you have the colors and one set of position information. I think you know how solve this one for positions too.

    For Case 2, you have three possible choices for the colors, you should know which one of the three it is after the third Guess. (Guess 2 (ade) if it is not (ade) then Guess 3 (bde) if it not (bde) then it is (cde).)

    For Case 3, you have six possible choices for the colors. Guess 2 (abd)
    Case 4: three correct colors – (abd) is the answer
    Case 5: one correct color – the answer is (ace) or (bce). (Guess 3 (ace) if it is not then it is (bce))
    Case 6: two correct colors – the answer is (abe), (acd) or (bcd). (Guess 3 (abe) if not then Guess 4 (acd) if not then (bcd).

    So in every case we can get identify the three colors within five guesses.

    L. Flynn

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Anton Shepelev@21:1/5 to All on Fri Mar 12 14:00:56 2021
    leflynn:

    Lets start with five colors represented by a, b, c, d and
    e. There are 10 possible choices for the three colors not
    counting positioning. I would list the ten cases and then
    cross them off when they are eliminated.

    Yes, finding a combination instead of a permutation is much
    easier. I still, however, am at a loss about solving the
    original problem as fast as the game requires.

    --
    () ascii ribbon campaign - against html e-mail
    /\ http://preview.tinyurl.com/qcy6mjc [archived]

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From leflynn@21:1/5 to Anton Shepelev on Sun Mar 14 14:51:30 2021
    On Friday, March 12, 2021 at 6:01:00 AM UTC-5, Anton Shepelev wrote:
    leflynn:

    LetТs start with five colors represented by a, b, c, d and
    e. There are 10 possible choices for the three colors not
    counting positioning. I would list the ten cases and then
    cross them off when they are eliminated.
    Yes, finding a combination instead of a permutation is much
    easier. I still, however, am at a loss about solving the
    original problem as fast as the game requires.

    --
    () ascii ribbon campaign - against html e-mail
    /\ http://preview.tinyurl.com/qcy6mjc [archived]

    I see finding a combination as part of the problem fortunately, you can do both at once.

    Here is a decision table. You'll need to view it with a fixed width font.
    G1 G2 G3 G4 G5
    (ABC)
    3p0 (BCA)(CAB)
    3p1 (ACB)(BAC)(CBA)
    2p0 (BAD)
    3p1 (BDA)(DAB)
    2p0 (CDA)
    3p0 (DCA)
    2p0 (DCB)
    2p2 (CDB)
    2p1 (BEA) (EAB)
    2p2 (BAE)(BCD)(CAD)
    1p0 (CEA)
    3p1 (ECA)
    2p0 (ECB)
    2p2 (CEB)
    1p1 (CAE)(BCE)
    2p1 (ADB)
    3p0 (DBA)
    2p0 (CBD)
    3p0 (BDC)
    1p0 (DAC)
    1p1 (EBA)
    2p1 (ACD)
    2p2 (AEB)
    1p0 (BEC)
    3p0 (CBE)
    2p1 (EAC)
    1p1 (ACE)
    2p2 (ABD)
    2p2 (ABE)
    2p1 (ADC)(DBC)
    1p1 (AEC)(EBC)
    1p0 (DAE)
    3p0 (EDA)
    3p1 (DEA)(EAD)
    2p0 (BED)
    3p0 (EDB)
    3p1 (DEB)
    2p0 (ECD)
    2p2 (CED)
    2p1 (BDE)(CDE)
    2p2 (DCE)
    1p1 (ADE)
    3p1 (AED)
    2p0 (DEC)(EBD)
    2p1 (DBE)(EDC)

    Is (DCE) in the correct place?
    Guess1(ABC) 1p0; Guess2(DAE) 2p2; Guess3 (DCE)
    Is (CAD) in the correct place?
    Guess1(ABC) 2p0; Guess2(BAD) 2p2; Guess3(BAE) 1P; Guess4(BCD) 2p1; Guess5(CAD)

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From leflynn@21:1/5 to leflynn on Sun Mar 14 15:39:36 2021
    On Sunday, March 14, 2021 at 5:51:31 PM UTC-4, leflynn wrote:
    On Friday, March 12, 2021 at 6:01:00 AM UTC-5, Anton Shepelev wrote:
    leflynn:

    LetТs start with five colors represented by a, b, c, d and
    e. There are 10 possible choices for the three colors not
    counting positioning. I would list the ten cases and then
    cross them off when they are eliminated.
    Yes, finding a combination instead of a permutation is much
    easier. I still, however, am at a loss about solving the
    original problem as fast as the game requires.

    --
    () ascii ribbon campaign - against html e-mail
    /\ http://preview.tinyurl.com/qcy6mjc [archived]

    I see finding a combination as part of the problem fortunately, you can do both at once.
    When I play mastermind I keep a grid of the possible locations for the remaining colors.
    It would start as
    a a a
    b b b
    c c c
    d d d
    e e e
    Rows would be removed or crossed out as colors were eliminated and x’s would be placed in positons that are not allowed. Some obvious results:
    Guess results with one correct color either tell us where the three colors present in the guess may go or where they cannot go. For example if Guess 1 (abc) had one correct color and it was not in the correct positon, the position chart would become:
    x a a
    b x b
    c c x
    d d d
    e e e

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From leflynn@21:1/5 to leflynn on Sun Mar 14 16:05:36 2021
    On Sunday, March 14, 2021 at 5:51:31 PM UTC-4, leflynn wrote:
    On Friday, March 12, 2021 at 6:01:00 AM UTC-5, Anton Shepelev wrote:
    leflynn:

    LetТs start with five colors represented by a, b, c, d and
    e. There are 10 possible choices for the three colors not
    counting positioning. I would list the ten cases and then
    cross them off when they are eliminated.
    Yes, finding a combination instead of a permutation is much
    easier. I still, however, am at a loss about solving the
    original problem as fast as the game requires.

    --
    () ascii ribbon campaign - against html e-mail
    /\ http://preview.tinyurl.com/qcy6mjc [archived]
    I see finding a combination as part of the problem fortunately, you can do both at once.

    Here is a decision table. You'll need to view it with a fixed width font.
    Well, that post turned into a mess...
    _G1___G2___G3___G4___G5
    (ABC)_
    _3p0_(BCA)(CAB)
    _3p1_(ACB)(BAC)(CBA)
    _2p0_(BAD)
    ______3p1_(BDA)(DAB)
    ______2p0_(CDA)
    ___________3p0_(DCA)
    ___________2p0_(DCB)
    ___________2p2_(CDB)
    ______2p1_(BEA)
    ___________3p0_(EAB)
    ______2p2_(BAE)(BCD)(CAD)
    ______1p0_(CEA)_
    ___________3p1_(ECA)
    ___________2p0_(ECB)
    ___________2p2_(CEB)
    ______1p1_(CAE)(BCE)
    _2p1_(ADB)
    ______3p0_(DBA)
    ______2p0_(CBD)
    ___________3p0_(BDC)
    ___________1p0_(DAC)
    ___________1p1_(EBA)_
    ______2p1_(ACD)
    ______2p2_(AEB)
    ______1p0_(BEC)
    ___________3p0_(CBE)
    ___________2p1_(EAC)
    ______1p1_(ACE)
    _2p2_(ABD)
    ______2p2_(ABE)
    ______2p1_(ADC)(DBC)
    ______1p1_(AEC)(EBC)
    _1p0_(DAE)
    ______3p0_(EDA)
    ______3p1_(DEA)(EAD)
    ______2p0_(BED)
    ___________3p0_(EDB)
    ___________3p1_(DEB)
    ___________2p0_(ECD)
    ___________2p2_(CED)
    ______2p1_(BDE)(CDE)
    ______2p2_(DCE)
    _1p1_(ADE)_
    ______3p1_(AED)
    ______2p0_(DEC)(EBD)
    ______2p1_(DBE)(EDC)

    Is (DCE) in the correct place?
    Guess1(ABC) 1p0; Guess2(DAE) 2p2; Guess3 (DCE)
    Is (CAD) in the correct place?
    Guess1(ABC) 2p0; Guess2(BAD) 2p2; Guess3(BAE) 1P; Guess4(BCD) 2p1; Guess5(CAD)

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Anton Shepelev@21:1/5 to All on Mon Mar 15 14:41:17 2021
    leflynn:

    Well, that post turned into a mess...

    Not at all: http://al.howardknight.net/?ID=161580839200
    but I still wonder why anyone who knows about monospace
    fonts should access Usenet via crippled GoogleGropus...

    I have not yet studied your solution yet, but will do.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From leflynn@21:1/5 to Anton Shepelev on Mon Mar 15 09:51:09 2021
    On Monday, March 15, 2021 at 7:41:19 AM UTC-4, Anton Shepelev wrote:
    leflynn:
    Well, that post turned into a mess...
    Not at all: http://al.howardknight.net/?ID=161580839200
    but I still wonder why anyone who knows about monospace
    fonts should access Usenet via crippled GoogleGropus...

    I have not yet studied your solution yet, but will do.

    Anton,
    The solution I gave picks from the available responses at each stage in alphabetical order. It’s not the way I would play a real game. When I play Mastermind, I keep piles of colors with information in my head.

    For colors here, I start with A,B,C.
    If all three are right, then good.
    If one is right, then D and E are colors and one of A,B,C. My round two guesses for these cases use A,D,E.
    (If your guess is always three distinct colors, then one color right always tells you that the other two are good colors. Once you know two for sure, every guess either tells you that one of the other three colors is right or wrong.)
    If two are right, then I would have a pile of A,B,C and remember that my answer needs two of them and a pile of D,E and I need one of them.
    My round two guess for these cases use A,B,D for colors.

    For round three
    If your A,B,D guess has one color right, then you know C and E and are correct and one A and B. I chose to check A,C,E first
    If you’re A,B,D guess has two colors right, then you either need to swap A or B for C or you need to swap D for E. I chose to swap B for C and guess using A,C,D. You can tell right away by whether number of correct colors stayed the same, decreased
    or increased whether the B -> C swapped good for good (B,C is right and so is D), good for bad (A,B is right and so is E) or bad for good (A,C is right and so is D).

    For positions, I just choose the first one alphabetically to go in the earliest correct position or not go in the earliest wrong position working through to the current guess, and then move on to the next color.

    L. Flynn

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Anton Shepelev@21:1/5 to sitions. I on Tue Mar 16 20:55:03 2021
    leflynn:

    Here is a decision table.

    If I understand your notation, xpy encodes a guess with x
    correct colors and y colors at the correct position. I fear
    it is not the feedback the player receives in the puzzle I
    posted. Instead of two numbers, the machine answers to each
    guess with a single number -- the quanitity of colors at
    correct positions. It does not reveal how many colors in the
    guess are present in the sought combination at incorrect po-
    sitions. I wrote in my original post:

    In response to each guess, he receives the number of
    matching items, which have the right color at the right
    position.

    I am sorry that my English was not clear enough. Perhaps a
    screenshot would help:

    https://freeshell.de/~antonius/img_host/colorpuzzle.png

    The five avaialbe colors are present above label that reads
    "possible color synapse nodes". The guess log is on the
    right, with a single number annotating each of the four
    spend guesses: 0,1,0,1. This is not a binary code, but the
    nubmers of matching color-positions.

    --
    () ascii ribbon campaign - against html e-mail
    /\ http://preview.tinyurl.com/qcy6mjc [archived]

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From leflynn@21:1/5 to Anton Shepelev on Tue Mar 16 11:20:36 2021
    On Tuesday, March 16, 2021 at 1:55:06 PM UTC-4, Anton Shepelev wrote:
    leflynn:
    Here is a decision table.
    If I understand your notation, xpy encodes a guess with x
    correct colors and y colors at the correct position. I fear
    it is not the feedback the player receives in the puzzle I
    posted. Instead of two numbers, the machine answers to each
    guess with a single number -- the quanitity of colors at
    correct positions. It does not reveal how many colors in the
    guess are present in the sought combination at incorrect po-
    sitions. I wrote in my original post:

    In response to each guess, he receives the number of
    matching items, which have the right color at the right
    position.
    I am sorry that my English was not clear enough. Perhaps a
    screenshot would help:

    https://freeshell.de/~antonius/img_host/colorpuzzle.png

    The five avaialbe colors are present above label that reads
    "possible color synapse nodes". The guess log is on the
    right, with a single number annotating each of the four
    spend guesses: 0,1,0,1. This is not a binary code, but the
    nubmers of matching color-positions.
    --
    () ascii ribbon campaign - against html e-mail
    /\ http://preview.tinyurl.com/qcy6mjc [archived]
    Ouch. Sorry, I've been solving the Mastermind problem, not yours.
    L. Flynn

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Lawrence E Flynn - NOAA Federal@21:1/5 to Anton Shepelev on Wed Mar 17 13:30:58 2021
    On Sunday, March 7, 2021 at 1:02:42 PM UTC-5, Anton Shepelev wrote:
    Hello, all

    I am stuck at the Master-Mind-like puzzle from Journeyman
    Project, where the player must find a sequence of three
    *different* colors out of five available, in five attempts,
    so that there are 60 combinations to choose from. In re-
    sponse to each guess, he receives the number of matching
    items, which have the right color at the right position. To
    complicate matters, the puzzle has to be solved within a
    very short time limit.

    Has anybody an idea how to make it honestly, without the as-
    sistance of a computer program that keeps track of exlcuded
    combinations and suggests a next attempt that efficiently
    narrows-down the remaning possiblities? I can't find a
    strategy simple enough to employ in mind & memory.

    --
    () ascii ribbon campaign -- against html e-mail
    /\ http://preview.tinyurl.com/qcy6mjc [archived]
    I don't have one I would want to memorize but I have a table that wins 51/60 times for an individual puzzle.
    It's essentially a printout of a computer program, so it doesn't even meet that limiting part of your request.
    I see some patterns and choices but by the third round it's too messy.
    G1 G2 G3 G4 G5
    123
    _0_231
    ____0_314
    _______0_542
    __________1_452
    _______1_342
    __________0_415
    __________1_412 x512
    __________2_345 x352
    _______2_315
    __________1_354 x514
    __________2_312
    ____1_541
    _______0_215
    __________0_432
    __________1_254 x435
    __________2_214
    _______1_531 x (only non-feasible guess)
    __________0_245
    __________1_351 x451
    __________2_532 x534
    _______2_341
    ____2_531
    _______1_234
    __________1_241 x251
    __________2_235
    _______2_431
    _1_543
    ____0_425
    _______0_134
    __________1_152
    __________2_132 x154
    _______1_321
    __________0_135
    __________2_324
    _______2_325
    __________1_421
    ____1_253
    _______0_524
    __________0_142 x145
    __________2_521
    _______1_413
    _______2_213
    __________1_453
    ____2_513
    _______1_243
    _2_124
    ____1_523
    _______1_143
    __________2_153
    _______2_423
    ____2_125

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From leflynn@21:1/5 to Anton Shepelev on Wed Mar 17 15:46:06 2021
    On Sunday, March 7, 2021 at 1:02:42 PM UTC-5, Anton Shepelev wrote:
    Hello, all

    I am stuck at the Master-Mind-like puzzle from Journeyman
    Project, where the player must find a sequence of three
    *different* colors out of five available, in five attempts,
    so that there are 60 combinations to choose from. In re-
    sponse to each guess, he receives the number of matching
    items, which have the right color at the right position. To
    complicate matters, the puzzle has to be solved within a
    very short time limit.

    Has anybody an idea how to make it honestly, without the as-
    sistance of a computer program that keeps track of exlcuded
    combinations and suggests a next attempt that efficiently
    narrows-down the remaning possiblities? I can't find a
    strategy simple enough to employ in mind & memory.

    --
    () ascii ribbon campaign -- against html e-mail
    /\ http://preview.tinyurl.com/qcy6mjc [archived]
    I don't have one I would want to memorize but I have a table that wins 51/60 times for an individual puzzle.
    It's essentially a printout of a computer program, so it doesn't even meet that limiting part of your request.
    I see some patterns and choices but by the third round it's too messy.
    G1 G2 G3 G4 G5
    123
    _0_231
    ____0_314
    _______0_542
    __________1_452
    _______1_342
    __________0_415
    __________1_412 x512
    __________2_345 x352
    _______2_315
    __________1_354 x514
    __________2_312
    ____1_541
    _______0_215
    __________0_432
    __________1_254 x435
    __________2_214
    _______1_531 x (only non-feasible guess)
    __________0_245
    __________1_351 x451
    __________2_532 x534
    _______2_341
    ____2_531
    _______1_234
    __________1_241 x251
    __________2_235
    _______2_431
    _1_543
    ____0_425
    _______0_134
    __________1_152
    __________2_132 x154
    _______1_321
    __________0_135
    __________2_324
    _______2_325
    __________1_421
    ____1_253
    _______0_524
    __________0_142 x145
    __________2_521
    _______1_413
    _______2_213
    __________1_453
    ____2_513
    _______1_243
    _2_124
    ____1_523
    _______1_143
    __________2_153
    _______2_423
    ____2_125

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Anton Shepelev@21:1/5 to All on Sun Mar 21 17:50:32 2021
    Lawrence E Flynn:

    Ouch. Sorry, I've been solving the Mastermind problem, not
    yours.

    You indeed have.

    I don't have one I would want to memorize but I have a
    table that wins 51/60 times for an individual puzzle.
    It's essentially a printout of a computer program, so it
    doesn't even meet that limiting part of your request. I
    see some patterns and choices but by the third round it's
    too messy.

    Thank you very much anyway. It helped me solve the puzzle
    two times! It is randomly generated and the first time I
    didn't save my game correctly. I will propose your decision
    tree to for inclusion into an existing walkthrough, if I
    find a way to do it, with due credits.

    --
    () ascii ribbon campaign -- against html e-mail
    /\ http://preview.tinyurl.com/qcy6mjc [archived]

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From leflynn@21:1/5 to Anton Shepelev on Sun Mar 21 09:46:53 2021
    On Sunday, March 21, 2021 at 10:50:34 AM UTC-4, Anton Shepelev wrote:
    Lawrence E Flynn:
    Ouch. Sorry, I've been solving the Mastermind problem, not
    yours.
    You indeed have.
    I don't have one I would want to memorize but I have a
    table that wins 51/60 times for an individual puzzle.
    It's essentially a printout of a computer program, so it
    doesn't even meet that limiting part of your request. I
    see some patterns and choices but by the third round it's
    too messy.
    Thank you very much anyway. It helped me solve the puzzle
    two times! It is randomly generated and the first time I
    didn't save my game correctly. I will propose your decision
    tree to for inclusion into an existing walkthrough, if I
    find a way to do it, with due credits.
    --
    () ascii ribbon campaign -- against html e-mail
    /\ http://preview.tinyurl.com/qcy6mjc [archived]
    Anton,
    The approach I gave is probably not optimal. It was not an exhaustive search. One question, are you only allowed to make guesses with three different colors, or could you guess, for example, (A,A,B)?
    L. Flynn

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Anton Shepelev@21:1/5 to All on Sun Mar 21 20:38:22 2021
    Lawrence E Flynn:

    The approach I gave is probably not optimal. It was not an
    exhaustive search.

    I suspected it, because the walkthrough gives a higher
    chance of winning thatn 51/60.

    One question, are you only allowed to make guesses with
    three different colors, or could you guess, for example,
    (A,A,B)?

    No, I am not, and nor are any other players :-)

    I have thought of writing a program that at every step
    chooses a guess that invalidates the most combinations, by
    brutally trying every one of them.

    --
    () ascii ribbon campaign -- against html e-mail
    /\ http://preview.tinyurl.com/qcy6mjc [archived]

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From leflynn@21:1/5 to Anton Shepelev on Sun Apr 4 07:06:00 2021
    On Sunday, March 21, 2021 at 1:38:24 PM UTC-4, Anton Shepelev wrote:
    Lawrence E Flynn:
    The approach I gave is probably not optimal. It was not an
    exhaustive search.
    I suspected it, because the walkthrough gives a higher
    chance of winning thatn 51/60.
    One question, are you only allowed to make guesses with
    three different colors, or could you guess, for example,
    (A,A,B)?
    No, I am not, and nor are any other players :-)

    I have thought of writing a program that at every step
    chooses a guess that invalidates the most combinations, by
    brutally trying every one of them.
    --
    () ascii ribbon campaign -- against html e-mail
    /\ http://preview.tinyurl.com/qcy6mjc [archived]
    Anton,
    Part of the approach I used looks at trying to minimize the maximum number of possible choices in in each of the branches of possible answers (0, 1, 2, or 3 correct) from the current guess given the solution subset for the responses so far. It is
    usually the 0 and 1 sets that are competing in size, although, once we have a 2 response from then on it is the 1 and 2 sets that compete.
    L. Flynn

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From leflynn@21:1/5 to leflynn on Wed Apr 7 18:29:59 2021
    On Sunday, April 4, 2021 at 10:06:01 AM UTC-4, leflynn wrote:
    On Sunday, March 21, 2021 at 1:38:24 PM UTC-4, Anton Shepelev wrote:
    Lawrence E Flynn:
    The approach I gave is probably not optimal. It was not an
    exhaustive search.
    I suspected it, because the walkthrough gives a higher
    chance of winning thatn 51/60.
    One question, are you only allowed to make guesses with
    three different colors, or could you guess, for example,
    (A,A,B)?
    No, I am not, and nor are any other players :-)

    I have thought of writing a program that at every step
    chooses a guess that invalidates the most combinations, by
    brutally trying every one of them.
    --
    () ascii ribbon campaign -- against html e-mail
    /\ http://preview.tinyurl.com/qcy6mjc [archived]
    Anton,
    Part of the approach I used looks at trying to minimize the maximum number of possible choices in in each of the branches of possible answers (0, 1, 2, or 3 correct) from the current guess given the solution subset for the responses so far. It is
    usually the 0 and 1 sets that are competing in size, although, once we have a 2 response from then on it is the 1 and 2 sets that compete.
    L. Flynn

    Anton,
    I have an improvement for the cases where the initial guess has zero hits. It only loses three of those 32 cases. It uses two guesses that are not in the possible solution set at the point they are used.
    G1 G2 G3 G4 G5
    123
    _0_412
    ____0_235
    _______0_354
    __________0_541
    __________1_341
    __________2_351
    _______1_241
    __________0_534
    __________1_254 x531 x345
    __________2_251
    _______2_134 (not a possible solution at this point)
    __________0_245
    __________1_231
    __________2_234
    ____1_314
    _______0_431
    __________0_542
    __________1_532
    __________2_435 x451
    _______1_142 (not a possible solution at this point)
    __________0_215
    __________1_352
    __________2_342
    _______2_214
    __________1_315
    __________2_514
    ____2_312
    _______1_432
    __________1_415
    __________2_452
    _______2_512

    L. Flynn

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From leflynn@21:1/5 to leflynn on Wed Apr 7 21:59:31 2021
    On Wednesday, April 7, 2021 at 9:30:00 PM UTC-4, leflynn wrote:
    On Sunday, April 4, 2021 at 10:06:01 AM UTC-4, leflynn wrote:
    On Sunday, March 21, 2021 at 1:38:24 PM UTC-4, Anton Shepelev wrote:
    Lawrence E Flynn:
    The approach I gave is probably not optimal. It was not an
    exhaustive search.
    I suspected it, because the walkthrough gives a higher
    chance of winning thatn 51/60.
    One question, are you only allowed to make guesses with
    three different colors, or could you guess, for example,
    (A,A,B)?
    No, I am not, and nor are any other players :-)

    I have thought of writing a program that at every step
    chooses a guess that invalidates the most combinations, by
    brutally trying every one of them.
    --
    () ascii ribbon campaign -- against html e-mail
    /\ http://preview.tinyurl.com/qcy6mjc [archived]
    Anton,
    Part of the approach I used looks at trying to minimize the maximum number of possible choices in in each of the branches of possible answers (0, 1, 2, or 3 correct) from the current guess given the solution subset for the responses so far. It is
    usually the 0 and 1 sets that are competing in size, although, once we have a 2 response from then on it is the 1 and 2 sets that compete.
    L. Flynn
    Anton,
    I have an improvement for the cases where the initial guess has zero hits. It only loses three of those 32 cases. It uses two guesses that are not in the possible solution set at the point they are used.
    G1 G2 G3 G4 G5
    123
    _0_412
    ____0_235
    _______0_354
    __________0_541
    __________1_341
    __________2_351
    _______1_241
    __________0_534
    __________1_254 x531 x345
    __________2_251
    _______2_134 (not a possible solution at this point)
    __________0_245
    __________1_231
    __________2_234
    ____1_314
    _______0_431
    __________0_542
    __________1_532
    __________2_435 x451
    _______1_142 (not a possible solution at this point)
    __________0_215
    __________1_352
    __________2_342
    _______2_214
    __________1_315
    __________2_514
    ____2_312
    _______1_432
    __________1_415
    __________2_452
    _______2_512

    L. Flynn
    And an improvement where the first guess has one hit that loses one out of 21. G1 G2 G3 G4 G5
    123
    _1_134
    ____0_413
    _______0_325
    __________1_521
    __________2_321
    _______1_253
    __________0_421 x425
    __________1_543
    __________2_243
    _______2_213
    __________1_453
    __________2_513
    ____1_124 NP
    _______1_145
    __________1_152
    __________2_142
    _______2_324
    __________2_524
    ____2_135
    _______1_154
    _______2_132

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Anton Shepelev@21:1/5 to All on Fri Apr 9 16:59:45 2021
    L. Flynn:

    Anton,
    I have an improvement for the cases where the initial
    guess has zero hits. It only loses three of those 32
    cases. It uses two guesses that are not in the possible
    solution set at the point they are used.
    [...]
    And an improvement where the first guess has one hit that
    loses one out of 21.

    I have already solved the puzzle with the help of your initial
    decision tree. If you will care to provide the entire
    decision tree with these improvements, then I will update it
    in the walkthrough on this WIKI:

    https://strategywiki.org/wiki/The_Journeyman_Project:_Pegasus_Prime/Morimoto_Colony#Shield_Generator
    https://pastebin.com/raw/NCnpAz3s

    --
    () ascii ribbon campaign - against html e-mail
    /\ http://preview.tinyurl.com/qcy6mjc [archived]

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From leflynn@21:1/5 to All on Fri Apr 9 10:52:25 2021
    An improved tree that gets to the answer 56/60 times. The ##Y and ##X are my checks on success or failure for the list of 60 possible answers in numerical order. The odds of winning all of the games in a five-game match is slightly over 70%. It gets to
    the answer for 30 of the cases with four or fewer guesses.
    L. Flynn

    ____G1 G2 G3 G4 G5
    01Y123
    37Y_0_412
    18Y____0_235
    33Y_______0_354
    58Y__________0_541
    31Y__________1_341
    34Y__________2_351
    19Y_______1_241
    57Y__________0_534
    24Y__________1_254 55X531 36X345 28
    22Y__________2_251
    16Y_______2_231
    21Y__________1_245
    17Y__________2_234
    26Y____1_314
    43Y_______0_431
    59Y__________0_542
    56Y__________1_532
    45Y__________2_435 46X451
    35Y_______1_352
    15Y__________0_215
    32Y__________2_342
    14Y_______2_214
    27Y__________1_315
    51Y__________2_514
    25Y____2_312
    44Y_______1_432
    39Y__________1_415
    47Y__________2_452
    49Y_______2_512

    05Y_1_134
    38Y____0_413
    30Y_______0_325
    52Y__________1_521
    28Y__________2_321
    23Y_______1_253
    40Y__________0_421 42X425
    60Y__________1_543
    20Y__________2_243
    13Y_______2_213
    48Y__________1_453
    50Y__________2_513
    09Y____1_145
    29Y_______0_324
    54Y__________2_524
    10Y_______1_152
    07Y_______2_142
    06Y____2_135
    12Y_______1_154
    04Y_______2_132

    02Y_2_124
    53Y____1_523
    08Y_______1_143
    11Y__________2_153
    41Y_______2_423
    03Y____2_125

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From James Dow Allen@21:1/5 to Anton Shepelev on Tue Apr 13 13:16:41 2021
    On Monday, March 8, 2021 at 1:02:42 AM UTC+7, Anton Shepelev wrote:
    Hello, all

    I am stuck at the Master-Mind-like puzzle ...

    Better yet, try this webpage:
    https://james.fabpedigree.com/mmind/mmind.htm
    Five pegs, seven colors. Pretty tough! How many guesses do you need?
    Although details are little known, it's always possible to solve with six guesses.

    At the webpage you can go for computer's secret, have it go for yours, or both concurrently.
    Try it and tell me what you think.

    --
    () ascii ribbon campaign -- against html e-mail
    /\ http://preview.tinyurl.com/qcy6mjc [archived]

    Dearest regards, jamesdowallen

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Anton Shepelev@21:1/5 to All on Mon May 3 13:23:03 2021
    L. Flynn:

    An improved tree that gets to the answer 56/60 times. The
    ##Y and ##X are my checks on success or failure for the
    list of 60 possible answers in numerical order. The odds
    of winning all of the games in a five-game match is
    slightly over 70%. It gets to the answer for 30 of the
    cases with four or fewer guesses.

    Thanks for an improved solution. Hopefully I will get around
    to updating the game's walkthrough.

    --
    () ascii ribbon campaign -- against html e-mail
    /\ http://preview.tinyurl.com/qcy6mjc [archived]

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From leflynn@21:1/5 to leflynn on Wed Jul 7 12:31:51 2021
    On Friday, April 9, 2021 at 1:52:26 PM UTC-4, leflynn wrote:
    An improved tree that gets to the answer 56/60 times. The ##Y and ##X are my checks on success or failure for the list of 60 possible answers in numerical order. The odds of winning all of the games in a five-game match is slightly over 70%. It gets to
    the answer for 30 of the cases with four or fewer guesses.
    L. Flynn
    Here is a version that is converted from colors 1-5 to colors YBGPR
    G1 G2 G3 G4 G5
    YBG
    _0_PYB
    ____0_BGR
    _______0_GRP
    __________0_RPY
    __________1_GPY
    __________2_GRY
    _______1_BPY
    __________0_RGP
    __________1_BRP RGY GPR
    __________2_BRY
    _______2_BGY
    __________1_BPR
    __________2_BGP
    ____1_GYP
    _______0_PGY
    __________0_RPB
    __________1_RGB
    __________2_PGR PRY
    _______1_GRB
    __________0_BYR
    __________2_GPB
    _______2_BYP
    __________1_GYR
    __________2_RYP
    ____2_GYB
    _______1_PGB
    __________1_PYR
    __________2_PRB
    _______2_RYB

    _1_YGP
    ____0_PYG
    _______0_GBR
    __________1_RBY
    __________2_GBY
    _______1_BRG
    __________0_PBY PBR
    __________1_RPG
    __________2_BPG
    _______2_BYG
    __________1_PRG
    __________2_RYG
    ____1_YPR
    _______0_GBP
    __________2_RBP
    _______1_YRB
    _______2_YPB
    ____2_YGR
    _______1_YRP
    _______2_YGB

    _2_YBP
    ____1_RBG
    _______1_YPG
    __________2_YRG
    _______2_PBG
    ____2_YBR

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From Anton Shepelev@21:1/5 to All on Tue Jul 13 19:35:21 2021
    L. Flynn:

    Here is a version that is converted from colors 1-5 to colors YBGPR

    That makes more sense, thank you. I can use it now with less work.

    --
    () ascii ribbon campaign - against html e-mail
    /\ http://preview.tinyurl.com/qcy6mjc [archived]

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)
  • From HenHanna@21:1/5 to James Dow Allen on Wed Mar 20 04:12:47 2024
    James Dow Allen wrote:

    On Monday, March 8, 2021 at 1:02:42 AM UTC+7, Anton Shepelev wrote:
    Hello, all

    I am stuck at the Master-Mind-like puzzle ...

    Better yet, try this webpage:
    https://james.fabpedigree.com/mmind/mmind.htm
    Five pegs, seven colors. Pretty tough! How many guesses do you need? Although details are little known, it's always possible to solve with six guesses.

    At the webpage you can go for computer's secret, have it go for yours, or both concurrently.
    Try it and tell me what you think.

    --
    () ascii ribbon campaign -- against html e-mail
    /\ http://preview.tinyurl.com/qcy6mjc [archived]

    Dearest regards, jamesdowallen







    ------ Great... Within a few days, i'll try to sic my Python-solver on it.

    --- SoupGate-Win32 v1.05
    * Origin: fsxNet Usenet Gateway (21:1/5)