Category: Puzzles

Farewell Raymond Smullyan

Martin Gardner was a kind of gateway drug. When I was a kid I went to the same secondary school that my dad taught at. That wasn’t a particular problem for me but it meant I had to hand around school until he was heading home.

This meant sitting in the school library by myself reading, which, being a bookish sort, was not any great hardship. I ploughed through the sci-fi books but on other occasions, I’d just look at random books.  It was in this way I stumbled across old Penguin (or Pelican?) editions of Gardner’s Mathematical Puzzle’s and Diversions books. Now, I was not much of one for mathematics at the time but I liked puzzles and I liked the odd, arcane nature of these books.

As an older teen Gardner led me to other writers: Douglas Hofstadter (notably I read Metamagical Themas first then Godel, Escher, Bach) but also Raymond Smullyan.

Like Gardner, Smullyan combined a love of puzzles and magic but also the absurd. That logic and absurdity are natural companions is something that people find paradoxical. People think it odd that Lewis Carroll was both an eminent logician and author of Alice in Wonderland despite the absurdism of the Alice books often relying on wordplay and uncooperative literalism.

Smullyan, who died last Monday (Feb 6th 2017) tied absurdity more closely to logic in his complex puzzle books. The connection is overt – using weird settings and strange kinds of people (knights, vampires among others) with proscriptive approaches to communication. The New York times has a substantial sample here https://www.nytimes.com/interactive/2017/02/11/obituaries/smullyan-logic-puzzles.html

So why the close connection between the absurd and logic?

Two elements are at play. Firstly the necessary insistence on literalism. Exactly WHAT is being said? It is an insistence that when applied to normal conversation is a breach of normal social conventions. Secondly, the use of absurd or nonsensical propositions and conclusions helps highlight FORMAL aspects of an argument from informal and empirical aspects. To see how a syllogism functions (for example) it can be misleading to use a string of commonplace truths.

For example, Wikipedia uses this example for Felapton:

No flowers are animals.  All flowers are plants. ∴ Some plants are not animals. (SoP)

But to see the formal connection writers like Carroll might use more odd juxtapositions

No elephants are professors.  All elephants are stamp collectors.  ∴ Some stamp collectors are not professors.

Formal truths need to hold even in absurd worlds.

Smullyan wrote a whole bunch of wonderfully weird books, that used puzzles and odd juxtapositions to exemplify logic and reasoning. I think his most substantial achievement was Forever Undecided: A Puzzle Guide to Godel that used his signature style to lead the reader to Godel’s incompleteness theorems in a charmingly accessible manner.

Logician, Taoist, Magician – a 97 year life of tricks, puzzles and deep thoughts.

 

The Griefer Matrix: Griefers Reloaded

The purpose of focusing on a set of issues is not intended to catalogue ‘victory conditions’ for the GSBA. Such a person/group may have also sorts of other things that they regard as victories or positive results. Rather the focus is on those things which impact negatively on everybody else or provide strong incentives for bad-actors in general to target the Hugo Awards.

Going through the solutions and judging them against the issues is the name of the game. I was going to present this as a literal matrix but it didn’t fit in a visible way. So here it is as a list. Part 1 is here. You probably need it open in a spare tab to cross reference the numbers.
2 Membership

a) Restrict nominations to attending members.
b) Restrict nominations to a jury.
c) Remove second year nominating rights.
d) Increase supporting membership fees.
e) Ban bad actors.
f) Restrict nominating to members with consecutive years of membership.

Each of these attempt to deal with the GSBA at a high level by disrupting their ability to join, nominate, vote or influence voting. Each one attempts to act on the issues listed as A,B,C, and D. Arguably each runs the risk of point E. The GSBA forces a change in the character of the awards.

2a. may achieve A, B, C, D by effectively increasing the cost of nominating. As voting in the final awards is still open to supporting members many of the specific issues may still apply. Specifically 3 & 6 (and hence 10, 12, 13) are still possible actions by a GSBA.
2b. Stops all the issues (assuming the jury functions) except issue E. Changing to a jury award is a major change in the character of the Hugo Awards. It is a single bad outcome but a significant one.

2c. limits the overall capacity of the GSBA to cause issues but does not any specific one.

2d. Raises the cost of the GSBA’s actions but at the expense of members and may result in E.

2e. Would in theory succeed on all issues but it is unclear how this would be done pre-emptively.

2f. This makes it harder for a GSBA to attempt entryist tactics with the award. Again this reduces the ease with which the GSBA can act but may not stop a particularly determined bad-actor.

3. Nomination
a) Implement slate detection algorithms
This may be technically feasible with few false positives. This would only stop issue 5 – i.e. forcing No Award by sweeping a category. Without No Award as an option it would limit the chance of 10 (but No Award is an option of course). It limits point 9 (publicity) by stopping sweeps. However it is easy for a GSBA to adapt to and doesn’t stop 1, 2, 3, 4, 6, 7 or 8. It might encourage 3.5 depending on how it works (i.e. voters might avoid nominating a work on a Griefers slate for fear that their ballot will be ‘detected’ as a slate work – voters may act this way even if the algorithm is smart enough to spot the difference)

b) Allow the membership to confirm nominations through a long list (DN).
This limits/stops 1 but a harassing work would still appear on the long list and this might be sufficient incentive for the GSBA. Unclear of the impact on 2 but would probably stop it. Doesn’t stop 3. Does stop 4, 5, 7 and 8. Probably stops 4.5 and 5.5. Doesn’t stop 6. Possibly encourages 3.5 (hostage taking)

c) Allow membership to deny nominations through a long list (3SV)
This more clearly stops 1 but a harassing work would still appear on the long list and that might be sufficient incentive for the GSBA. May need a quorum or other measures to prevent this system making 2 easier for the GSBA by down voting enemies. Doesn’t stop 3. Does stop 4, 5, 7 and 8. Might not stop 4.5 (weak nominees via a slate) and possibly even 5.5 (sweeps due to a slate) if people are shy about using it on marginal cases. Doesn’t stop 6. Possibly encourages 3.5 (hostage taking)?

d) Empower the admins to remove ballots of bad actors.
Unclear how this could be done and may just result in a shift in tactics. Probably stops 1, 2, 4, 5, 7, 8 depending on what tactics the GSBA is using and how public they are initially. May encourage hostage taking for the same reason as given in (a).

e) Empower the admins to add a limited number of nominees in years with bad actors (A+2)
Stops 5. Limits 2 but only partly. Doesn’t stop 1, 3, 4, 6, 7, 8 (and hence 9, 11, 12)

f) Empower a jury to add a limited number of overlooked nominees to the short list (Juried+2)
Stops 5. Limits 2 perhaps better than (e) as a jury may feel moe able to add nominees. Doesn’t stop 1, 3, 4, 6, 7, 8 (and hence 9, 11, 12)

g) Algorithmically filter the nominations
Hard to say without knowing more. Probably stops 5 and possibly 2. Limits 7 and 8.

h) Restrict the nominations relative to the field (4/6)
Stops 5. Limits 2 a little. Limits 7 and 8. Encourages 3.5 possibly?

j) Empower the admins or a jury to remove works which tend to bring discredit on WorldCon
Assuming other works take the place of the removed ones, stops 1, 4, 5, limits 2. Combing both jury powers in (f) and (j) stops 1, 2, 4, 5, limits 7 and 8. May be perceived as a change in character of the awards.

k) Allow the membership to vote on expanding just the number of nominees on a given year’s shortlist.
A more complex way of doing (e) and (f) with a similar effect.

No solution deals with all issues. #3 and #6 are the hardest issues to deal with via a voting reform. 3b DN and 3c 3SV deal with several high impact issues. A combination jury (3f, 3j) would probably deal with the same range of issues but with less administration and with more capacity to adapt to different circumstance.

Other possibilities beyond voting reform:
A code of conduct around voting/nominating? This would help empower solution 2e (Ban bad actors) but also would act as a deterrent for various kinds of campaigning – particularly negative campaigning. However, may encourage hostage taking (i.e. the GSBA campaigning FOR a declared enemy aka ‘black knighting’)

An anti-shenagigans jury/panel with various discretionary powers might be the only sufficiently flexible solution but it may be perceived as changing the character of the awards – with a remit to ensure the validity and integrity of the voting?

3SV looks like the best bet as a voting change but it needs a quorum limitation to prevent 2 (down voting a GSBA enemy) which in turn may limit its capacity to deal with 4.5 (weak nominees) and even 5.5 (sweeps). Those issues may be solved by EPH or a combination of measures (e.g. +2).

The Griefer Matrix [Update]

[Update: started writing part 2 and discovered I’d left off three important things; hostage taking, nominating poor quality (but not actually discreditable) works, and general sweeping categories. In other words VDs 2016 tactic plus Sad Puppy issues. I’ve added those as x.5 issues in places because I’d already used the other numbers in a whole section I’d written.]

This post will get a bit involved, so be warned.

At File 770 there has been a number of discussions about further reforms to deal with briefing from VD or some future obnoxious person. Regular Filer Stoic Cynic collated a list of possible solutions: http://file770.com/?p=29029

1. Do Nothing
2. Membership
a) Restrict nominations to attending members.
b) Restrict nominations to a jury.
c) Remove second year nominating rights.
d) Increase supporting membership fees.
e) Ban bad actors.
f) Restrict nominating to members with consecutive years of membership.
3. Nomination
a) Implement slate detection algorithms (These were mentioned last year. What would they look like?).
b) Allow the membership to confirm nominations through a long list (DN).
c) Allow membership to deny nominations through a long list (3SV)
d) Empower the admins to remove ballots of bad actors.
e) Empower the admins to add a limited number of nominees in years with bad actors (A+2)
f) Empower a jury to add a limited number of overlooked nominees to the short list (Juried+2)
g) Algorithmically filter the nominations (EPH, EPH+, Diluted Nomination, Frozen Nomination, NOTE: Jameson Quinn is working on another algorithm variant, as yet unnamed, to incorporate a ‘satisfaction index’ and achieve A+2 type results in conjunction with EPH programmatically).
h) Restrict the nominations relative to the field (4/6)
j) Empower the admins or a jury to remove works which tend to bring discredit on WorldCon
k) Allow the membership to vote on expanding just the number of nominees on a given year’s shortlist.
4. Voting
a) Filter the votes by preference ranking (IRV)
b) Veto bad actors (No Award)

There is much active discussion on these – particularly 3c “3SV” which has a 2nd round of voting in which members get to throw out some works from a longlist. In my earlier post about EPH I said that I thought it works better than people were saying because it prevents a briefer forcing No Award on a category. Others (particularly Jim Henley) have pointed out that it doesn’t stop a whole range of other bad-effects. The bad-effects are what motivates a griefer to grief. While they can still gain one of these negative achievements they still have an incentive to make other people miserable.

So based on what other people have said here is a list. Again Jim Henley originated some of these but this is how I understood them – so credit to him for good things and blame to me for bad things about the list 🙂

List of possible things to be solved.

In the list GSBA stands for griefer or slater or bad actor of some kind.

Three meta-level issues

A. The GSBA undermining the reputation of the awards
B. The GSBA demoralising the participating members (and hence also achieve A)
C. The GSBA makes the awards administratively to complex
D. The GSBA uses the awards process to harass/defame either specific individuals or groups
E. The GSBA forces a change in the character of the awards

Specific issues

  1. Harassment/libel of members of the community by the GSBA via nominated works. e.g. stories in which a member of the community is defamed or violence is acted out towards them
  2. Blocking the nomination of a declared enemy of the GSBA  e.g. stopping Tropes vs Women getting a nomination for best related work by using a slate to sweep the category
  3. Blocking by ‘king making’ of a declared enemy of the GSBA e.g. throwing the weight of a block of voters against one finalist to stop somebody else winning and then boasting about it.
    3.5 Hostage taking. Nominating or otherwise associating the GSBA with a legitimate work to cause consternation and baffle voters.
  4.  Ridiculous nominations e.g. nominating obviously not good works to make the award look ridiculous.
    4.5 Nominating weak or poor quality works (e.g. some of the SP3 slate)
  5.   Forcing No Award on a category i.e. using a slate and a tactic like #4 so that members have no option other than to vote No Award. The GSBA doesn’t win but then neither does anybody else.
    5.5 Sweeping a category with works for reasons other than those covered elsewhere (again SP3 issues)
  6. General king making – I’m distinguishing this from 3 in so far as a minority faction may do this in good-faith or even without really thinking about it. The difference is the GSBA will tend to boast about it.
  7. Finalist label-bogarting i.e. getting to be a finalist with no hope of winning just so they can say ‘Hugo finalist’. In particular when this is done by a publisher.
  8. Hugo-packet ride-hitching i.e. becoming a finalist means that you get to have Worldcon distribute your work/samples to a wider audience.
  9. Publicity stunting i.e. the act itself gets you coverage and hence sales, blog hits, notoriety etc
  10.  Winning an award i.e. actually winning a Hugo award
  11.  Radicalization of sympathizers: the Rabids use a kind of Trotskyist model with the Sad – get them involved in a conflict which they will lose but gain more radical sympathizers as a result.
  12. In-group bonding. The Rabids use these kinds of actions as a way of making the group feel involved in something and actively fighting their perceived enemy

Of those #10 is a solved problem. No Award and the final voting system pretty much make it hard for the GSBA to win an award without massive effort. 11 and 12 are not directly solvable but by dealing with many of the others the process by which these work is disrupted – an attack on the Hugos that goes nowhere. The same is true of 10. Undermine success in the others and you undermine in general the publicity enjoyed by the GSBA. To get publicity they have to do something noteworthy.

That leaves 9 specific issues. How do each of the solutions do against the 9 issues? I’ll try and make a matrix of issues versus solutions and see what we see. But that is for next time…

 

A modern take on the classic Lewis Carrol style logic puzzle

Lewis Carroll was a noted logician and wrote intricate logic puzzles that combined his sense of the absurd with his interest in logic. At io9’s  regular puzzle spot they have a modern (1960s) version by Hubert Phillips.

  1. Pickled walnuts are always provided at Professor Piltdown’s parties.
  2. No animal that does not prefer Beethoven to Mozart ever takes a taxi in Bond Street
  3. All armadillos can speak the Basque dialect.
  4. No animal can be registered as a philatelist who does not carry a collapsible umbrella.
  5. Any animal that can speak Basque is eligible for the Tintinnabulum Club.
  6. Only animals that are registered philatelists are invited to Professor Piltdown’s parties.
  7. All animals eligible for the Tintinnabulum Club prefer Mozart to Beethoven.
  8. The only animals that enjoy pickled walnuts are those who get them at Professor Piltdown’s.
  9. Only animals that take taxis in Bond Street carry collapsible umbrellas.

The idea is to pick your way through the chains of logic to see what kinds of conclusions can be drawn. Re-ordering them into related chains can help piece together some of the connections. Dealing with the sentences on a purely abstract basis is one way to go but letting the bizarre imagery pile up on you is more fun.

No animal can be registered as a philatelist who does not carry a collapsible umbrella. Only animals that take taxis in Bond Street carry collapsible umbrellas. No animal that does not prefer Beethoven to Mozart ever takes a taxi in Bond Street.

Can we infer that animals registered as philatelists prefer Beethoven to Mozart?

http://io9.com/this-puzzle-is-completely-absurd-but-you-can-still-sol-1722971133

Area maze puzzle

Another fun puzzle article in The Guardian here: http://www.theguardian.com/science/2015/aug/03/alex-belllos-monday-puzzle-question-area-maze-smarter-than-japanese-schoolchild

You can find simpler styles of this kind of problem in mathematics tests from English speaking countries but in Japan this style of puzzle has become a phenomena. Essentially you just need three things 1. all the numbers in the solution are positive integers (but not all the numbers in the problem) 2. there are no hidden gaps etc in the diagrams 3. all the shapes really are rectangles (they aren’t slightly trapezoidal). From this you can infer missing lengths by lucking at adjacent rectangles and by factorizing the areas. This also allows for hugely complex puzzle to be created that are formed from a chain of shapes.

Gender puzzle

There is a neat puzzle on The Guardian website that has interesting implications for evolution or fictional social policy.spread

The government – which wants to increase the ratio of girls to boys – introduces a law for all couples that states:

If a couple has a boy, they have another child.
If a couple has a girl, they have no more children afterwards.

In other words, couples must carry on producing babies until they have a girl. Then they stop.

Rather than working it out from first principles, this puzzle allows for an experimental approach on a spreadsheet. Here is what I did.

    • Column A: A column of numbers 1 to 100 for each of 100 families
    • Column B: =IF(RAND()>0.5,"G","B")
    • Column C:=IF(OR(B2="G",B2=""),"",IF(RAND()>0.5,"G","B"))
    • Column D etc: fill across from C to about 15 columns (it isn’t likely to go beyond that)

This generates a big matrix of B’s and G’s. The B’s stop when there is a G just as the rule above suggests. The number of children in 100 families will vary but it quickly reveals that the proportion of Boys to Girls under this regime tends towards 50-50.

The policy doesn’t change the overall proportion of boys and girls but it does change the kinds of families there are.

  • lots of families with only one child (which is a girl)
  • lots of families with many children most of whom are boys

The article leaves with this conundrum: What policy can you come up with that produces the highest possible ratio of girls to boys, and what is that ratio? A solution (of sorts) is here http://www.theguardian.com/science/video/2015/jul/20/family-planning-puzzle-solution-video