# A game you will probably solve only by brute force

31 views

### David Baggett

Nov 16, 1994, 5:36:35 PM11/16/94
to
In article <FLEE.94No...@jovial.cse.psu.edu>,
Felix Lee <fl...@cse.psu.edu> wrote:

>Hmm. In what sense? So far, I've figured out how the calculator
>works, partly by trial and error. Does this count as exhaustive
>search?

Well, obviously this is all fairly subjective, but basically what I mean by
"requires exhaustive search" is that you have to find the solution by
serendipity rather than by reasoning, just as many people have had to find
the way into the _Balances_ temple by serendipity rather than deduction.
(I am aware that there is a hint, but the hint was obviously not enough for
many players, including me.)

Probably "requires random search" is more accurate.

If the calculator does something that helps you, and you figure this out
only by trying random things with the calculator and then discovering that
your score has inexplicably gone up, then I'd say you have solved the
puzzle by random search.

If, on the other hand, you determine the rules governing the calculator's
behavior by random search and *then* figure out that it'll do said helpful
thing *by deduction*, then I'd say that you've solved the puzzle
deductively rather than with random search.

But, again, this is subjective. I think you will know what I mean when you
see (or figure out) a walkthrough for += 3.

Dave Baggett
__
d...@ai.mit.edu MIT AI Lab He who has the highest Kibo # when he dies wins.
ADVENTIONS: We make Kuul text adventures! Email for a catalog of releases.

### David Baggett

Nov 15, 1994, 10:57:18 PM11/15/94
to
In response to Russ Bryan's challenge to my claim that I can make a simple
and logical game that he won't be able to solve without exhaustive search
(if at all), Carl de Marcken and I have uploaded "+= 3: A Logical

This is a TADS 2.1 game file. You can play it with any version of TADS >=
2.1, including the new TADS 2.2.

If you want to play on a Mac, you will have to use ResEdit or some similar
program to change the file type and owner to "TADR" and "TADG". (Maybe a
kind Mac user can do this for me and upload a MacBinary file to the archive
if there is sufficient interest.)

Carl and I feel that all solutions follow logically from information the
game gives you, or from common sense. No special knowledge is required.
However, we suspect that few people will be able to solve this game without
simply trying random commands until something works.

On the other hand, we put a reasonable amount of thought into this, and
have tried to pick puzzles that have *very sensible solutions* in order to
avoid complaints that the game is too obscure. Given this, it probably
*is* possible someone may get lucky and solve the game directly. (Those
damn TABU people always cause trouble!)

We'll be surprised if anyone does this, but even if it happens, we still
think the game makes the point well; that is, that it is very easy to make
puzzles that are perfectly logical and require no special knowledge, and
"make sense" in a text adventure context, yet are incredibly (if not
impossibly) difficult to solve directly (i.e., without a random search over
all reasonable input commands).

Once people have either solved it or have given up trying, I'll post the
source and a walkthrough.

A final word: don't expect this to be tremendously fun. It's the product
of about two hours of work and is just a joke (and only a mildly amusing
one at that).

Have at it, kids.

### Felix Lee

Nov 16, 1994, 4:39:15 PM11/16/94
to
Dave Baggett:

>However, we suspect that few people will be able to solve this game without
>simply trying random commands until something works.

Hmm. In what sense? So far, I've figured out how the calculator

works, partly by trial and error. Does this count as exhaustive
search?

Yes, I'll agree that it's very easy to make puzzles that are very hard
to solve. But this doesn't mean they're necessarily bad puzzles,
although you probably have to be something of a puzzle freak to tackle
them even if they're good puzzles.

Hard puzzles that are good puzzles tend to be hard to make.
--

### David Baggett

Nov 17, 1994, 3:29:29 PM11/17/94
to
In article <baf.78...@max.tiac.net>,
Carl Muckenhoupt <b...@max.tiac.net> wrote:

>I'd characterize ... "satisfying the troll" as a problem requiring
>exhaustive search.

To clarify a bit:

The solution is logical; in that sense, you do not *in theory* need to
resort to exhaustive/random search to solve the problem. There is a
deductive solution, and in fact many people would find that solution very
quickly. However, knowing what we do about the people who will play this
game, we designed a problem that the players would be unlikely to solve by
deduction. So for these players, the game will (we suspect) degenerate
into exhaustive search. To our amusement, it appears that this has already
happened.

It is precisely examples like this which show that simple, logical problems
-- even those designed very quickly -- are not necessarily directly
solvable.

You may end up calling the solutions unfair; but many other people who
*could* solve the game directly wouldn't feel this way at all.

### Felix Lee

Nov 17, 1994, 7:47:21 PM11/17/94
to
just some more boring rambling about my thought processes.

various things that Dave said about the game is now leading me down
different lines of thought. most of the stuff I thought might be
significant is probably just misleading. like, the behavior of the
calculator is probably useless.

so now I'm thinking about the red herring, and about "common sense"
approaches.

for a while, I was going down many different complicated paths, using
my own interpretation of "logical" and "exhaustive search". (anyone
want to hear about balanced-trinary numbers? :) I forgot that I should
be approaching those terms from Dave's point of view.

urg. the problem with "common sense" approaches is, adventure games
aren't rich enough to support realistic interaction; it's usually much
more productive to deal with the puzzles on a rationally artificial
basis.

this I guess is an argument against realism in adventure games. :)

I want to talk with the troll, shake his hand, grunt at him, gesture
wildly, smile, dance a jig, rob him blind--anything that will
communicate with him, start some sort of dialog. But he just stands
there, holding out his hand, and grunts.

Is he standing? I can't tell.

oh well. Time to go do something else, and see what my subconscious
will come up with.
--

### Greg Ewing

Nov 17, 1994, 6:59:56 PM11/17/94
to

In article <FLEE.94No...@smalltalk.cse.psu.edu>, fl...@cse.psu.edu (Felix Lee) writes:
|> (Is it fair to use the
|> information that the puzzle is unfair in order to solve it? :)

If it's unfairly unfair (unfair-complete?) then this

|> So far, I don't think I can solve this. But I like the calculator.
|> I'd like to see more things like that in games.

Interesting you should mention that... I have a few ideas
in that direction... (Not unfair ones, I hasten to add!)

Greg Ewing, Computer Science Dept, +--------------------------------------+
University of Canterbury, | A citizen of NewZealandCorp, a |
Christchurch, New Zealand | wholly-owned subsidiary of Japan Inc.|
gr...@cosc.canterbury.ac.nz +--------------------------------------+

### Felix Lee

Nov 17, 1994, 1:22:49 PM11/17/94
to
so today I decided to figure out how large the calculator's stack is,
and it's at least 100 deep. Perhaps arbitrarily deep.

At the moment, this is what I know:

the way the calculator works. (though I'm a little hazy on the
specifics of deClear.)

The troll probably wants something. The troll is ticking, so maybe
there's a time dependency.

3 is probably significant, because it appears all over. ("+= 3",
three troll, 9 buttons, 3 non-? states, 99 points, 1587 moves.)

and I don't think anything else is significant. The bridge is
probably scenery. The troll's hand, umm, I'm not sure. "listen to
self" is surreal but probably meaningless. oh, what about the
footnote?

If this were a fair puzzle, I'd try to find a relationship between all
this information. But since this is an unfair puzzle, I'm probably
going to have to approach this some other way. (Is it fair to use the

information that the puzzle is unfair in order to solve it? :)

So far, I don't think I can solve this. But I like the calculator.

I'd like to see more things like that in games.

--

### Carl Muckenhoupt

Nov 17, 1994, 12:03:32 AM11/17/94
to
fl...@cse.psu.edu (Felix Lee) writes:

>Dave Baggett:
>>However, we suspect that few people will be able to solve this game without
>>simply trying random commands until something works.

>Hmm. In what sense? So far, I've figured out how the calculator
>works, partly by trial and error. Does this count as exhaustive
>search?

I've gotten this far as well - I've figured out how the rules governing
the calculator (I think), but not how they can be used to get past the
troll. Figuring out the calculator's rules is a good puzzle in itself,
and illustrates the difference between exhaustive search (or random
search, whichever you want to call it) and experiment. Experiments give
you feedback that helps you to figure things out. Exhaustive search is
necessary in the absence of feedback. Thus, I'd characterize "figuring
out how the calculator works" as a problem requiring experiment, and

"satisfying the troll" as a problem requiring exhaustive search.

But what do I know, I haven't solved the thing yet.

--
Carl Muckenhoupt
Breakfast Dragon
-==(UDIC)==-

### Gareth Rees

Nov 21, 1994, 1:15:07 PM11/21/94
to
David Baggett (d...@ai.mit.edu) wrote:
> Well, if it would help to make the point, I can write you a three room
> game that you'll never solve in a million years and then make fun of
> your pathetic attempts to solve it. :)

Russ Bryan (rbr...@netcom.com) wrote:
> Please do, Dave. Honestly, I'd love the challenge.

David Baggett (d...@case.ai.mit.edu) writes:
> In response to Russ Bryan's challenge to my claim that I can make a

> simple and logical game that he won't be able to solve [...]

So what does Russ Bryan think of "+=3"? I think we should be told...

My own opinion is that this whole discussion is silly. An interactive
fiction author *wants* his readers to solve the game (what's the point
of writing all that endgame if no-one ever gets that far?). Perhaps he
also wants to make his readers work hard for their rewards -- if so, in
order to increase the enjoyment of some of his readers (those who enjoy
solving difficult puzzles) he must run the risk of putting off others.

The balance of difficulty will come out in playtesting, and if too few
playtesters can solve the puzzles the author will make them easier until
the balance seems right to him.

"Balances", over which this discussion arose, was not (I think)
playtested. Hence the complaints.

--
Gareth Rees

### David Baggett

Nov 21, 1994, 2:11:07 PM11/21/94
to
In article <3aqo3b\$m...@lyra.csx.cam.ac.uk>,
Gareth Rees <gd...@cl.cam.ac.uk> wrote:

>So what does Russ Bryan think of "+=3"? I think we should be told...

First of all, Russ Bryan was not the one who really set off this debate;
that dubious honor goes to Matt Russotto, who, in flaming someone who
couldn't figure out a puzzle in _Curses_, argued that any puzzle that a
reasonable person can design, an intelligent person can solve.

Second of all, why do you care what Russ thinks if you think the whole
discussion is silly? :)

>My own opinion is that this whole discussion is silly.

I agree that it's pretty silly, but only because people are *still*
debating the basic truths that we're talking about here:

1) Some problems are inherently intractable.
2) People are only capable of solving very simple problems.
3) "Logical and simple" are necessary but not sufficient requirements of
*good* puzzles.
4) There are objective criteria for evaluating fairness in adventure games.

It *should* be silly. These things should be totally obvious, certainly
given the half-dozen examples that have been provided.

>An interactive fiction author *wants* his readers to solve the game (what's
>the point of writing all that endgame if no-one ever gets that far?).

That doesn't seem to be the case with _Balances_. It is only a demo! We
certainly can't blame Graham for giving us a free demo, but your claim
about IF authors is certainly not true of _Balances_.

People write games for lots of different reasons. I'm sure Russ Bryan's
game will be a twisted mass of impossible brain-teasers, and that he'll
flame you to hell and back for complaining that the game's too hard. :)

For my part, I'm tired of making puzzle games and have been trying to get
away from the brain-teaser aspect of these games. (Though _Legend_ is
certainly still puzzle-based.) As well as illustrating the four truths
above, I think +=3 demonstrates the fundamental absurdity of
puzzle-oriented works of literature.

>Perhaps he also wants to make his readers work hard for their rewards -- if
>so, in order to increase the enjoyment of some of his readers (those who
>enjoy solving difficult puzzles) he must run the risk of putting off
>others.

IMHO: I do not think that struggling to get through something makes the

>The balance of difficulty will come out in playtesting, and if too few
>playtesters can solve the puzzles the author will make them easier until
>the balance seems right to him.

Yes; so why are so many people objecting to claims that many of the puzzles
in _Balances_, a game that has not been playtested, are unfair and overly
difficult?

Finally, I think _Balances_ is pretty nifty. Just because I think some of
the puzzles are unfair doesn't mean I think it sucks, or that I don't like
Graham, etc. etc.

Dave Baggett
__
MIT AI Lab ADVENTIONS: Kuul text adventures! Email for a catalog of releases

### DBlaheta

Nov 22, 1994, 6:55:12 AM11/22/94
to
Well... since I don't have TADS, and I don't have a big enough hard drive
to download it, could someone post (or email, I *guess*) a transcript of a
session in which they do all the major stuff they've figured out? I'm
starting to get lost in this conversation, but I'm very interested in
it...

Don Blaheta
dbla...@aol.com

### Gareth Rees

Nov 22, 1994, 6:41:18 AM11/22/94
to
I wrote:
> An interactive fiction author *wants* his readers to solve the game
> (what's the point of writing all that endgame if no-one ever gets that
> far?).

David Baggett (d...@case.ai.mit.edu) writes:
> That doesn't seem to be the case with _Balances_. It is only a demo!
> We certainly can't blame Graham for giving us a free demo, but your
> claim about IF authors is certainly not true of _Balances_.

Perhaps the lesson to be learned from this is that playtesting is
incredibly important. Even if you think it's only a demonstration game
with eleven rooms and that people can always look at the source if they
want to solve it, then you should still playtest it!

The original version of "Curses" had no hints at all about the romantic
novel puzzle, and it was the complaints of the playtesters that made
Graham put in the clues.

--
Gareth Rees

### Damien P. Neil

Nov 22, 1994, 10:59:43 AM11/22/94
to
In article <3asm70\$j...@newsbf01.news.aol.com>,

I'll post this, since I suspect that you may not be the only one who

Warning! Complete and utter spoilers coming up! (As if it matters with
a game like this... :>)

------- begin transcript
Wow! This sure has been one tough game so far. But you're almost there --
99 points out of 100 and what looks to be a run-of-the-mill troll on the
bridge puzzle. Should be trivial for an adventure game god like you. And
besides, any puzzle with a logical solution can't be *too* hard to solve.
It's about time you got busy and finished the last puzzle and wrote up a
walkthrough for...

Version 1.0
Copyright (C) 1994 David M. Baggett
WorldClass library version 1.3.0.
Difficulty Rating: Requires Exhaustive Search (10 out of 10)

***

For instructions, type "instructions".

Read footnotes [1] with the note command; e.g., type "note 1."

***

On the Three Troll Bridge

You are standing on a rickety wooden bridge. A burly Three Troll blocks
your passage north, across the bridge.

Something is ticking.

>look at troll
He holds his hand out expectantly.

>i
You have a qualitative calculator.

>remove shirt

>give shirt to troll
The troll accepts your kind offer.

>remove glasses

>give glasses to troll
The troll accepts your kind offer.

>remove shoes

>give shoes to troll
The troll accepts your kind offer. The troll now seems less overbearing.

>n
You stride confidently past the troll and onto a glorious but drafty victory.

*** Your score just went up by a point. ***

In a total of 1595 turns, you have achieved a score of 100 points out of a
possible 100.

You may restore a saved game, start over, quit, or undo the current command.
Please enter RESTORE, RESTART, QUIT, or UNDO: >quit
Thanks for playing this game!
------- end transcript

The calculator is useless -- you can play with it, but nothing else. The
troll won't accept it.

There is a red herring (literally) in the water under the bridge.

In the `easier' version, the troll will say `One!', `Two!', and `Three!'
as you give him three items of clothing. In addition (due to a bug, I
believe), he will accept the calculator, and items given to him will not

- Damien

### Darin Johnson

Nov 22, 1994, 1:37:09 PM11/22/94
to
> The calculator is useless -- you can play with it, but nothing else. The
> troll won't accept it.

Hmm, the 3easier did. I assumed this was the "correct" solution
for both versions.

If you indeed are supposed to remove articles of clothing, then
I'd have to say that it's an illogical game! Of course, there
were no real instructions, so one assumes that like all other
adventure games experienced, items that don't show up don't
exist. Sure, it's an implicit rule, but it's still a rule :-)
Sort of like having a "mate in 3" puzzle where the solution is
to whittle down a bishop into a pawn...
--
Darin Johnson
djoh...@ucsd.edu
"You used to be big."
"I am big. It's the pictures that got small."

### David Baggett

Nov 23, 1994, 12:22:04 PM11/23/94
to
In article <3at4hf\$d...@usenet.ins.cwru.edu>,

Damien P. Neil <dam...@b63519.student.cwru.edu> wrote:

>In the `easier' version, the troll will say `One!', `Two!', and `Three!'
>as you give him three items of clothing. In addition (due to a bug, I
>believe), he will accept the calculator, and items given to him will not

He should accept the calculator (this seems a bit more fair, and
guarantees that you will get the "One!" hint), but the stuff should
not stick around when you give it to him.

### David Baggett

Nov 23, 1994, 5:28:22 PM11/23/94
to
In article <DJOHNSON.94...@seuss.ucsd.edu>,
Darin Johnson <djoh...@seuss.ucsd.edu> wrote:

>If you indeed are supposed to remove articles of clothing, then
>I'd have to say that it's an illogical game! Of course, there
>were no real instructions, so one assumes that like all other
>adventure games experienced, items that don't show up don't
>exist. Sure, it's an implicit rule, but it's still a rule :-)

You have not argued that the puzzle is illogical; you have argued that it
is difficult, and perhaps unfair (which I would agree with). That you
assumed things that prevented you from finding the solution does not have
any bearing on whether or not the solution is logical. I don't know how I
can say this any more plainly.

What's more, your claim that "items that don't show up don't exist" is also
not true of many adventure games. HHGTG plays very similar games (e.g.,
[no] tea), and there are hundreds of objects that exist in the "text-based
virtual reality" that are never explicitly mentioned, and which you can
manipulate; e.g., your body in general ("x me"), your hands ("get X"), your
eyes ("examine X"), your ears ("listen to X"), your legs and feet
("north"), the floor/ground/ceiling/walls, the air, the ambient light
and sources thereof, etc. etc.

*** UU2 spoiler ***

In UU2 I've got a puzzle that requires you to put a clothespin on your
nose. Nobody has any trouble with this once they see the clothespin and
realize that they must avoid a bad smell. However, the game never
explicitly tells the player that he has a nose. This again contradicts
your claim -- the puzzle is certainly fair, and pretty easy too.

There's even an amusing twist to this. In UU2 you can transmogrify
yourself into various animals, only some of which have noses. People never
cease to be amused at the message "As a newt, you do not have a nose to
clamp the clothespin onto." Not only do people *assume* that they've got
noses, they find it funny when they're reminded that they're noseless.

(And furthermore, no one has ever wondered how an animal with no hands
can pick things up, turn dials, type passwords, etc.)

We *do* use a great deal of real-world common sense knowledge when playing
these games. The fact that, in general, items that are not listed are not
"important" never stops players from finding solutions to easy puzzles that
require manipulations of unlisted items. This is a matter of difficulty,
not logic.

The puzzle in +=3 is difficult, and it is pretty unfair, but it is not
illogical.

### Greg Ewing

Nov 23, 1994, 9:22:24 PM11/23/94
to

In article <3b0fm6...@life.ai.mit.edu>, d...@case.ai.mit.edu (David Baggett) writes:
|>
|> What's more, your claim that "items that don't show up don't exist" is also
|> not true of many adventure games. HHGTG plays very similar games

But nobody claimed those puzzles were "logical". To me a logical
puzzle is one where you know all the valid moves beforehand.
By describing your puzzle as "logical", I think you have
misled a lot of people into thinking it was a puzzle of this
sort, when it was not. That is what is being perceived as
unfair - not the puzzle itself, but the way it was presented.

|> The puzzle in +=3 is difficult, and it is pretty unfair, but it is not
|> illogical.

I think it is more accurate to say it is *sensible*, which
is not quite the same thing. We were all expecting some
fiendishly difficult logic puzzle, whereas it was just
a fairly ordinary guess-the-verb puzzle. We feel cheated!

|> Dave Baggett
|> __
|> d...@ai.mit.edu "Your ruse; your cunning attempt to trick me." -- Clerks

Greg Ewing, Computer Science Dept, +--------------------------------------+

### David Baggett

Nov 23, 1994, 11:05:08 PM11/23/94
to
In article <3b0td0\$r...@cantua.canterbury.ac.nz>,
Greg Ewing <gr...@cosc.canterbury.ac.nz> wrote:

>But nobody claimed those puzzles were "logical".

This has nothing to do with my argument; I mentioned those puzzles only in
reponse to the claim that players assume that unlisted items are not
manipulatable.

>To me a logical puzzle is one where you know all the valid moves
>beforehand.

You may define "logical" as you wish, but when I said "logical" I meant
"logical" as it is typically defined. I said "logical", meant "logical",
and the puzzle is "logical."

Folks, I'm really disappointed. I go to considerable effort to read
the messages here carefully and completely before responding. Likewise,
I try not to post something in a thread unless I actually know what the
topic is, and know something about the topic.

It seems like no one is bothering to try to understand what Carl and
I are trying to argue here. This has been one off-the-cuff remark
after another. We've also had quite a few people post to ask what
the point of this discussion is, or to complain that the discussion
is stupid, without bothering to go back and read the discussion
itself before posting.

Let's review, so everyone will know what's going on:

The original claim was quite simply stated: A puzzle that a person designs
can be solved by another (smart) person. Originally, I just posted a few
*examples* that rather plainly illustrated how untrue this was. I didn't
expect this to be a big deal. It seems quite obvious to me. There are
many well-known problems that are easily posed and known to be incredibly
difficult to solve.

Given continuing dissent, I jokingly suggested that I write an insanely
difficult game to make the point, and I included a few more examples of
possible puzzles for such a game.

Since there was *still* some argument, Carl and I thought about the
challenge for a minute, and came up with a very basic design. I
implemented the puzzle in about two hours, and posted the game.

The point the game makes is that in practise people are only capable of
solving a very small subset of all possible adventure game puzzles. In
particular, it is not enough for a puzzle to be "logical". It must also
fair, and it must not violate certain vague playability criteria.

Again, we did not expect this to be a controversial claim. For those who
have asked what the point of this debate is, there you go. Read the
previous paragraph again to make sure you understand it. That's all
there is to it. No big deal.

Yet there are *still* arguments. No, the game was not designed to
prove that Dave and Carl are smarter than Russ Bryan. No, it was not
supposed to show how much better TADS is than Inform. No, it is not
making fun of Balances. It is not supposed to be analogous to Balances.
Balances has nothing to do with +=3.

It's just supposed to show that a very straightforward common sense puzzle
can still be very hard to solve. In the process, it does point out that
veteran adventure game players make a lot of assumptions when they play
games.

Don't try to tell me that logical doesn't mean what I know it means,
and what a dictionary will tell you it means. It does not mean any
of the following:

1) The sequence of events required to solve the puzzle is known a priori
2) The puzzle does not violate any of the standard advenutre game assumptions
3) The puzzle is one that we could expect most players to solve
4) The puzzle is one Russ Bryan could solve
5) The puzzle is fair
6) The puzzle is a good puzzle
7) The puzzle is like a puzzle in Balances
8) The puzzle is no harder than puzzles in some Unnkulian games

If you want to define logical in any of these ways, just say so and I will
know not to argue with you (in English, at least). Otherwise, you will
have to go by what the word "logical" means in English, not what it
evidently means in adventure game veteran argot.

>We were all expecting some fiendishly difficult logic puzzle, whereas it
>was just a fairly ordinary guess-the-verb puzzle.

This is *exactly* the kind of off-the-cuff remark that I'm talking about.
Clearly you do not mean "guess-the-verb". Determining the correct syntax
for the command is most certainly *not* the hard part. The hard part is
deducing the *logical* common sense solution required, even though that
solution is, by adventure game standards, atypical.

Dave Baggett
__

### DBlaheta

Nov 25, 1994, 1:15:16 AM11/25/94
to
In article <3b13d...@life.ai.mit.edu>, d...@case.ai.mit.edu (David
Baggett) writes:

> This is *exactly* the kind of off-the-cuff remark that I'm talking
> Clearly you do not mean "guess-the-verb". Determining the correct
syntax
> for the command is most certainly *not* the hard part. The hard part is
> deducing the *logical* common sense solution required, even though that
> solution is, by adventure game standards, atypical.

Clearly he *does* mean "guess-the-verb". Given not a hint, or even
acknowledgement of the existence of the clothing you are to manipulate,
the *only* way to solve the puzzle is "guess-the-verb". If I had seen a
single message *anywhere* in that transcript which remotely mentions
clothing, that puzzle would be a fiendishly difficult but logical puzzle.
As it was, there was nothing logical about it. If you think it's so
logical, prove it. I want to see the *exact* Premises you use in your
proof, so I can refute them (or not).

This is a challenge, of sorts.

Don Blaheta
dbla...@aol.com

### Paul Francis Gilbert

Nov 25, 1994, 8:02:45 PM11/25/94
to
dam...@b63519.student.cwru.edu (Damien P. Neil) writes:

>In the `easier' version, the troll will say `One!', `Two!', and `Three!'
>as you give him three items of clothing. In addition (due to a bug, I
>believe), he will accept the calculator, and items given to him will not

> - Damien

In addition, in the easier version the same bug allows you to throw the
calculator at the troll three times (since you don't actually GIVE it to him it
doesn't disappear). I'm unsure whether Dave has fixed it up yet, since it did it

--
Paul Gilbert | s940...@yallara.cs.rmit.edu.au
Bach App Sci, Bach Eng | The opinions expressed are my own, all my own, and
Year 1, RMIT Melbourne | as such will contain no references to small furry
Australia | creatures from Alpha Centauri.

### David Baggett

Nov 27, 1994, 10:47:54 AM11/27/94
to
In article <3b3vdk\$1...@newsbf01.news.aol.com>,
DBlaheta <dbla...@aol.com> wrote:

>This is a challenge, of sorts.

Challenge declined. If you are really interested in why I think the puzzle
is logical, read the other dozen messages I've posted on the same topic.
If they have already expired at your site, never fear; this newsgroup is
archived on ftp.gmd.de.

### Andrew Southwick

Nov 28, 1994, 5:50:22 PM11/28/94
to
A thought experiment from the peanut gallery.

dbla...@aol.com (DBlaheta) writes:
>d...@case.ai.mit.edu (David >Baggett) writes:

>> This is *exactly* the kind of off-the-cuff remark that I'm talking
>> Clearly you do not mean "guess-the-verb". Determining the correct
>syntax
>> for the command is most certainly *not* the hard part. The hard part is
>> deducing the *logical* common sense solution required, even though that
>> solution is, by adventure game standards, atypical.

>Clearly he *does* mean "guess-the-verb". Given not a hint, or even

"guess-the-verb" means "I know how to solve this puzzle, and what I want to
do, but this game just doesn't understand any of the verbs I've tried."
You are using g-t-v here as "I have no clue, I'll just try random verbs
until something interesting happens," which is by no means the standard
interpretation.

>acknowledgement of the existence of the clothing you are to manipulate,
>the *only* way to solve the puzzle is "guess-the-verb". If I had seen a
>single message *anywhere* in that transcript which remotely mentions
>clothing, that puzzle would be a fiendishly difficult but logical puzzle.

He didn't *say* it was a logical puzzle. He said it had a logical solution.
You hear the solution and say, yeah, that makes sense. You give the troll
the shirt off your back. It *is* fiendishly difficult to solve because it
is not very fair and no hint is made of your clothing. It plays on the
assumptions that IF-players make to speed up their games.

The actions needed to solve the game itself are logical. Is there any
logical method by which to arrive at this solution? Remeber that logical
does not mean easy.

>As it was, there was nothing logical about it. If you think it's so
>logical, prove it. I want to see the *exact* Premises you use in your
>proof, so I can refute them (or not).

I would wager that most of those who failed to solve the puzzle directly
did so because they gave up. Their thoughts went something like this:
- He wants something.
- What does he want?
- What can I give him?
- Is there anything around?
No.
- What am I carrying?
A calculator.
- Does he want the calculator?
No.
- Well, there's nothing I can give him.
It is at this point that the player has given up. He has gone with his
assumption that if it isn't mentioned, it isn't there. Or, more
appropriately, if it isn't mentioned, it isn't manipulable. Which is
sometimes an effective assumption, but it is not an absolute one. If the
game said you had on a pair of jeans, a Oxford, and a jacket, would you
have thought to search your pockets? Or are you also assuming that
>i
will search your pockets for you?

Now imagine that the game, at some point, told you that the Troll spits
on your shirt. Would you have been surprised to learn that you had a shirt
on? I doubt it. But this *would* have made the puzzle easier, because it
would have challenged your assumption that there was nothing else in the
game-universe that you could manipulate.

The question to ask yourself, in this game, is "What assumptions am I
making here?" Good luck, because this is often an extremely difficult
question.

>Don Blaheta

Andrew R. Southwick asout...@vnet.ibm.com
I am not a lawyer, but I've seen one on TV I speak not for IBM
-- Freedom by permission is a contradiction in terms --

### Felix Lee

Nov 29, 1994, 8:23:51 AM11/29/94
to
Andrew Southwick:

>It is at this point that the player has given up. He has gone with his
>assumption that if it isn't mentioned, it isn't there. Or, more
>appropriately, if it isn't mentioned, it isn't manipulable.

well, almost. I think I was working from an assumption that the
puzzle was "regular" and "consistent" in ways that it wasn't. At one
point I explored the possibility that the player was "special" in some
way, but the response to "listen to me" closed off that line of inquiry.

btw, I don't think I would have solved "3easier" either; the counting
from the troll is pretty redundant. Once you find something the troll
will accept, it's pretty obvious what the name "three troll" means.

The name "three troll" is itself a strange violation of realism that
led me to prefer nonrealistic approaches to solving the puzzle.

but anyway, this is all beside the point, right? wasn't the original
point to make a "logical" puzzle that's impossible to solve without
exhaustive search? I think there are too many other things going on
in +=3 for it to be a good example of this..

a lot of the argument seems to be because people are using different
definitions of "logical puzzle".
--

### John Holder

Nov 29, 1994, 11:12:26 AM11/29/94
to
Into the void.... ;^)

Seriously, I've refrained from posting on this thread because of the very
same off-the-cuff responses Dave mentions, obvoiusly few people here know
what logic really means. To continue:

Thus spake David Baggett (d...@case.ai.mit.edu):
] In article <3b0td0\$r...@cantua.canterbury.ac.nz>,
] Greg Ewing <gr...@cosc.canterbury.ac.nz> wrote:

] >To me a logical puzzle is one where you know all the valid moves
] >beforehand.

Ouch! This cannot hold. Does the existance of a logical mathematical
proof mean you know all of the correct steps of the proof beforehand?

] You may define "logical" as you wish, but when I said "logical" I meant

] "logical" as it is typically defined. I said "logical", meant "logical",
] and the puzzle is "logical."

Amen, Dave.
In case others are unclear and wonder what logical _really_ means, try
this(these) definition(s):

logical: According to or agreeing with the principles of logic.
logic: The science that investigates the principles governing CORRECT or
RELIABLE inferance.
inferance: The process of deriving from ASSUMED premises either the strict
logical conclusion or one THAT IS TO SOME DEGREE PROBABLE.

Notice that these definitions show that nothing about a logical puzzle
must be explicitly stated before encountering the puzzle (assumed premises
are valid), as long as the puzzle can be solved by a somewhat probable
solution. In the case of 3+=, The solution is probable if you assume
enough things. The thing that makes this puzzle hard is knowing what to
assume, not whether the puzzle is logical.

] [---deletia---]

] The point the game makes is that in practise people are only capable of

] solving a very small subset of all possible adventure game puzzles. In
] particular, it is not enough for a puzzle to be "logical". It must also
] fair, and it must not violate certain vague playability criteria.

Once again, I must agree with Dave. His puzzle was certainly a logical one,
but it was very unfair since we weren't explicitly told what to assume,
unlike many other adventuring puzzles. In fact, we were unfairly led to
assume an incorrect set of assumptions.

] Again, we did not expect this to be a controversial claim. For those who

] have asked what the point of this debate is, there you go. Read the
] previous paragraph again to make sure you understand it. That's all
] there is to it. No big deal.

] [---deletia---]

] It's just supposed to show that a very straightforward common sense puzzle

] can still be very hard to solve. In the process, it does point out that
] veteran adventure game players make a lot of assumptions when they play
] games.

My point exactly, which is good since I agree with Dave.

] [--deletia--] (definitions of what logical is NOT)

] >We were all expecting some fiendishly difficult logic puzzle, whereas it

] >was just a fairly ordinary guess-the-verb puzzle.

Hmmm. Perhaps what you think of as a guess-the-verb puzzle is really a lack
of the correct assumptions for a game, although if that is the case you
might as well say so instead of hiding behind the term "guess-the-verb puzzle."

John
__ __
__/\_\ John Holder - jho...@nmsu.edu /_/\__
/\_\/_/ Computer Science - New Mexico State University \_\/_/\
\/_/\_\ Coffee should be as black as hell, as strong /_/\_\/
\/_/ as death, and as sweet as love. (Turk. prov.) \_\/

### Greg Ewing

Nov 29, 1994, 5:53:08 PM11/29/94
to

In article <3bdmre\$r...@watnews1.watson.ibm.com>, and...@csgrad.cs.vt.edu (Andrew Southwick) writes:
|>
|> dbla...@aol.com (DBlaheta) writes:
|> >d...@case.ai.mit.edu (David >Baggett) writes:
|>
|> >> Clearly you do not mean "guess-the-verb".

|> >Clearly he *does* mean "guess-the-verb".

Actually, David Baggett was right - that wasn't quite the
right term for what I wanted to say.

Let me try again. As I understood it, the purpose of +=3 was
to exhibit a puzzle whose only *reliable* means of solution
is an exhaustive search, and thereby to prove by example that
such puzzles exist.

But, to carry out an exhaustive search, you have to know
what the search space is! If you know what all the valid
basic moves are, then you can systematically try out all
possible finite sequences of such moves, and you will
eventually find the solution.

In +=3, we don't even know what all the valid moves are.
There is *no* reliable way of solving such a puzzle.
The only way to "solve" a puzzle like that is to be
lucky enough to guess the answer, which is why I described
it (somewhat inaccurately) as a "guess-the-verb" puzzle.
What I really meant was that it's a "guess-the-answer"
puzzle, rather than a "find-the-answer" puzzle.

I'm not saying that the author of +=3 is to blame for
use of the word "logical": we thought he meant it in
the rigorous sense, whereas he only meant it in the
everyday, non-rigorous sense.

I think the only reason we're making a fuss about it is
that we're a little disappointed. A puzzle requiring
exhaustive search is uninteresting, but at least a
tiny amount of masochistic pleasure can be obtained from
bashing one's head against it. A puzzle you can only
guess at is even less interesting, and my reaction to
such a thing is "why bother".

The point I want to make is that there are two types of
very difficult puzzle, and to make it clear which type
+=3 is an example of.

|> >Don Blaheta

|> Andrew R. Southwick asout...@vnet.ibm.com

Greg Ewing, Computer Science Dept, +--------------------------------------+

### Greg Ewing

Nov 29, 1994, 6:33:22 PM11/29/94
to

In article <3bfjta\$o...@dns1.NMSU.Edu>, jho...@nmsu.edu (John Holder) writes:
|>
|> ] >To me a logical puzzle is one where you know all the valid moves
|> ] >beforehand.
|>
|> Ouch! This cannot hold. Does the existance of a logical mathematical
|> proof mean you know all of the correct steps of the proof beforehand?

No, but you know all of the valid transformations which
can occur at each step of the proof. The only thing you
don't know is what order to apply them in.

|> In case others are unclear and wonder what logical _really_ means, try
|> this(these) definition(s):
|>

|> logic: The science that investigates the principles governing CORRECT or
|> RELIABLE inferance.
|> inferance: The process of deriving from ASSUMED premises either the strict
|> logical conclusion or one THAT IS TO SOME DEGREE PROBABLE.

What you've just described is the everyday, non-rigorous
usage of the term "logic". There is also another meaning
to the term which is widely used and understood. Neither
one is better or more correct than the other - they're
just different uses of the same word.

|> Notice that these definitions show that nothing about a logical puzzle
|> must be explicitly stated before encountering the puzzle

Consider this scenario: A friend gives you an object made
from pieces of wire bent into strange shapes, with a loop
of string entwined among them. He says, "Remove the string."

You try every contortion you can think of, but are unable
to free the string. Eventually you give up and say, "Okay,
I'm stumped - show me how it's done."

He takes out a pair of scissors and cuts the string.

I know what my reaction would be - and it wouldn't be
favourable! Note that the friend never actually said that it
was logical, or even that it was a puzzle. But that wouldn't
make me feel any better. The form of the object very
strongly implies that it is a puzzle, and one with a
rigorously logical solution. Even though he didn't say
so explicitly, I would feel that my friend had lied
to me.

It's exactly the same reaction people are having to Dave's
puzzle.

|> Hmmm. Perhaps what you think of as a guess-the-verb puzzle is really a lack
|> of the correct assumptions for a game,

It's precisely this lack of reliable assumptions that makes
the puzzle non-logical (in the rigorous sense). If there are
no axioms you can rely on, then you can't even apply logic
to the problem.

|> __/\_\ John Holder - jho...@nmsu.edu /_/\__
|> /\_\/_/ Computer Science - New Mexico State University \_\/_/\

Greg Ewing, Computer Science Dept, +--------------------------------------+

### Neil Demause

Nov 29, 1994, 3:48:08 PM11/29/94
to

: Consider this scenario: A friend gives you an object made

: from pieces of wire bent into strange shapes, with a loop
: of string entwined among them. He says, "Remove the string."

: You try every contortion you can think of, but are unable
: to free the string. Eventually you give up and say, "Okay,
: I'm stumped - show me how it's done."

: He takes out a pair of scissors and cuts the string.

Heh. That actually sounds like a neat trick to me. But then, I liked
Dave's puzzle too.

A lot of this reminds me of the story about the science test with the
question: "How can you tell the height of a building with a barometer?"
One student, so the story goes, writes, "Drop the barometer off the top
of the building, time how long it takes to fall, apply 32ft/sec/sec, etc."

The teacher marks it wrong. Student complains. Teacher agrees to give
student another shot at it. This time the student writes down a whole
series of "wrong" answers, ranging from "Measure the height of the
barometer and length of its shadow, and the length of the building's
shadow" to "Walk up the stairs counting off the height in
barometer-lengths" to my absolute favorite: "Go knock on the door of the
building superintendent. Say, 'Mr. Superintendent, I have a very fine
barometer here that I will give you if you can tell me the height of this
building...'"

This, it seems to me, is not just in the spirit of i-f games, it is the
very *essence* of them. It's all about misdirection and playing off your
assumptions, and if Dave just happened to pick some deep-rooted
assumptions to pick on, that doesn't make it any less kosher.

Neil deMause

### John Holder

Nov 30, 1994, 10:37:17 AM11/30/94
to
Thus spake Greg Ewing (gr...@cosc.canterbury.ac.nz):

] In article <3bfjta\$o...@dns1.NMSU.Edu>, jho...@nmsu.edu (John Holder) writes:
] |>
] |> ] >To me a logical puzzle is one where you know all the valid moves
] |> ] >beforehand.
] |>
] |> Ouch! This cannot hold. Does the existance of a logical mathematical
] |> proof mean you know all of the correct steps of the proof beforehand?

] No, but you know all of the valid transformations which
] can occur at each step of the proof. The only thing you
] don't know is what order to apply them in.

If some else tells you about the existance of a proof, all you necessarily
know is that a solution exists - just because a given proof in mathematics
exists doesn't guarantee that you or I know all of the necessary
ttransformations or even where to find the necessary transformations, all that
is guaranteed is that any person who has PROVED the proof knows the necessary
tranformations. I fine hiar to split, but an important one. Given
enough time, you could theoretically learn all possible transformations and
the play "guess-the-transformation", much like the "guess-the-puzzle" you
suggest.

] |> In case others are unclear and wonder what logical _really_ means, try

] |> this(these) definition(s):
] |>
] |> logic: The science that investigates the principles governing CORRECT or
] |> RELIABLE inferance.
] |> inferance: The process of deriving from ASSUMED premises either the strict
] |> logical conclusion or one THAT IS TO SOME DEGREE PROBABLE.

] What you've just described is the everyday, non-rigorous
] usage of the term "logic". There is also another meaning
] to the term which is widely used and understood. Neither
] one is better or more correct than the other - they're
] just different uses of the same word.

I'm afraid most puzzles of the adventure game variety that I have run into
require the use of what you are calling "everyday logic" rather than rigorous
mathmatical logic.

] |> Notice that these definitions show that nothing about a logical puzzle

] |> must be explicitly stated before encountering the puzzle

] Consider this scenario: A friend gives you an object made
] from pieces of wire bent into strange shapes, with a loop
] of string entwined among them. He says, "Remove the string."

] You try every contortion you can think of, but are unable
] to free the string. Eventually you give up and say, "Okay,
] I'm stumped - show me how it's done."

] He takes out a pair of scissors and cuts the string.

My reaction happens to be a smile. I am impressed that I didn't ask if
I was allowed to cut the string to solve the puzzle, and because I didn't
ask, I didn't make a correct assumption.

] I know what my reaction would be - and it wouldn't be
] favourable!

You obviously don't like making poor assumptions, or not being able to assume
everything about a given situation. Doesn't change things.

] The form of the object very

] strongly implies that it is a puzzle, and one with a
] rigorously logical solution. Even though he didn't say
] so explicitly, I would feel that my friend had lied
] to me.

Gee, I thought the solution was rigorously logical. Especially once you
know the right transformation. All you do is change the topology of the
string...

] It's exactly the same reaction people are having to Dave's
] puzzle.

]|> Hmmm. Perhaps what you think of as a guess-the-verb puzzle is really a lack
]|> of the correct assumptions for a game,

] It's precisely this lack of reliable assumptions that makes
] the puzzle non-logical (in the rigorous sense). If there are
] no axioms you can rely on, then you can't even apply logic
] to the problem.

Did you assume anything about having clothes? If you did, did you think
a) you were naked or
b) you were clothed?

If you didn't, you had the same problem as with the string puzzle - all of the
transformations were not immediately obvious to you.

John
__ __

__/\_\ John Holder - jho...@nmsu.edu /_/\__
/\_\/_/ Computer Science - New Mexico State University \_\/_/\

### Neil K. Guy

Dec 1, 1994, 3:40:25 AM12/1/94
to
gr...@cosc.canterbury.ac.nz (Greg Ewing) writes:

>Consider this scenario: A friend gives you an object made
>from pieces of wire bent into strange shapes, with a loop
>of string entwined among them. He says, "Remove the string."

>You try every contortion you can think of, but are unable
>to free the string. Eventually you give up and say, "Okay,
>I'm stumped - show me how it's done."

>He takes out a pair of scissors and cuts the string.

Hm. Sounds like cutting the Gordian knot to me. If it was good enough
for Alexander the Great surely it's good enough for anyone else. That
is, unless you have more parochial concerns related to solving fiddly
puzzles rather than conquering entire continents...

Obligatory smiley -> :)

- Neil K.

### r...@vectorbd.com

Nov 30, 1994, 10:56:08 PM11/30/94
to
Greg Ewing (gr...@cosc.canterbury.ac.nz) wrote:

: In article <3bdmre\$r...@watnews1.watson.ibm.com>, and...@csgrad.cs.vt.edu (Andrew Southwick) writes:
: |> dbla...@aol.com (DBlaheta) writes:
: |> >d...@case.ai.mit.edu (David >Baggett) writes:
: |> >> Clearly you do not mean "guess-the-verb".
: |> >Clearly he *does* mean "guess-the-verb".
:
: Actually, David Baggett was right - that wasn't quite the
: right term for what I wanted to say.
:
: [Deletia]
:
: The only way to "solve" a puzzle like that is to be

: lucky enough to guess the answer, which is why I described
: it (somewhat inaccurately) as a "guess-the-verb" puzzle.
: What I really meant was that it's a "guess-the-answer"
: puzzle, rather than a "find-the-answer" puzzle.

No, no, it's a "guess-the-noun" puzzle.

<grin>

-- Katy

--
Katy or Rich Mulvey
r...@vectorbd.com

### Felix Lee

Dec 1, 1994, 6:35:31 AM12/1/94
to
Katy Mulvey:

>No, no, it's a "guess-the-noun" puzzle.

nah. It's a "guess-the-puzzle" puzzle.
--

### Felix Lee

Dec 1, 1994, 8:36:28 AM12/1/94
to
Greg Ewing:

> Consider this scenario: A friend gives you an object made
> from pieces of wire bent into strange shapes, with a loop
> of string entwined among them. He says, "Remove the string."

Actually, I always carry a pocket knife for just that eventuality.
Whenever someone hands me a string-and-wire puzzle, the first thing I
do is cut the string. I also carry a permanent black marker for
Rubik's-Cube-type puzzles, and a butane torch for those pesky wooden
burrs. These work great on crosswords too. Noone's given me any
really challenging puzzles for a long time.

By the way, I solved "+=3" by pushing the troll off the bridge.
Unfortunately, the program didn't understand my solution and wouldn't
give me credit for it. So I hired a hit man to take care of Dave.

:)
--

### axl119

Dec 1, 1994, 10:17:10 PM12/1/94
to
Well, just to barge in where I'm not known...

I'd suggest that the problem lies a little differently than the
consensus seems to be heading to. The coathanger/Gordian Knot bit was
a great metaphor, but like all metaphors was slightly flawed.

Someone gave me a puzzle like that once, except it looked fairly
professional. I worked at it for over an hour (on a bus ride), and
then, while my friend was in the bathroom, untied the knot holding the
cord onto the linked horseshoes. (Yes, it looked strange.) When he got
back, I told him I'd completed it. He said he bet I'd just untied it.
I denied it vigorously, at which point he pointed out that that was the
only way 'unless you can bend horseshoes with your bare hands'.

Well, I solved it, right? I didn't think it was the RIGHT way to solve
it, but I knew it was POSSIBLE to untie the cord. I mean, it had a
knot in it, this is Real Life, I can untie the cord. And, if (somehow)
this puzzle happened in real life, or in a virtual reality that was
flexible enough to truly react as such... well, my clothes wouldn't be
the first thing I thought of, but they'd be there.

Whereas, given *common* interactive fiction assumptions, you wouldn't
imagine that it WAS possible to remove your clothes, if they weren't
ever mentioned. There are certain common assumptions about what is and
isn't manipulatable in i.f., and unmentioned clothes are just in the
'not manipulatable' class. If you can point out a single instance to
the contrary... well, then congratulations.

By the definition of 'logic' et al that was given, it is the process of
deriving from ASSUMED premises... assumed premises based on observation
about how the world works. If you observe that you can't ever
manipulate clothing that is never mentioned in I.F., it becomes an
'assumed premise', and a perfectly valid one so far as I can see.

Now, the poster of the definitions goes on to say the following:

>Notice that these definitions show that nothing about a logical puzzle

>must be explicitly stated before encountering the puzzle (assumed premises
>are valid), as long as the puzzle can be solved by a somewhat probable
>solution. In the case of 3+=, The solution is probable if you assume
>enough things. The thing that makes this puzzle hard is knowing what to
>assume, not whether the puzzle is logical.

See... a 'somewhat probable solution'. Now take as some of your basic
assumptions the following:

1) I am solving an interactive fiction puzzle
2) Interactive fiction clothing is not manipulatable unless described
in some
manner, however briefly.

Where is your 'somewhat probable' solution? Is there any indication
that they aren't valid assumptions, aside from a total inability to
find any other solution? No. Therefore, it is 'logical' to conclude
that the answer lieth not in the clothing.

Let me speculate a little further. Suppose the puzzle were the same,
but the person was wearing... say, a kilt, a sporran, and a cloak. And
each article of clothing had to be specified by name. Still a
perfectly logical puzzle, by your definition... you just had to work on
the assumption that the main character was Scottish. (It *ALSO*
wouldn't be fair, as there are many people out there who don't even
know what a sporran IS, but that's another issue entirely.)

Now, most people wouldn't work on this assumption, just like most
fairly experienced IF gamers wouldn't work on the assumption that you
can manipulate nonexistant clothing in interactive fiction. So is
making the character a Scotsman perfectly logical? How about an
ancient Etruscan? (Anyone know what they wore?)

What I think should correctly have been claimed was 'If you work only
with the assumptions that I want you to have, this is a logical
puzzle.' Now, that's fine to say, but it's true of ANY puzzle. If I
write a puzzle that involves calling a cow 'fred', and I assume that
all cows are named 'fred', it's a perfectly logical puzzle, whether I
mention it anywhere or not. It's your problem that you don't share my
assumption. If that's what you think... well, I guess you could make
an argument for it. But that's not what I'd offhand say they meant
when they defined 'logical'.

Hmm.

I have the feeling I had a perfectly good argument there and I screwed
it up by shooting a cow and a Scotsman into it. Well, whatever. Make
of this what you will, I'm going to bed early today. I desparately
need it.