|
|
|
Archive
Since there has been interesting strategy discussion on these puzzles and by popular request,
I have opened an archive on the last thirty weeks. The discussion section is open for continued posting.
Share and Enjoy
Puzzles 1 to 50 are now available as a , just $5.49.
|
Discussion...
... by: kolumbus
@John: I agree with You, and I liked Candiru's solution very much, because it is based on logic and it is short. That means "elegant" for me. When we discover a swordfish, we are used to exclude digits OUTSIDE the swordfish fields – from the remaining rows/columns. Here we can exclude a digit from a field covered by the swordfish itself. I think it's not a new rule, but an intelligent combination of two methods in apparently one step.
Essentialy the solution contains two steps: swordfish and range check. In a first step we can declare 2 as a SURE CANDIDATE on F89 (because of the swordfish), and in a second step we can exclude 4 from F8 (because of range check).
I am not an expert in programming – but I imagine it should be possible to write the software in this way (?)
... by: Klaus
Hi John, i totally agree with you that looking for rules is a challeging and interesting thing to do, of course perfectly ok. my post was more adressed to the people here, who complain whenever a puzzle needs a bit more than just apllying rules. i do not agree with your last sentences. eliminating a candidate because of a chain that leads to contradiction, is of course a perfectly logic way to proceed. and often, especially when combined with other arguments as Setti, more original and creative than, say, by finding a Jellyfish. Swordfish, i think, comes originally from Sudoko and there, of course, things like musts do not exist. but in Str8ts it means a bit more, namely that the digit becomes a must in the related rows/columns. i think SlowThinker mentioned this in his strategy papers. by the way, i have written a Str8ts solver myself, purely rule based, not a single assumption, or much better guess, in it. for me a challenging thing too. Best wishes
... by: John
@Klaus You say "I never quite got what people saying "I just don’t like guessing" want to say." As one who doesn't like guessing, perhaps I can explain my motivation. I am not so much interested in solving the puzzle as understanding how and why the puzzle can be solved. Every Str8t can be solved by guessing right from the start and I think we all agree that isn't much fun. Trying to find the rules which apply is deeply fascinating and is what reseaqrch scientists and mathematicians do for a living. When I get stuck, I would like to be able to find a rule which allows me to go forward. At present I doubt that we know all the possible rules – not so long ago someone called Setti told us a wonderful rule - and I doubt we will be able to discover rules which allow every puzzle to be solved. Andrew creates these puzzles by random and I doubt he has any idea whether they will be solvable or not - all he knows (I hope!) is that they have just one solution. My second reason for wanting to discover the rules is that I am interested in writing a "rules based" puzzle solver program for Str8ts. I must have a very clear understanding of exactly what a rule is before writing the program because the program must tell the computer each tiny step it must take. Running a program soon tells you if your rule is right or not! So, in this puzzle, although Candiru says he has used 3x3 Swordfish on 2 to delete the 4 in F8, I do not think that is the "classic" Swordfish rule. I think it is an interesting extension of Swordfish based on compartment check, and one which I missed when I tried solving it :-( Loosely, I think the vertical Swordfish rule says "If there are 3 columns in which a mandatory digit appears in only 3 rows, then there can be no other occurrences of that digit in those 3 rows". That is very easy to program - just scan the rows and columns and if you find the pattern, delete all the extra digits. This is a rule – I know what I am looking for before I start looking for it. But if I write my Swordfish rule like this in my program, it does not allow the exclusion of the 4 in F8. The 4 is excluded by the extension which loosely says "Because we have a vertical Swordfish, there must an instance of the digit in each row." This makes 2 mandatory in F89 and compartment check now eliminates the 4. Of course, I can accept that many who don’t want to write software will be very pleased to continue by guessing and they will enjoy the rest of the puzzle. I sometimes do myself. But I don’t find it nearly as satisfying as saying “Hey – that’s a great idea!” when someone like Candiru finds a logical reason for being able to go forward in a puzzle where I am stuck.
... by: JimKnopf
@tambil: its has to be a 6 in column 3 (settis rule).
1. A 6 in ABC3 leads to 4, 5 and 6 in ABC3. That leads to 8 and 9 in HJ3. And that leads to 1, 2 or 3 in HJ3, which is impossible, because J3 must be 5, 6, 7,8 or 9.
2. A 6 in EF3 is impossible because no 5, 6 and 7 in E3.
3. There has to be a 6 in HJ3: J3 = 6
... by: Tambil
Thx everybody, got it now..... I dont normally have this much trouble, brain just not functioning correctly this week!!
... by: Klaus
@Tambil: EF3=2,3 would mean F3=2, this would force F9 > 5, but A9 forces F9 <= 5, so its not possible
... by: E.Bandit
e,f and j
... by: E.Bandit
@tambil: the compartment ranges (left and right!) of d e and j dont allow this! setti 6: 6 in all so 6 in 3 but not top compartments so HJ3 so J3
... by: Tambil
@JimKnopf& @Stefan I'm still confused about being so certain about J3, as if there was a 6 in either BC3, couldnt there be a st8rt of 6,5,4 in those 3 cells, why can't HJ3 be 9,8 or 7, leaving 3,2 for EF3?
Having tried this puzzle a few times, I can see that later on there are things that may point away from this theroy, just wondering how we can be so sure at this point that 6 is in J3.
... by: E.Bandit
@Ilman: the 9live way to solve the puzzle XD, decipher and you get the solution.haha so, attack the weak spot: E23, high/low str8ts possible, now my very complicated way (again) ABC3 not 123? -> no 3 and 4 in E (apply setti on 2,3,4 now to 3,4,D,F!) leads to 5 in H3( ABC3 -234) or 5 in BC4 (abc3 - 345) first not possible -> H89 ~ 45 2nd not possible -> UR GH89 i see better ways now (the starf on 2 fgj789, damn didnt saw it) but i did it .... my way
... by: Ilman
I lived in No 11 Sunnyfield Rd asorccs the avenue from the Bookers. I remember Michael, Harry and Daphne. Next door to us was the Curry’s. Stephen was my mate. He had a brother David and sister Dawn at that time. We used to go in the field and play in the brook. I left when I was aged 5 and went to Speke but not before spending some time at Joseph Williams School. My dad was a joiner and he helped build the school. My name is Joseph William Barlow and I used to think the school was named after me cos. my dad worked on it
... by: JimKnopf
@Stefan: setti tells you, that there is a 6 in column 3. In EF3 its not possible. A 6 in ABC3 will lead to no possible digits in HJ3. So a 6 must be in J3.
... by: Stefan
@Ulrich regarding "after the first 6 solved (C8,C9,E8,G1,H6 and J3)"... How do you solve J3 ?!?
... by: Goues
There is a lot of talk abourt sea creatures. I didn't read all, but I think the solution simplier than going fishing in the ocean.
First of all, I used Setti on almost all numbers, which left be with rather not much candidates in cells at the bottom right of the puzzle. Then, I applied two simple chains (length about three or four, both starting on G8), one more Setti and the puzzle is solved completely.
If I find time, I'll post diagrams when the next puzzle is available.
... by: kolumbus
@Julia: It's a column swordfish: columns 789 have to contain a 2 in row FGJ
... by: JimKnopf
@Julia: Beginning with the 6 solved cells and the solved B7 its just a few settis to go. I dont need a swordfish to solve the puzzle.
... by: Julia
@Candiru I can't see the swordfish. in F the 2 is a maybe and what is with the 2 in G5? @Ulrich: thanks for the hint for the first obvious cells. I haven't thought much about Col 3, but I found J3 but I got stuck here..
... by: Klaus
I never quite got what people saying "i just dont like guessing" want to say. when it comes down to a puzzle not solvable by simply rules what then? is it just a bad puzzle? boring? on the other hand i would not call a puzzle thats solvable by simply applying rules very hard, and surely not extreme. It may be extreme for someone who dont know the rules, because he has to be inventive, but for others its just a matter of accurate routine, no thinking. basically those people want extreme puzzles that are relatively easy to solve. i would rule this out by contradiction. hey, come on, fuck the rules, take up the challenge and be creative. i hope Andrew continues with these kind of puzzles.
... by: kolumbus
I didn't see it either, I confess. My chain was (leading to the same result as Candiru's elimination of 4 without assumption): if J7=2 -> J8=1 -> J9=3 -> F9=2 -> F8=??? -> so G7=2 The other one: if ABC3=456 -> EF3=789 -> HJ3=??? -> so J3=6 (Setti) @Frank: If you call this a guess, you may call any advanced elimination a guess... (my 1st posting WAS polite, you don't know me when I'm NOT polite 8-])
... by: Klaus
thanks Candiru, now i see it. i forgot that the Swordfish (i think 3 x 3 is Swordfish, no?) makes 2 a must in F89 and therefore eliminates 4. thank you!
... by: Candiru
@Klaus: Setti on 5 leads to two naked pairs in Col.9 (23/45) Then you get F89, G78 and J789 as possible places for 2. Three 2's needed = no 4 in F8, no 2 in G5, nekaed Pair 12 in FJ8. So you get GH89 = 456, with G8 = 6 (UR). After that everything goes str8t ahead !
... by: Klaus
@Ulrich: i like your "balanced argument" very much. you can make it going by argueing that no 8 in row F means F4=6, G4=5, B4=4 and therefore B7=8 anyway. @Candiru: i'm confused. probably i miss something obvious but i see no UR at GH89.
... by: Frank
@ kolumbus: I have had an solution, using swordfish on 7 and two settis, but then i did not find a way to solve the puzzle without guessing (which you call chain or presumption - well, this is what I call guessing). I do not like puzzles where I do have to guess to find a solution, this was my point. If you cannot deal with critical remarks, stay calm or at least polite.
... by: Ulrich
@Candiru you are right; now you can solve it without any chains and assumptions!!
... by: Candiru
@ Ulrich How can you be sure there is an 8 in F12345 at this point ? Could still be 3...7. I think it works more easily with jellyfish on 2 in FGJ789 an then UR in GH89.
... by: Ulrich
@Klaus you can solve it without chains and assumptions: after the first 6 solved (C8,C9,E8,G1,H6 and J3) you can use setti on 8: row E is maybe and columns 3 and 7 are maybes; but if a 8 in row E then there is a 8 in column 3 and if not a 8 in row E there is not a 8 in column 3; row E and column 3 are balanced. That means: B7 = 8!! This should help you to lift the secret.
... by: Klaus
Thanks Kolumbus, that sounds a bit better than my way. I could eliminate 8 and 9 from E2, which solves. i liked especially the first one, as it opens a big door to Setti. But the chains are not too short, so that its more like trying. I have to take a second look.
... by: Lawanda
Insights like this liven thigns up around here.
... by: kolumbus
@Klaus: Not quite. I used Setti's rule, 1 UR, and 2 VERY short chains which start with an assumption, but almost could be called normal eliminations. For the 2nd half of the puzzle 1 x-wing (I'm not sure if it's necessary)
... by: Klaus
Has anybody a solution without the new word asumption?
... by: JimKnopf
@kolumbus: I agree. The way you solve the puzzle is the goal, not the solution.
... by: kolumbus
@ Pat and Frank: If your skills or your patience are not sufficient to find a logic solution, you should not put the blame on Andrew's puzzles. You better exercise or go to the 1-4 stars section ...
|
| Facebook fan? Now available on |
 |
| A great place to recommend the 'Extreme' |
- 1h16m after 5th attempt. No guessing
- Keith - 1 day ago - ca. 2 hours; finally something within my reach
- schmidi - 3 days ago - 1 guess 40:28
- Mtunza - 3 days ago - 23:07
one guess B4
- osprey - 3 days ago - according to timer, almost 5 hrs, but had long break in between. took 2 guesses on G4 and F3
- str8ter - 4 days ago - easy going ;-)
- wirbeide - 4 days ago - 17 mins, only Setti and other logic strategies
- Goues - 5 days ago - 1:14 - the first one I could solve - was it really *****?
- Felix - 5 days ago - 31:00 seti on 5; one guessing
- Joe Reitzl - 5 days ago - 22:55 no guess, very nice one
- Scholle - 5 days ago - 56:27 one guessing on G4, no Setti, just logical. First solve for me since 3 weeks.
- kandel - 6 days ago - only trial and error helps: third attempt after two unsuccessful guessings: 21' in 3rd, after 41' in 1st and 16' in 2nd approach.
- grisu - 6 days ago - 27:19
one giuess
- dominic - 6 days ago - 19:50 but second attempt to solve it.
- krissie - 6 days ago - 1:40 with Seti
- zergerdan - 6 days ago - 42 min without guessing
- Tobsi - 6 days ago - 34 mins and about 5 tries....one 50/50 guess after the first 5 givens
- Glenn - 6 days ago - 31 min....very nice one...not too hard
- janmorgenstern - 6 days ago - 31 min....very nice one...not too hard
- janmorgenstern - 6 days ago - 53 min ;-) Seti helps; no guessing only logic
- Jutta - 6 days ago - 1:05: no guess; settis help, its possible!
- john - 6 days ago - 37min on the fifth attempt; solvable with logic and one setti on 5
- ursa - 6 days ago - solvable without guess, 2 h
- kolumbus - 6 days ago - Again a trial and error str8ts, extremely boring.
- Frank - 6 days ago - Only for brute force solvers or guessers (which is the same after all), no fun. Setti doesn't help either.
- Pat - 6 days ago - doable but tough,
enjoy ;)
- vic - 6 days ago - i give up after 2h, could only solve 7 without guessing
- Mikel - 6 days ago - 1:15: again very hard
- JimKnopf - 6 days ago - 2 guesses, 17' on 3rd try. Yeah! (Could not solve last 2 weeks' extremes.)
- A B Traherne - 6 days ago - 47 min with one early guess at g4
- iwl - 6 days ago
|