Protection Racket Efficiency Challenge
Derek Ray
For the pro-PR proponents, I posit the postfixed proposal*:
__ Stop telling us why it should work; just show us that it does. __
I know, that's a revolutionary concept around RGRN, where "much
discussion, little action" is often the order of the day, but you guys
can beat each other over the head with anecdotal corner-cases all
day long and never get anywhere.
The devnull trophies are an excellent measurement tool; let's use that
as our yardstick. If the protection racket is really as consistent and
reliable as just playing normally, then attaining any of the trophies
should be quite possible using ONLY characters who are obeying the
Rules of the Protection Racket.
I further propose these rules be defined as:
* The player's primary objective shall be to proceed to Minetown and
obtain a minimum of 7 points of divine protection from the priest
prior to proceeding onward towards ascension.
* The player's experience level shall not exceed 3 prior to obtaining
the divine protection, or the current game shall be deemed a failure.
* The player shall not intentionally descend below the Sokoban level
in the main dungeon branch until the condition of divine protection
has been satisfied. If circumstances drop the player below Sokoban
(hole trap, level teleporter), he must make immediate effort to return
above the Sokoban level.
These should satisfy the concept of what a "protection racket" is quite
adequately, even allowing for variations in the details (some prefer
XL1, some say XL2 or XL3 are OK) and unexpected circumstance (lost pet).
I challenge the protection racket proponents to, under these rules, play
13 games on a public vanilla server (nethack.alt.org == ideal) and reveal
their results, as well as which devnull trophies they would've won as a
result of those 13 games. (Birdie, Double Top, Hat Trick, Grand Slam,
etc.)
Is anyone willing to take this challenge up? Or is the protection
racket one of those hypotheses that only works on paper, not in practice?
--
Derek
* Yes, I know that's shoddy grammar. It's intended to be humorous
alliteration, not an epic work of literature.
XyZZy
wrote:
> * The player shall not intentionally descend below the Sokoban level
> in the main dungeon branch until the condition of divine protection
> has been satisfied. If circumstances drop the player below Sokoban
> (hole trap, level teleporter), he must make immediate effort to return
> above the Sokoban level.
for no reason than my sheer curiosity, why this rule?
Derek Ray
To help maintain the "proceed to minetown directly" concept. I'm
undecided at this point whether "doing Sokoban first" should be
considered as a true racket -- but all that can be up for discussion,
since this is just the initial proposal
The most important part would be that all parties agree to the
conditions before any games were played, if someone wishes to accept.
If that includes "do soko first, then the mines", so be it; but that
should clearly be spelled out, too.
--
Derek
Game info and change log: http://sporkhack.com
Beta Server: telnet://sporkhack.com
IRC: irc.freenode.net, #sporkhack
Carl
wrote:
I'm willing to give it a try just to add another data point to the
experiment, though I am not a "proponent" of the protection racket.
But, does the game have to go on to completion? That's more
experiment than I'm able to take on at the moment. I wonder if there
is a different completion milestone that could be established. For
example, since the Protection Racket is intended to get one safely to
the Mid-game, could success of the Protection Racket be defined as
surviving to a set point, such as arrival on Quest Level 2?
sjde...@yahoo.com
wrote:
I'd say drop point #3 and stipulate "no level draining". Seems fine
for a racketeer to run around looking for money as much as they want,
but it's not in the spirit to level up normally then go drain levels
and buy protection.
Derek Ray
> On Jan 21, 2:09 pm, Derek Ray <de...@moot.its.only.a.spamtrap.org>
>> the divine protection, or the current game shall be deemed a failure.
>
This is actually handled by point #2 (quoted above) -- if you go to XL4
and come back to XL3 via drain, you've exceeded 3 and it's a failure.
> Seems fine for a racketeer to run around looking for money as much as
> they want,
Well, that's why point #3 exists. There's a certain point at which
you're stepping well outside the idea of "Go to Minetown at low level
and grab protection -- because it's safer to have divine protection when
you go deep into the dungeon."
If you're running around deep in the dungeon without protection _or_
having gained levels normally, you're demonstrating that the protection
racket is ineffective -- since you didn't need the protection even
though you were at low XL.
Technically that works out in favor of my side, but I'm more interested
in an actual set of data points here from play as opposed to "winning"
one way or the other. If it's possible, then it's possible, and the
most convincing evidence will _always_ be "It was done." -- witness the
11-conduct ascension, yes?
> but it's not in the spirit to level up normally then go drain levels
> and buy protection.
Agreed. Hence point #2. :)
XyZZy
i'm wondering about this one myself. personally, by that point in
games i have ascended, i am untouchable,and the rest of the game is
pretty much a mindless run, with occasionally checking for grave
statuses (ill/stoning). i can see why you would want complete
ascensions (for the trophies), but at the same time, i (as well as
others) don't have the kind of time to dedicate to this. Getting a
character to a point where things are simply going through the motions
is a much quicker game i'd be glad to try, but i doubt i have time to
even run 13 non half-hearted games in a month.
(on a side note, while it certainly doesn't follow these rules, i'm
probably pretty much the poster child FOR protection racketeering. my
first ascension took like 70 tries, the last 20-30 of which were
protection rackets. the next ascension took only 15 tries, and this
one i'm finishing right now/hopefully submitting for this "constest"
as the first of 13, only took 2 (with the first try actually getting
well past the protection racket stage, but lag + stupidity means about
100 turns of nurses bashing me while i wield a weapon and have no
armor). "ilya86" account on alt.org, if you want to change the
dumplogs. the characters haven't exactly been properly fit for
protection racketeering either, other than the first one (tourist).
the second was rogue, and this one is a knight.) but again, this
probably falls well within your "corner cases" so i doubt it means
much.
David Damerell
>(on a side note, while it certainly doesn't follow these rules, i'm
>probably pretty much the poster child FOR protection racketeering. my
>first ascension took like 70 tries, the last 20-30 of which were
>protection rackets. the next ascension took only 15 tries, and this
>one i'm finishing right now/hopefully submitting for this "constest"
>as the first of 13, only took 2
This is not a particularly remarkable pattern of improvement, PR or no PR.
--
David Damerell <dame...@chiark.greenend.org.uk> flcl?
Today is Sunday, January - a weekend.
XyZZy
wrote:
> Quoting XyZZy <EliminateCali...@gmail.com>:
>
> >(on a side note, while it certainly doesn't follow these rules, i'm
> >probably pretty much the poster child FOR protection racketeering. my
> >first ascension took like 70 tries, the last 20-30 of which were
> >protection rackets. the next ascension took only 15 tries, and this
> >one i'm finishing right now/hopefully submitting for this "constest"
> >as the first of 13, only took 2
>
> This is not a particularly remarkable pattern of improvement, PR or no PR.
> --
> Today is Sunday, January - a weekend.
i never really said it was remarkable, although would you agree with
me it's average-ish, correct? thus showing, that (at least in one
case) PR works as a strategy for an average player. which is probably
what i really meant to say, sorry if it came off self-aggrandizing.
the games themselves are QUITE unremarkable, downright newbish in the
amount of over-preparation.
David Damerell
>On Jan 23, 3:22 am, David Damerell <damer...@chiark.greenend.org.uk>
>>>(on a side note, while it certainly doesn't follow these rules, i'm
>>>probably pretty much the poster child FOR protection racketeering. my
>>>first ascension took like 70 tries, the last 20-30 of which were
>>>protection rackets. the next ascension took only 15 tries, and this
>>>one i'm finishing right now/hopefully submitting for this "constest"
>>>as the first of 13, only took 2
>>This is not a particularly remarkable pattern of improvement, PR or no PR.
>me it's average-ish, correct?
The variation is huge, but it doesn't seem unusual.
>thus showing, that (at least in one case) PR works as a strategy for
>an average player.
I don't think it's in dispute that the PR _can_ work - merely if it's the
best strategy.
--
David Damerell <dame...@chiark.greenend.org.uk> flcl?
Jakob Creutzig
> I don't think it's in dispute that the PR _can_ work - merely if it's the
> best strategy.
Depends on the personality, I reckon. I studied marvin's moves
and in principle I believe I understood almost all tactical and
strategic issues. I survive much better with PR, however, because
I lack patience for proper implementation of them.
A perfect player should have better odds without PR, anyway;
but some people's defects will make PR attractive for them.
Cheers,
Jakob
David Damerell
>David Damerell <dame...@chiark.greenend.org.uk> writes:
>>I don't think it's in dispute that the PR _can_ work - merely if it's the
>>best strategy.
>Depends on the personality, I reckon. I studied marvin's moves
>and in principle I believe I understood almost all tactical and
>strategic issues. I survive much better with PR, however, because
>I lack patience for proper implementation of them.
I don't think that's a sensible way to compare strategies. Banging your
head on the keyboard is the best strategy if you have a personality that
abhors typing, for example.
Aki Lumme
"Derek Ray" <de...@moot.its.only.a.spamtrap.org> kirjoitti
viestissä:slrnfp9rf5...@still.just.a.spamtrap.org...
> I further propose these rules be defined as:
Actually, these rules seem to be so definite that the game
could check them, and if successful, the player would
gain the title "racketeer", carrying it along to the
high score table.
Zork the racketeer, killed by a soldier ant
Or even more cruelly:
Zork the racketeer-wannabe, killed by a newt
Of course, this sort of thinking leads into some other
titles as well:
Zork the farmer (>1000 puddings killed)
etc.
AL
3cn...@gmail.com
many of the pieces necessary for a smooth run. Namely,
-1. LARGE DOG/CAT, preferably with more pets to spare.
You will *need* a large dog/cat. You will need to keep your large dog/
cat as far away from the Minetown watch as possible. You must not
enter the mines unless your pet is in the second stage, and only then
in an emergency. A large dog/cat can kill zap-happy dwarves and gnomes
before they can zap you. He can lift an entire chest out of a store,
or a bag laden with valuable items. He will be your best friend. Get a
leash or magic whistle if you can, and carry a few tripe rations or
eggs on your person to keep your pet close.
0. FOOD.
The protection run takes around 3000 turns on a good day. You will
*need* a good supply of food rations to succeed. Moreover, even with
food,
it's useful to stay close to your pet to snatch corpses from under his
nose, at least when he's fighting weak stuff like jackals and coyotes.
1. pick-axe.
In Vanilla nethack, gold vaults are a near-foolproof way to get money.
You will need the pick-axe. Moreover, there's been some confusion
about how the pick-axe is used to "dig downward" in the mines. Here's
how: you flee to the edge of the map, wait for your dog to kill most
of the monsters close by, dig two squares into the wall, displace your
dog so he's trapped in the newly-dug tunnel behind you, [E]ngrave
several Elbereths, and dig. This setup minimizes the chances of your
dog falling through a trap door, because you're not exploring much of
the mines.
2. 17 HP.
You *cannot* reliably finish the protection run with fewer than 17 HP.
One stray zap from a wand of striking and you're a bones file. If you
have positive alignment and are low on HP, with max HP lower than 19,
your god will always donate some "spare" HP. Using this trick and
gaining EXP:2, you can easily make it to 25 HP or more, and the run
will be a breeze. The more HP you have, the less risky the run is,
period.
3. Credit cloning.
In Sporkhack, vaults are a Bad Idea because they're infested with
*giant eels*. If I'm doing the run in Spork, I avoid the vaults.
Period. I credit-clone from shops instead. A mere sack works wonders
for credit cloning. Otherwise, I'll lug around the single most
valuable item I've found, usually a high-level spellbook or an early
ID'd gem.
4. light source.
A light source is a luxury during the run. It's very useful to see
exactly where your dog is and what he's fighting. And if he falls
through a trap door, you'll know so you can start "damage control",
chucking daggers at some nearby gnomes to gain XP and prematurely end
the run.
I don't think the protection run is a good strategy in general. I use
it when I feel the circumstances are right, i.e I have plenty of food,
a good source of cash, and an early pick-axe from the first level of
the mines. There's *no* downside to exploring the first few dungeon
levels without gaining substantial XP. In fact, since you'll get
weaker monsters (no **** floating eyes and only a few gas spores),
staying inexperienced levels your pet much more quickly than usual,
and allows you a good supply of jackal and lichen corpses in lieu of a
better food source.
Right now, I'm experimenting with a rogue named Run2 on the Spork
server - a *rogue*, who starts with no food, no pick-axe, and weak HP.
His strategy is to level up to EXP:2 or 3 and credit-clone extra gold.
While his first two games were flukes - killing a gas spore by
accident and getting hit by a poisoned dart from a kobold - his third
game managed 7 units protection despite horrible circumstances, and
might be a fun ttyrec to watch.
There were no stores outside the mines, and but a single vault which I
later found to contain a giant eel, but Run2 bravely descended to a
bones-laden minetown complete with a wandering white unicorn, rock
mole, floating eye and nymph, and credit cloned around 6000 gold
pieces at EXP:3 (was going for EXP:2 but accidentally killed a gnome,
damn it). His large dog bit the dust to a giant mimic after 6 units
protection, sadly, but Run2 then found a bones file near the upstairs
with 1200 gold pieces, allowing him to buy the final unit at EXP:4.
Woo-hoo!
Right now, Run2 is living a happy, healthy Sporkish life, having
gained +5 plate mail and a +5 short sword from the generosity of Kos,
and an early pair of speed boots. I'm still unsure of whether I'll
take up Derek's challenge, but hope this post - and Run2's latest game
- give some healthy food for thought.
XyZZy
> 0. FOOD.
> The protection run takes around 3000 turns on a good day. You will
> *need* a good supply of food rations to succeed. Moreover, even with
> food,
> it's useful to stay close to your pet to snatch corpses from under his
> nose, at least when he's fighting weak stuff like jackals and coyotes.
Prayer can generally take care of the food "problem" very nicely,
unless you're not eating ANY corpses or finding foods
> 1. pick-axe.
> In Vanilla nethack, gold vaults are a near-foolproof way to get money.
> You will need the pick-axe. Moreover, there's been some confusion
> about how the pick-axe is used to "dig downward" in the mines. Here's
> how: you flee to the edge of the map, wait for your dog to kill most
> of the monsters close by, dig two squares into the wall, displace your
> dog so he's trapped in the newly-dug tunnel behind you, [E]ngrave
> several Elbereths, and dig. This setup minimizes the chances of your
> dog falling through a trap door, because you're not exploring much of
> the mines.
I'd only do this on the more dangerous unlit levels. checking each
mine level until mintetown for at least the guaranteed tool can easily
net you that much neede magic whistle, or bag.
> 4. light source.
> A light source is a luxury during the run. It's very useful to see
> exactly where your dog is and what he's fighting. And if he falls
> through a trap door, you'll know so you can start "damage control",
> chucking daggers at some nearby gnomes to gain XP and prematurely end
> the run.
don't prematurely end the run if something like this happens. try
following the dog through the trap door, or make a (pacifistic) break
for minetown with what you've got. you'd be surprised how often you
can get away with things like this, and even recover (small cats/dogs
can be generated in even the highest minetown, and you should have the
food to tame them.
i also generally ADORE getting floating eyes on the run, because i can
kill one of them (17 XP means you won't gain even one level) and get
telepathy which can really help assess danger levels upon entering a
level. just do what you can to keep your stupid pet the hell away from
them, unless you like hitting "n100." in hopes of them reviving.
Aki Lumme
children. The floor was very crowded ... four of the girls were
in a room by themselves at the back of the house. They were all
huddled up together, and seemed frightened. The defendants were
in the front part of the house. The girls at the back part of the
house could not have got out without passing through the room
where the defendants were. This house has been known to me for a
long time as one where young girls were kept to be shipped off to
California."
A watch-repairer and jeweler who had resided opposite this place
for three or four years declared that he knew the first defendant,
A-Neung, and that she had lived there some years, on the first
floor; that he had seen a number of girls going in and out of
the house, seeming to arrive by steamer, some in chairs and some
walking, and that he knew from what he had seen of her and the
girls that she was a buyer and seller of girls. A carpenter living
below in the same house deposed: "I have always seen a number of
young girls being taken in and out of the house. The age of the
girls ranged from 10 to 20 years. There was always a great deal of
crying and groaning amongst the girls up-stairs. I have not heard
any beating, but the girls were constantly crying. The crying was
a
Derek Ray
> On Jan 21, 11:09 am, Derek Ray <de...@moot.its.only.a.spamtrap.org>
>> 13 games on a public vanilla server (nethack.alt.org == ideal) and reveal
>> their results, as well as which devnull trophies they would've won as a
>> result of those 13 games. (Birdie, Double Top, Hat Trick, Grand Slam,
>> etc.)
>
> I'm willing to give it a try just to add another data point to the
> experiment, though I am not a "proponent" of the protection racket.
> But, does the game have to go on to completion?
Sorry for the delay in responding; I wanted to think about this a bit
(and, frankly, I've actually been somewhat busy :).
While I understand the point made (there's a point beyond which the
protection racket is very little influence), I think that yes, the game
should in fact go on to completion.
Partly, this is to adhere strictly to the manner in which real games would
be played; any deviation could potentially raise doubt about the
results, such as "sure, he got to the Quest, but he couldn't have ascended
from there, he burned up all his stuff." If all games finish in
ascension or death, no quibbling is available.
Partly, if you know you only have to make it to the Quest, you aren't
necessarily playing for the long-term and this opens up a number of
potential loopholes (once you get protection, DfV, bang the drawbridge
to reach XL14, and use the wand to wish for enough teleport tools to make
it back to QL2. This isn't an "ascension" approach, it's an approach to
deliberately game the mechanics of the challenge. There are surely
other such approaches I haven't though of yet.)
And partly, I actually _do_ think the results could be different; if
you play 13 short PR games in a row and #quit immediately upon reaching
a predefined point, you don't have to spend any mental time or
effort on the rest of the game -- which (as i'm sure you're aware) is
not particularly insignificant, _and_ bears absolutely no relation to
the tactics used for the PR.
This is part of why even the Hat Trick with Bells On is a particularly
good achievement; the player held it all together and didn't drop it on
his foot for _five consecutive games_, no breaks in concentration.
Removing the latter half to two-thirds of the game reduces the
concentration required and effort expended; I think this would distort
the results.
So realistically, no, I think all the games must be fully completed.
By the same token, theoretically all these games should take place within
a single month, but I worry that if I attempt to set that condition that
nobody will even try -- or, possibly, someone would wait until June to
try. (Which isn't such a bad plan.)
Derek Ray
> Quoting XyZZy <Eliminat...@gmail.com>:
>>thus showing, that (at least in one case) PR works as a strategy for
>>an average player.
>
> I don't think it's in dispute that the PR _can_ work - merely if it's the
> best strategy.
I am directly challenging the assertion that the protection racket is as
consistent a path to ascension as simply playing normally.
We know "playing normally" is a consistent path from the performances in
devnull, individual player streaks of 20+ ascensions, etc. If the PR
really is as consistent, a PR player should be able to measure himself
against the devnull trophies effectively.
(Note that the devnull trophies might also require you to use more than
one class...)
Carl
> This has long passed the point of coherent discussion.
>
> For the pro-PR proponents, I posit the postfixed proposal*:
>
> __ Stop telling us why it should work; just show us that it does. __
>
> I know, that's a revolutionary concept around RGRN, where "much
> discussion, little action" is often the order of the day, but you guys
> can beat each other over the head with anecdotal corner-cases all
> day long and never get anywhere.
>
> The devnull trophies are an excellent measurement tool; let's use that
> as our yardstick. If the protection racket is really as consistent and
> reliable as just playing normally, then attaining any of the trophies
> should be quite possible using ONLY characters who are obeying the
> Rules of the Protection Racket.
>
> I further propose these rules be defined as:
>
> * The player's primary objective shall be to proceed to Minetown and
> obtain a minimum of 7 points of divine protection from the priest
> prior to proceeding onward towards ascension.
A suggestion for rewording would be: "player's primary objective shall
be to proceed to a temple and obtain a minimum of 7 points of divine
protection from the priest prior to proceeding onwards toward
ascension." After all, if a temple is found on level 9 or 10 there is
no reason to hazard the dangers of the mines.
>
> * The player's experience level shall not exceed 3 prior to obtaining
> the divine protection, or the current game shall be deemed a failure.
>
> * The player shall not intentionally descend below the Sokoban level
> in the main dungeon branch until the condition of divine protection
> has been satisfied. If circumstances drop the player below Sokoban
> (hole trap, level teleporter), he must make immediate effort to return
> above the Sokoban level.
>
I don't think this rule makes so much sense. If I find the downstairs
on the level below the Sokoban entrance then I am certainly going to
go downstairs just to check them out and maybe grab some more loot.
Especially if it looks like the coast is clear, and then what if I
hear that there is a shop or vault on that level? Or, what if I see
yet another downstair and it looks like the coast is clear yet again?
Certainly it makes sense to take a quick peak and see if there is some
gold. Sneaking around looking for gold does not necessarily imply not
needing protection, but if gold on lower levels is sufficiently easy
to grab-and-go, then why not grab-it-and-go?
Carl
wrote:
> On 2008-01-22, Carl <carlhnel...@gmail.com> wrote:
>
> > On Jan 21, 11:09 am, Derek Ray <de...@moot.its.only.a.spamtrap.org>
> >> I challenge the protection racket proponents to, under these rules, play
> >> 13 games on a public vanilla server (nethack.alt.org == ideal) and reveal
> >> their results, as well as which devnull trophies they would've won as a
> >> result of those 13 games. (Birdie, Double Top, Hat Trick, Grand Slam,
> >> etc.)
>
> > I'm willing to give it a try just to add another data point to the
> > experiment, though I am not a "proponent" of the protection racket.
> > But, does the game have to go on to completion?
>
> Sorry for the delay in responding; I wanted to think about this a bit
> (and, frankly, I've actually been somewhat busy :).
>
> While I understand the point made (there's a point beyond which the
> protection racket is very little influence), I think that yes, the game
> should in fact go on to completion.
>
> Partly, this is to adhere strictly to the manner in which real games would
> be played; any deviation could potentially raise doubt about the
> results, such as "sure, he got to the Quest, but he couldn't have ascended
> from there, he burned up all his stuff." If all games finish in
> ascension or death, no quibbling is available.
Well, I still think you're measuring the whole game unnecessarily. As
for burning up all of the resources, I don't think that's particularly
relevant; there's always more gear to be found in the dungeon.
Though, one challenge of having a non-ascension ending would of course
being able to easily tally success.
>
> Partly, if you know you only have to make it to the Quest, you aren't
> necessarily playing for the long-term and this opens up a number of
> potential loopholes (once you get protection, DfV, bang the drawbridge
> to reach XL14, and use the wand to wish for enough teleport tools to make
> it back to QL2. This isn't an "ascension" approach, it's an approach to
> deliberately game the mechanics of the challenge. There are surely
> other such approaches I haven't though of yet.)
>
> And partly, I actually _do_ think the results could be different; if
> you play 13 short PR games in a row and #quit immediately upon reaching
> a predefined point, you don't have to spend any mental time or
> effort on the rest of the game -- which (as i'm sure you're aware) is
> not particularly insignificant, _and_ bears absolutely no relation to
> the tactics used for the PR.
Well if it bears no relation to the tactics used for the PR, why
should it bear any relation to this experiment? The Nethack endgame
for a veteran player is hardly an exercise in skill; it's an exercise
in patience.
>
> This is part of why even the Hat Trick with Bells On is a particularly
> good achievement; the player held it all together and didn't drop it on
> his foot for _five consecutive games_, no breaks in concentration.
> Removing the latter half to two-thirds of the game reduces the
> concentration required and effort expended; I think this would distort
> the results.
>
> So realistically, no, I think all the games must be fully completed.
> By the same token, theoretically all these games should take place within
> a single month, but I worry that if I attempt to set that condition that
> nobody will even try -- or, possibly, someone would wait until June to
> try. (Which isn't such a bad plan.)
>
Actually June wouldn't work for me, I've got other things I want to
work on then :)
Hmm, one month to do 13 games ... no can do for me. How about instead
we just set the challenge for ending on May 31?
By the way, is anyone else out there going to give it a try? And
Derek, what about you?
XyZZy
i'd totally go for it, but with the "month limit" stipulation, and
lack of time (college) i may as well just go for it in the devnull
tournament. at least then i'd get real trophies :P.
Henry J Cobb
> I agree with you Derek. Typically my breakpoint is DL5. The objective in
> the upper dungeon is to scout out the shops and to find the downstairs
> to the Mines. There's generally no need to go past the Oracle.
... because there's so little money at the end of Sokoban?
Can we exclude healers and wizards please or will there be per-class
competitions?
Uh, I'll take a gnome healer...
-HJC
XyZZy
the whole point of the trophies as a measuring scale is that getting
the more difficult ones REQUIRES all classes, not just the "easy"
ones. the 13-ascension run achieving the trophies requires you to play
all classes, never two ascensions of the same class. This in turn
shows that the protection racket is a very reilable strategy for all
classes, not just the ones "built for it". ascending one or two games
with a wizard or a healer and failing the rest will only prove the
point that the protection run is ineffecient for 11/13 classes. if
part of your run is a gnomish healer, neutral (or lawful, don't
remember which is better) wizard, then so be it, i hope you planned
your 13 runs out well, if you are using your gambit race on an already
easy character to do a racket with.
none Byron Jeff
I wasn't familiar with the tournament and its trophies. A description
for the trophies can be found here:
http://nethack.devnull.net/tournament/trophies.shtml
13 games?! I can barely find the patience to play one.
BAJ
none Byron Jeff
Carl <carlh...@gmail.com> wrote:
>On Jan 21, 11:09 am, Derek Ray <de...@moot.its.only.a.spamtrap.org>
>wrote:
>> This has long passed the point of coherent discussion.
>>
>> For the pro-PR proponents, I posit the postfixed proposal*:
>>
>> __ Stop telling us why it should work; just show us that it does. __
>>
>> I know, that's a revolutionary concept around RGRN, where "much
>> discussion, little action" is often the order of the day, but you guys
>> can beat each other over the head with anecdotal corner-cases all
>> day long and never get anywhere.
>>
>> The devnull trophies are an excellent measurement tool; let's use that
>> as our yardstick. If the protection racket is really as consistent and
>> reliable as just playing normally, then attaining any of the trophies
>> should be quite possible using ONLY characters who are obeying the
>> Rules of the Protection Racket.
>>
>> I further propose these rules be defined as:
>>
>> * The player's primary objective shall be to proceed to Minetown and
>> obtain a minimum of 7 points of divine protection from the priest
>> prior to proceeding onward towards ascension.
>
>A suggestion for rewording would be: "player's primary objective shall
>be to proceed to a temple and obtain a minimum of 7 points of divine
>protection from the priest prior to proceeding onwards toward
>ascension." After all, if a temple is found on level 9 or 10 there is
>no reason to hazard the dangers of the mines.
I disagree. The protection racket is specifically geared for the Minetown
priest(ess). This is for two reasons: that priest(ess) is guaranteed,
and Minetown is virtually guranteed to be generated within the first 8
levels from the top of the dungeon. Other temples have no such
guarantees.
I think Derek's definition is a perfect description of the process and
should be left unchanged.
>>
>> * The player's experience level shall not exceed 3 prior to obtaining
>> the divine protection, or the current game shall be deemed a failure.
>>
>> * The player shall not intentionally descend below the Sokoban level
>> in the main dungeon branch until the condition of divine protection
>> has been satisfied. If circumstances drop the player below Sokoban
>> (hole trap, level teleporter), he must make immediate effort to return
>> above the Sokoban level.
>>
>I don't think this rule makes so much sense. If I find the downstairs
>on the level below the Sokoban entrance then I am certainly going to
>go downstairs just to check them out and maybe grab some more loot.
>Especially if it looks like the coast is clear, and then what if I
>hear that there is a shop or vault on that level? Or, what if I see
>yet another downstair and it looks like the coast is clear yet again?
>Certainly it makes sense to take a quick peak and see if there is some
>gold. Sneaking around looking for gold does not necessarily imply not
>needing protection, but if gold on lower levels is sufficiently easy
>to grab-and-go, then why not grab-it-and-go?
Personally I think it's a moot point. It's a death trap down there at
low level, no pet, and no protection anyway. Maybe that rule can be
modified so that the player is limited to 1 level below their
circumstantial drop in the dungeon.
BAJ
Derek Ray
> On Jan 24, 8:32 am, Derek Ray <de...@moot.its.only.a.spamtrap.org>
>> > I'm willing to give it a try just to add another data point to the
>> > experiment, though I am not a "proponent" of the protection racket.
>> > But, does the game have to go on to completion?
>> protection racket is very little influence), I think that yes, the game
>> should in fact go on to completion.
>>
>> Partly, this is to adhere strictly to the manner in which real games would
>> be played; any deviation could potentially raise doubt about the
>> results, such as "sure, he got to the Quest, but he couldn't have ascended
>> from there, he burned up all his stuff." If all games finish in
>> ascension or death, no quibbling is available.
>
> Well, I still think you're measuring the whole game unnecessarily.
Past experience doing similar things has indicated that people will
quibble about _anything_ if they disagree with the results. I want
there to be as little interpretation of the results as necessary;
ideally, that would be "zero" interpretation, but that may or may not be
possible.
> for burning up all of the resources, I don't think that's particularly
> relevant; there's always more gear to be found in the dungeon.
I don't claim to be able to tell the future, so I can't know what
objections people might bring up. I prefer to remove the opportunity
for objections entirely.
> Though, one challenge of having a non-ascension ending would of course
> being able to easily tally success.
Well, ttyrecs and screenshots are not ideal but would suffice.
>> Partly, if you know you only have to make it to the Quest, you aren't
>> necessarily playing for the long-term and this opens up a number of
>> potential loopholes (once you get protection, DfV, bang the drawbridge
>> to reach XL14, and use the wand to wish for enough teleport tools to make
>> it back to QL2. This isn't an "ascension" approach, it's an approach to
>> deliberately game the mechanics of the challenge. There are surely
>> other such approaches I haven't though of yet.)
By the way, this is also a significant factor from _my_ perspective;
certainly much more so than the "resource burn" remark that I was mostly
bringing up just to try to find an example of something people would
gripe about.
>> And partly, I actually _do_ think the results could be different; if
>> you play 13 short PR games in a row and #quit immediately upon reaching
>> a predefined point, you don't have to spend any mental time or
>> effort on the rest of the game -- which (as i'm sure you're aware) is
>> not particularly insignificant, _and_ bears absolutely no relation to
>> the tactics used for the PR.
> Well if it bears no relation to the tactics used for the PR, why
> should it bear any relation to this experiment?
Because it forces the player to defocus on the PR skills and refocus on
actually completing the game -- which is exactly what he would be doing
if he were playing "for real".
As another (somewhat unrelated) example, which is easier -- to complete
World 1 of Super Mario World without losing a life 10 times in a row, or
to do it while _also_ completing the rest of the game right afterwards?
Obviously, the first way is easier; you can mentally discard anything
that doesn't relate to World 1. Now... which is more like the way
someone would normally play the game?
You see my point.
> The Nethack endgame for a veteran player is hardly an exercise in skill;
> it's an exercise in patience.
And the capacity for that patience is, in fact, part of the 'skill' of nethack.
>> Removing the latter half to two-thirds of the game reduces the
>> concentration required and effort expended; I think this would distort
>> the results.
>>
>> So realistically, no, I think all the games must be fully completed.
>> By the same token, theoretically all these games should take place within
>> a single month, but I worry that if I attempt to set that condition that
>> nobody will even try -- or, possibly, someone would wait until June to
>> try. (Which isn't such a bad plan.)
>>
> Actually June wouldn't work for me, I've got other things I want to
> work on then :)
>
> Hmm, one month to do 13 games ... no can do for me. How about instead
> we just set the challenge for ending on May 31?
I'm somewhat open to that, though I would require that all 13 games be
played in sequence on the given public server; "practice" games offline
are no big deal, but there should be no question about whether a
particular death is a "death", or a "practice" game. Once someone has
begun the challenge, each game played on that server should count as
part of the challenge, and games not on that server should not count.
"Practicing offline" sounds at first like something that should be
discouraged, but keep in mind that this challenge isn't intended to
evaluate any particular player's skill, but instead the efficiency of
the entire strategy. As such, refining or altering the strategy
(subject to the preconditions) should certainly be permissible, and
where else would someone do that but "offline"? However, I do want to
always make sure that a player, upon starting a game, knows _before
anything has happened in that game_ that the particular instance either
will or won't count in the results.
> By the way, is anyone else out there going to give it a try? And
> Derek, what about you?
Heh. I barely have enough time to code these days, much less blow
through 13 potential-ascensions.
However, there is another reason: As the issuer of the challenge, it
can be safely presumed that I have a pre-formed set of expectations of
the results. If I performed poorly, proponents would immediately raise
the objections that I deliberately didn't make my best effort in order
to skew the results.
This is why the original message was aimed at pro-PR proponents (or at
least those who believe it's possible to maintain equal efficiency);
there can be no question that those players are putting forth their best
effort to succeed -- again, reducing potential quibbling. :)
Derek Ray
> I wasn't familiar with the tournament and its trophies. A description
> for the trophies can be found here:
>
> http://nethack.devnull.net/tournament/trophies.shtml
Sorry about that. I probably should have linked that in the original
post, for clarity, but it was getting pretty long as-is.
> 13 games?! I can barely find the patience to play one.
Well, 13 games would only be necessary if you were going for the Grand
Slam. I have no delusions that people will aim for that first; this is
why I selected the devnull trophy structure, so that something less than
"13 out of 13" could be used as a measurement tool. Each of the trophies
requires a minimum number of plays, as follows:
Birdie (all sexes): 2 games
Double Top (Birdie + all alignments): 3 games
Hat Trick (Double Top + all races): 5 games
"bells on" for any -- games in consecutive sequence.
It's only once you start going for all roles that you're stuck at 13.
Derek Ray
>> obtain a minimum of 7 points of divine protection from the priest
>> prior to proceeding onward towards ascension.
>
> A suggestion for rewording would be: "player's primary objective shall
> be to proceed to a temple and obtain a minimum of 7 points of divine
> protection from the priest prior to proceeding onwards toward
> ascension." After all, if a temple is found on level 9 or 10 there is
> no reason to hazard the dangers of the mines.
Byron has already answered this; I agree with him. Finding a temple on
level 9 or 10 is convenient, for sure, but doesn't test the efficacy of
the protection racket itself.
>> * The player shall not intentionally descend below the Sokoban level
>> in the main dungeon branch until the condition of divine protection
>> has been satisfied. If circumstances drop the player below Sokoban
>> (hole trap, level teleporter), he must make immediate effort to return
>> above the Sokoban level.
>>
> I don't think this rule makes so much sense.
> [...]
> Sneaking around looking for gold does not necessarily imply not
> needing protection, but if gold on lower levels is sufficiently easy
> to grab-and-go, then why not grab-it-and-go?
This is here simply to constrain focus and limit future objections to
results on the basis of "look how deep you went without protection,
obviously the PR is meaningless." The bar has to be set somewhere.
(If it helps, Byron's message makes it clear that he considers levels
that deep to be a death trap at low XL, and he is one of the biggest
PR proponents around; when possible, I would prefer to use the PR
proponents' definitions of what the PR constitutes as it facilitates a
lack of need to interpret the results.)
(Did that last sentence make sense? Hmm.)
XyZZy
wrote:
i just realized, wouldn't it be impossible to get grand slam with
bells on, since PR breaks atheist conduct? (if i remember right).
Carl
wrote:
> On 2008-01-25, Carl <carlhnel...@gmail.com> wrote:
>
>
>
>
>
> > On Jan 24, 8:32 am, Derek Ray <de...@moot.its.only.a.spamtrap.org>
> >> On 2008-01-22, Carl <carlhnel...@gmail.com> wrote:
> >> > I'm willing to give it a try just to add another data point to the
> >> > experiment, though I am not a "proponent" of the protection racket.
> >> > But, does the game have to go on to completion?
> >> While I understand the point made (there's a point beyond which the
> >> protection racket is very little influence), I think that yes, the game
> >> should in fact go on to completion.
>
> >> Partly, this is to adhere strictly to the manner in which real games would
> >> be played; any deviation could potentially raise doubt about the
> >> results, such as "sure, he got to the Quest, but he couldn't have ascended
> >> from there, he burned up all his stuff." If all games finish in
> >> ascension or death, no quibbling is available.
>
> > Well, I still think you're measuring the whole game unnecessarily.
>
> Past experience doing similar things has indicated that people will
> quibble about _anything_ if they disagree with the results. I want
> there to be as little interpretation of the results as necessary;
> ideally, that would be "zero" interpretation, but that may or may not be
> possible.
I see your point, but I don't think going to ascension at all removes
the "shoulda-coulda-woulda" factor. Consider what would happen if a
player has successfully completed the PR but then on the way up with
the amulet (thousands of turns later) decides to fly across Medusa
with levitation boots and YASD's with a foocubus. Are you saying that
the "drowned in a pool" ending would count against the PR? That
doesn't seem like a fair way to evaluate the PR.
Another consideration is that requiring the game to be Ascended in
order to count for PR excludes less consistent players from
contributing towards making a good showing for the PR. If there is a
player who has never ascended but is able to use the PR to
consistently get them past the early game, that would tend to speak
favorably for the PR.
Yes, I do, despite my parents refusing to buy me a Super Nintendo as a
kid :)
>
> > The Nethack endgame for a veteran player is hardly an exercise in skill;
> > it's an exercise in patience.
>
> And the capacity for that patience is, in fact, part of the 'skill' of nethack.
>
> >> Removing the latter half to two-thirds of the game reduces the
> >> concentration required and effort expended; I think this would distort
> >> the results.
>
> >> So realistically, no, I think all the games must be fully completed.
> >> By the same token, theoretically all these games should take place within
> >> a single month, but I worry that if I attempt to set that condition that
> >> nobody will even try -- or, possibly, someone would wait until June to
> >> try. (Which isn't such a bad plan.)
>
> > Actually June wouldn't work for me, I've got other things I want to
> > work on then :)
>
> > Hmm, one month to do 13 games ... no can do for me. How about instead
> > we just set the challenge for ending on May 31?
>
> I'm somewhat open to that, though I would require that all 13 games be
> played in sequence on the given public server; "practice" games offline
> are no big deal, but there should be no question about whether a
> particular death is a "death", or a "practice" game. Once someone has
> begun the challenge, each game played on that server should count as
> part of the challenge, and games not on that server should not count.
Sounds appropriate. One could even do like I did with
'Tenaya11' (eleven conduct account) and 'Tenaya12' (in-progress twelve
conduct project) on NAO.
>
> "Practicing offline" sounds at first like something that should be
> discouraged, but keep in mind that this challenge isn't intended to
> evaluate any particular player's skill, but instead the efficiency of
> the entire strategy. As such, refining or altering the strategy
> (subject to the preconditions) should certainly be permissible, and
> where else would someone do that but "offline"? However, I do want to
> always make sure that a player, upon starting a game, knows _before
> anything has happened in that game_ that the particular instance either
> will or won't count in the results.
>
> > By the way, is anyone else out there going to give it a try? And
> > Derek, what about you?
>
> Heh. I barely have enough time to code these days, much less blow
> through 13 potential-ascensions.
>
> However, there is another reason: As the issuer of the challenge, it
> can be safely presumed that I have a pre-formed set of expectations of
> the results. If I performed poorly, proponents would immediately raise
> the objections that I deliberately didn't make my best effort in order
> to skew the results.
True, but if you performed well, then that would count.
>
> This is why the original message was aimed at pro-PR proponents (or at
> least those who believe it's possible to maintain equal efficiency);
> there can be no question that those players are putting forth their best
> effort to succeed -- again, reducing potential quibbling. :)
So, what do we have now? Two PR proponents responded and one of them
definitely too busy. Who ever is out there that really believes in
this PR-thing had better step up! :)
sjde...@yahoo.com
Grand slam with bells on is every race, alignment, gender, and role in
a row (13 games). No conducts are tracked.
Full Monty tracks conduct. It wouldn't be achievable, but certainly
if someone pulled off a Grand Slam (or even 10/13 or something) that'd
speak pretty well for the PR.
My concern with the PR is that it seems to add a 15% or so chance of
random early death, and that really eliminating randomness is the key
to ascending at a high ratio. Once you're into conducts, you're past
the point of trying to win normally into ways to introduce additional
challenges or randomness.
Derek Ray
> i just realized, wouldn't it be impossible to get grand slam with
> bells on, since PR breaks atheist conduct? (if i remember right).
Off the top of my head, I can't remember whether it's Full Monty that's
"all conducts" or Grand Slam. One of the two is just "all roles", the
other is "everything else + all conducts".
Obviously the PR precludes at least one and likely most conducts.
XyZZy
wrote:
(whoops, yeah i gotta read more carefully as far as the names of the
trophies and such goes).
which brings me to my question of whether the conducts in this case
are going to be considered or not? or will it be a grey area type of
thing ("5/13 with 5 conducts worth more than 6/13 with no conducts"
and such)
Derek Ray
>> >> be played; any deviation could potentially raise doubt about the
>> >> results, such as "sure, he got to the Quest, but he couldn't have ascended
>> >> from there, he burned up all his stuff." If all games finish in
>> >> ascension or death, no quibbling is available.
>>
>> quibble about _anything_ if they disagree with the results. I want
>> there to be as little interpretation of the results as necessary;
>> ideally, that would be "zero" interpretation, but that may or may not be
>> possible.
>
> I see your point, but I don't think going to ascension at all removes
> the "shoulda-coulda-woulda" factor.
It does in the important respect; the player won, or he didn't. From the
perspective of an unbiased watcher, there can be no questioning that an
ascension demonstrates "The PR worked in this game."
> Consider what would happen if a
> player has successfully completed the PR but then on the way up with
> the amulet (thousands of turns later) decides to fly across Medusa
> with levitation boots and YASD's with a foocubus. Are you saying that
> the "drowned in a pool" ending would count against the PR? That
> doesn't seem like a fair way to evaluate the PR.
This is exactly the kind of "interpretation" of game results that I seek
to avoid, though.
Games are measured in ascensions, deaths, or escapes (and
occasionally, celestial disgraces). Games are not measured in "I
would've ascended if I hadn't YASDed." A death is a death, not a
successful PR.
Running the PR requires far more concentration than actually playing
normally; I am not willing to concede that such an expenditure of
concentration in the early game (and over the course of 13 games) may
not show its effect later... perhaps in a YASD that would've normally
been avoided.
> Another consideration is that requiring the game to be Ascended in
> order to count for PR excludes less consistent players from
> contributing towards making a good showing for the PR.
Less consistent players will likely not be able to manage any of the
"bells on" varieties, but 5 out of 13 should certainly be possible for a
hat trick, shouldn't it? There aren't _that_ many hat tricks that occur
in devnull each year.
This challenge is obviously not aimed at newer players; but, then again,
newer players are unlikely to be able to consistently complete the early
game, whether they use the PR or play normally, so it's kind of moot.
> If there is a player who has never ascended but is able to use the PR to
> consistently get them past the early game, that would tend to speak
> favorably for the PR.
But again, this is "interpreting" the results.
A player who cannot ascend cannot ascend whether he uses the PR or
plays normally. This says nothing about either the PR or playing
normally, and only measures the skill of the individual player; exactly
what we do NOT want to do.
>> Because it forces the player to defocus on the PR skills and refocus on
>> actually completing the game -- which is exactly what he would be doing
>> if he were playing "for real".
>>
>> As another (somewhat unrelated) example, which is easier -- to complete
>> World 1 of Super Mario World without losing a life 10 times in a row, or
>> to do it while _also_ completing the rest of the game right afterwards?
>>
>> Obviously, the first way is easier; you can mentally discard anything
>> that doesn't relate to World 1. Now... which is more like the way
>> someone would normally play the game?
>>
>> You see my point.
>
> Yes, I do, despite my parents refusing to buy me a Super Nintendo as a
> kid :)
(Well, yes. I was hoping that the analogy was clear enough even for those
who had never played the game. :)
>> > Hmm, one month to do 13 games ... no can do for me. How about instead
>> > we just set the challenge for ending on May 31?
>> I'm somewhat open to that, though I would require that all 13 games be
>> played in sequence on the given public server; "practice" games offline
>> are no big deal, but there should be no question about whether a
>> particular death is a "death", or a "practice" game. Once someone has
>> begun the challenge, each game played on that server should count as
>> part of the challenge, and games not on that server should not count.
> Sounds appropriate. One could even do like I did with
> 'Tenaya11' (eleven conduct account) and 'Tenaya12' (in-progress twelve
> conduct project) on NAO.
I would certainly _recommend_ that someone create a new 'throwaway'
account to play, though some people may want to try on their normal. As
long as we know when they began (announced before they started playing
the first game, of course) and then took the next 13 games after that,
any nick would work fine.
>> However, there is another reason: As the issuer of the challenge, it
>> can be safely presumed that I have a pre-formed set of expectations of
>> the results. If I performed poorly, proponents would immediately raise
>> the objections that I deliberately didn't make my best effort in order
>> to skew the results.
> True, but if you performed well, then that would count.
I cannot predict the future. (Though I can guess; my best series of games
is 9 out of 10, all different roles; I have little confidence that I can
match that playing the PR exclusively, especially given that I don't
normally play the PR in the first place. I doubt I have a sufficiently
complete grasp of the nuances.)
>> This is why the original message was aimed at pro-PR proponents (or at
>> least those who believe it's possible to maintain equal efficiency);
>> there can be no question that those players are putting forth their best
>> effort to succeed -- again, reducing potential quibbling. :)
> So, what do we have now? Two PR proponents responded and one of them
> definitely too busy. Who ever is out there that really believes in
> this PR-thing had better step up! :)
On the plus side, there's no need for this to be a "february or bust"
thing; it can sit out there for time immemorial until it becomes either
obsolete (someone manages a Grand Slam with the PR, demonstrating the
pro- side quite adequately) or moot (Nethack v.NextVersion comes out
and wrecks the status quo sufficiently that all prior assumptions about
_all_ strategies must be re-evaluated.)
Derek Ray
>
> are going to be considered or not? or will it be a grey area type of
> thing ("5/13 with 5 conducts worth more than 6/13 with no conducts"
> and such)
Not considered. The protection racket itself is the conduct; it's
always best to test something with all players involved trying to win
"no holds barred" first, rather than leading off with "illiterate
foodless".