2 quick Crawl questions

On Sep 19, 5:11 am, JPEG <erina....TakeThisOut@hotmail.de> wrote:
> On Sep 19, 10:05 am, "Kenneth 'Bessarion' Boyd" <zaim....TakeThisOut@zaimoni.com>
> wrote:
>
> > On Sep 18, 12:51 pm, David Ploog <pl....TakeThisOut@mi.fu-berlin.de> wrote:
>
> > > When I encounter dangerous monsters, I just
> > > use full LOS exclusions, but I also go with autoexplore.
>
> > The exclusions have to be set before seeing the monsters, for this to
> > work.
>
> I'm sorry, I never understood that argument, so please enlighten me:
> What is the difference between trawling along by hand and enter a
> monster's los and between autoexplore doing the same? Either way, you
> only know a monster is there once you've seen it.

The difference currently, as illustrated in the diagram series I just
posted, is that there are at least eight mechanical cases (of which
two are diagrammed) where trawling around by hand can completely
prevent given the monster a free move by entering its LOS at range 1.

If that monster is a hydra in the midgame, that means autoexplore just
instakilled @. I was fortunate the first time this happened to me
(probability of instadeath from autoexploring into that O ambush was
about 10% on DL5).

> I know that there's one advantage to manual exploration that
> autoexplore will never be able to fill: the player can recognize
> minivault layout and try to avoid tricky areas before he is ready.

Which is fine.

On Sep 20, 6:39 pm, "Kenneth 'Bessarion' Boyd" <zaim....RemoveThis@zaimoni.com>
wrote:
> On Sep 19, 5:11 am, JPEG <erina....RemoveThis@hotmail.de> wrote:
> > I'm sorry, I never understood that argument, so please enlighten me:
> > What is the difference between trawling along by hand and enter a
> > monster's los and between autoexplore doing the same? Either way, you
> > only know a monster is there once you've seen it.
>
> The difference currently, as illustrated in the diagram series I just
> posted, is that there are at least eight mechanical cases (of which
> two are diagrammed) where trawling around by hand can completely
> prevent given the monster a free move by entering its LOS at range 1.

Thanks a lot for taking the time to explain! Those diagrams were
really helpful to finally understand what this is all about, and I'm
currently in the process of tweaking the travel code to take that into
account. All I still need to decide is whether this needs an option or
not, seeing how the old travel uses fewer turns. In any case, I do
think that the new one should be the default.

Might I please request that next time if you encounter a similar
situation where you apparently already know the solution, you don't
simply say that something is "broken" and leave us wondering what you
could possibly mean? Hoping for a fix is fairly pointless, after all,
as long as we don't know there is a problem in the first place. And
yes, it might seem obvious to you, but apparently it wasn't obvious to
us.
Rather, stuff like that is exactly what our feature requests tracker
(http://sourceforge.net/tracker/?group_id=143991&atid=757516) is there
for.

Thank you for the suggestion and subsequent explanations,

Johanna

On Sep 20, 9:40 pm, David Ploog <pl....RemoveThis@mi.fu-berlin.de> wrote:
> On Sun, 20 Sep 2009, Kenneth 'Bessarion' Boyd wrote:
> > What I want (and am considering a patch for) is a very simple scan of
> > the known map, Manhattan/L1 radius 2, that inspects whether a proposed
> > pathfinder move exposes @ to a melee ambush and veers if possible.
>
> I think this is the rule I just requested. But there are many maps that
> demand diagonal movement. What would you do then? Print a warning:
> "Further exploration is only possible at higher risk."?

No prompts, please! A prompt would serve the same purpose as simply
stopping whenever that situation came up but be more annoying.
For the record, I added a short section to the code with the following
behaviour.

..tx
?#@

t is the next step along our travel path (we only ever look at the
next step) and ? is a grid that is currently unseen (either unexplored
or an unseen non-wall grid).
If t has an adjacent unseen grid, it is changed to another grid on the
same orthogonal as @ (in this case, x). If x cannot be entered
(unpassable feature or excluded) t remains unchanged, danger and all.
Otherwise, x is used instead.

This behaviour should be documented somewhere ("trying to avoid
ambushes", if you will) and that's where we'll also want to mention
that sometimes only diagonal moves are possible (the checkerboard
vaults come to mind), so you cannot be entirely safe. But keep in mind
that the intent is not to guarantee completely safe travel but to try
to make autoexplore behave in a way the player would also move
manually. Faced with diagonals only, manual movement is no more safe
than automatic movement.

Johanna

On Sun, 20 Sep 2009, Kenneth 'Bessarion' Boyd wrote:

[dangers of autoexplore]

> You don't understand the original diagram either

[this is true]

> which illustrates that the *pathfinder* is broken. Fix the pathfinder,
> and autoexplore is automatically fixed: it heads to the nearest
> unexplored square without walking into avoidable melee ambushes.

Can you provide a simple rule the pathfinder should use to avoid that
problem?

David

On Sun, 20 Sep 2009, Kenneth 'Bessarion' Boyd wrote:

> What I want (and am considering a patch for) is a very simple scan of
> the known map, Manhattan/L1 radius 2, that inspects whether a proposed
> pathfinder move exposes @ to a melee ambush and veers if possible.
>
> Don't worry about diagonals, you have no choice but to risk instadeath
> ..#..
> ?#@#?

I think this is the rule I just requested. But there are many maps that
demand diagonal movement. What would you do then? Print a warning:
"Further exploration is only possible at higher risk."?

David

On Sun, 20 Sep 2009, JPEG wrote:
> On Sep 20, 9:40 pm, David Ploog <pl... DeleteThis @mi.fu-berlin.de> wrote:
>> On Sun, 20 Sep 2009, Kenneth 'Bessarion' Boyd wrote:
>>> What I want (and am considering a patch for) is a very simple scan of
>>> the known map, Manhattan/L1 radius 2, that inspects whether a proposed
>>> pathfinder move exposes @ to a melee ambush and veers if possible.
>>
>> I think this is the rule I just requested. But there are many maps that
>> demand diagonal movement. What would you do then? Print a warning:
>> "Further exploration is only possible at higher risk."?
>
> No prompts, please! A prompt would serve the same purpose as simply
> stopping whenever that situation came up but be more annoying.

Sure. The question about the prompt was rhetorical. However, what could
happen is this:

....
.t#.
?#@.

Here, t is risky, but it could be safer to walk around (going north). Of
course, that's an optimization game we won't even try to solve. Just
saying that the story does not really end here.

I like the change (thanks to Kenneth for the pictures) and I am against an
option.

David

On Sep 20, 2:40 pm, David Ploog <pl....TakeThisOut@mi.fu-berlin.de> wrote:
> On Sun, 20 Sep 2009, Kenneth 'Bessarion' Boyd wrote:
> > What I want (and am considering a patch for) is a very simple scan of
> > the known map, Manhattan/L1 radius 2, that inspects whether a proposed
> > pathfinder move exposes @ to a melee ambush and veers if possible.
>
> > Don't worry about diagonals, you have no choice but to risk instadeath
> > ..#..
> > ?#@#?
>
> I think this is the rule I just requested. But there are many maps that
> demand diagonal movement. What would you do then?

I'm fine with the pathfinder taking unavoidable risks for me. This
particular example isn't worth stopping and asking me for a rote
decision for.

On 2009-09-18, David Ploog <ploog.TakeThisOut@mi.fu-berlin.de> wrote:
>>> r/c/g?
>>
>> race/class/god
>
> We say species/job/god here.

We do? Maybe you do, but I don't think I've ever heard a player say
"job."

-r.

On Mon, 21 Sep 2009, Rachel Elizabeth Dillon wrote:
> On 2009-09-18, David Ploog <ploog.RemoveThis@mi.fu-berlin.de> wrote:

>>>> r/c/g?
>>>
>>> race/class/god
>>
>> We say species/job/god here.
>
> We do? Maybe you do, but I don't think I've ever heard a player say
> "job."

I am trying to pave the way for the change in terminology. It has been
said time and again how use of "class" is misleading. I agree that "job"
is not ideal but there haven't been better proposals as of yet.

David

On 2009-09-21, David Ploog <ploog.TakeThisOut@mi.fu-berlin.de> wrote:
> On Mon, 21 Sep 2009, Rachel Elizabeth Dillon wrote:
>> On 2009-09-18, David Ploog <ploog.TakeThisOut@mi.fu-berlin.de> wrote:
>
>>>>> r/c/g?
>>>>
>>>> race/class/god
>>>
>>> We say species/job/god here.
>>
>> We do? Maybe you do, but I don't think I've ever heard a player say
>> "job."
>
> I am trying to pave the way for the change in terminology. It has been
> said time and again how use of "class" is misleading. I agree that "job"
> is not ideal but there haven't been better proposals as of yet.

The ##crawl bot says "role" (and "race").

Cheers
Robert

On Sep 21, 6:10 pm, David Ploog <pl....DeleteThis@mi.fu-berlin.de> wrote:
> On Mon, 21 Sep 2009, Rachel Elizabeth Dillon wrote:
> > On 2009-09-18, David Ploog <pl....DeleteThis@mi.fu-berlin.de> wrote:
> >>>> r/c/g?
>
> >>> race/class/god
>
> >> We say species/job/god here.
>
> > We do? Maybe you do, but I don't think I've ever heard a player say
> > "job."
>
> I am trying to pave the way for the change in terminology. It has been
> said time and again how use of "class" is misleading. I agree that "job"
> is not ideal but there haven't been better proposals as of yet.

Well, there's profession, career and background, for example. The
trouble comes from finding a term (and "job" doesn't really fit
either) that can be used to refer to the large scale between Wizard
and Wanderer, Artificer and Berserker. Come to think of it, yes, even
role would be better than job.

jpeg

On 2009-09-21, JPEG <erinacea.DeleteThis@hotmail.de> wrote:
> On Sep 21, 6:10 pm, David Ploog <pl....DeleteThis@mi.fu-berlin.de> wrote:
>> On Mon, 21 Sep 2009, Rachel Elizabeth Dillon wrote:
>> > On 2009-09-18, David Ploog <pl....DeleteThis@mi.fu-berlin.de> wrote:
>> >>>> r/c/g?
>> >>> race/class/god
>> >> We say species/job/god here.
>> > We do? Maybe you do, but I don't think I've ever heard a player say
>> > "job."
>> I am trying to pave the way for the change in terminology. It has been
>> said time and again how use of "class" is misleading. I agree that "job"
>> is not ideal but there haven't been better proposals as of yet.
> Well, there's profession, career and background, for example. The
> trouble comes from finding a term (and "job" doesn't really fit
> either) that can be used to refer to the large scale between Wizard
> and Wanderer, Artificer and Berserker. Come to think of it, yes, even
> role would be better than job.

How about kit or template.
As only affects are on initial item/skill layout.

In article <h906tm$tpo$1@news.ryerson.ca>, jte <jengbrec.DeleteThis@ryerson.ca> wrote:
>
>also: how am I supposed to interpret the gold value listed on a
>ziggurat entrance? Is it part of the mystery of the ziggurat?

It's the entry price for the ziggurat: it will cost you that much to
enter.

--
Mark Mackey
The Association for the Advancement of Dungeon Crawling
Hints, tips and spoilers
http://www.swallowtail.org/crawl/