Re: Analysis of Janis' Dwa Cav Law game on 3/6 - (very long)

Derek Ray wrote:
Taken from observations of the following ttyrec:

Thanks for your analysis. (Frankly, after having me criticized so much
over the time I was really awaiting that you look at one of my games.)
I'll try to give my comments and opinions without extending the debate
we already had.

---------------------------------------------------------------------------
Frame 9: Neutral altar in the starting room. Super handy, major positive.

[I was lawful, BTW.]
Yes, I had a good feeling about that luck. But kept a cool head not to
start sacrificing edible food, because early altars are a death trap,
especially if you're lacking food. Intended to come back as soon as the
food question will have been settled.

Makes up nicely for your Dex:14, which is close to the worst I've seen on
a Caveman. On the plus side, Con:20 is excellent, and you get dwarven HP
boosts as well. You will have gallons of HP this game; never bad.

Indeed. Killing for HPs because of that super Dex was one point I tried
to achieve; a huge HP backup is worth a lot and can compensate the worse
parts of the character. Though, lacking poison resistance is still a major
issue, even if one has high HPs.

I would make my primary goals this game a) gain XL as fast as possible to
abuse the HP advantage, and b) convert this altar as soon as feasible.

Fine. We agree.

Frame 10: ...I would drop my starting rocks, by the way. They're junk.

As you like. (As long as I have no severe capacity problems I'd like to
keep them to be able to do ranged attacks. That surely wasn't a problem.)

Frame 325: You pick up two orcish daggers, one of which is cursed. You already have a lockpick -- there is no reason to do this. You also didn't leave your second +0 skull cap behind. Carrying junk objects around is a very bad habit.

I like to take the loot on the floor with me, if possible, to lock it away
from monsters; that's the primary concern. An additional piece of armor on
the head of some critter I'll try to avoid. And even a cursed weapon in a
monster's equipment may take the few HP that might decide between life and
death in the early game.

Frame 557: You pick up an empty large box and become Burdened as a result.
There isn't much point to the box right now, honestly; it's just a turn-waster
and a distraction from what should be your primary goals.

Yes, I agree. It might have been slightly better to delay that for later.
Though the box was not far from the stairs and I felt I could spend the few
turns (~50) to put it right to the altar and lock away the cursed dagger
and spare helmet as intended, and also the deathdrop loot from potential
sacrifice victims.

Frame 583: You pick up the kobold corpse and become Stressed. Why do you care at all about the kobold corpse? You can't eat it. Your dog won't eat it. It should stay on the floor and rot. If you are hoping to make it back to the altar _on another level, while burdened_, you are overly
optimistic. Plus, you are too low level to convert reliably, and -1 luck for failure is very bad for an XL2 character. This is what I mean above by
"as soon as feasible".

Having a poisonous corpse I'll try convert the altar, it's the appropriate
use for the corpse, since the other uses (as you also wrote) don't apply.
My experience with converting an altar is; you may sacrifice 200 corpses
and don't succeed in one game, you can convert it with your first newt and
get an artifact weapon with your second sacrificed newt in another game.
The bad armor and weapon makes this try worthwhile, IMO, since an early
artifact can easily compensate other early shortcomings in the game.

Frame 629: And sure enough, "nothing happens". Better than -1 luck though.

Sloppy play. Typically I count the distance and unburden myself if necessary
to reach the goal. Operator error, I admit.

Frame 669: Back to the box. It's not like you have anything worth stashing...

Frame 709: Dropped the box by the DL1 altar. For someone who is so concerned about food, you just wasted about 150 turns of nutrition or so doing things that you could've done later -- when you had an actual need.

Already agreed, that it might be slightly better. But I think the reasons
explained above are not unreasonable. (Actually it wasn't food that killed
me, neither directly nor indirectly, in that short game.) Each decision has
it's advantages or trade-offs.

Frame 740: Put the junk dagger and the junk orcish helm in the box; don't know
why, neither of these will be useful later. _Didn't_ put the scroll in the box, which is the only thing that needs protecting. Still carrying a second junk dagger despite having a lockpick. Orcish daggers do 1d3; cavemen suck with daggers.

First point explained above.

At that stage I don't want to protect scrolls; I think that is pointless
at that stage. Rather, I'd like to take it with me because it could become
useful, e.g. after price ID'ing it in a shop. The case that I hit, say, a
magic trap (emitting fire) and the scroll will get destroyed isn't likely.
And a creature with a wand of fire imposes a problem of more severe nature
than that a lost scroll would matter.

Projectiles in any amount are not worthless, IMO. (But I am a packrat,
I admit.) Having had that dagger in the inventory was not inherently bad.

(No, the intention was not to force locks with the daggers.)

Frame 762: Finally back to exploring new ground.

Frame 785: Could easily kill this hobbit and eat him, or better would be to let your dog do it to level itself up a bit.

HP gain is *the* primary advantage of that specific character. I wanted to
get the HP backup. Above we already agreed on that.

The point is; you can decide to boost your pet, or boost yourself, or to
some smaller degree you and your pet. People adviced all possibilities in
the past, and there's no inherently right or wrong way to go. (One argument
mentioned in the past was; pets can die, and then you are left with neither
the pet nor the HP gain.) I preferred to boost my character in the game
because of the high HP gains, because of the Caveman's high Con. I think
this was the right decision.

Frame 827: You reach XL3! This is kind of interesting, because you died at XL4; up til now, you are doing just fine with XL/DL progression, and you're not seeing any particularly unusual monsters.

Glad to hear.

Frame 877: Wererat in human form first spotted. It's the weakest werefoo, but
still somewhat of a threat to you if it gets off a summon, or, more annoyingly,
turns to rat form and gives you lycanthropy. Now would be a great time to start banging down Elbereth to protect yourself; once it changes back to human
form, you can melee it safely.

Frame 888: Letting your fresh starting pet deal with the wererat isn't a good option, though; if you had been putting down some Elbereth you could've been in there helping him by firing from your sling (less chance of disturbing dust).

Given that it is indeed the "weakest werefoo", as you correctly say, there's
nothing bad to let your dog handle it. And keeping up a distance between the
character and the werefoo is a Good Thing exactly because of all the reasons
you mentioned.

Frame 895: And sure enough, your pet dies; this is your first major mistake.

I explained my differing opinion.

Frame 903: You should be E-ngraving Elbereth instead of throwing rocks. There will be plenty of time to throw rocks when the wererat is running away.

I disagree to do that at this point; a zig-zag corridor is a bad place to
hit a fleeing monster often enough in a short time span.

Frame 911: You hit a dead end and are forced to do what you should've done in
the first place, when you would've had a live pet to help.

I explained my differing opinion, WRT the pet.

The straight corridor made it actually possible to fire a couple of stones
at the werefoo, while the creature had no possibility to escape to the left
or right when low on HP's. And indeed, just before he reached me to bring
me in potential trouble, he got killed. I think this was a good decision.

Frame 926: It succeeds first try. (Finger Elbereth ratio: 1/1, 100%, expected 70%)

Frame 964: You've fired some flint stones and engrave a second Elbereth. It succeeds first try as well. (Finger Elbereth ratio: 2/2, 100%, expected 70%)

Frame 996: And the flint stones finally kill the wererat, but it's cost you
your pet -- not a very good play here at all.

I explained my differing opinion.

Frame 1020: Giant bats are a pain. Better to have switched to the club here; your to-hit will be the same and your damage output will be much higher due to
strength bonuses.

Yes, they are a pain. Two options I had; either use melee, as you suggest,
to make the higher damage with the club in close contact, or continue with
ranged attack without losing a turn while keeping the bat at distance and
firing from an E-square. I think both options have their pros and cons.

*Here*, I have to say, I made a Bad Move; I wrote the E-word far too late,
giving the bat free attacks. (You may have missed that; doesn't mind.) The
point is that there's no escape route (contrary to the situation before
where I was able to flee), I was still in the dead-end corridor and would
have to pass the bat somehow. But despite the unlucky Elbereth's below I
survived.

Frame 1034: You attempt Elbereth; it fails. (F.E. ratio 2/3, 66%, expected 70%)

Frame 1066: Second Elbereth succeeds. (3/4, 75%, ex. 70%)

Frame 1080: Hungry, but you have a cactus so this is not an immediate concern.

Right.

Frame 1096: Elbereth again. (4/5, 80%)

Frame 1114: Slung stone finally hits and the bat is off your ass. And you're XL4 now!

Frame 1124: You should eat the bat corpse; being stunned is minor.

My own experience advised me to not eat the bat. As you correctly stated
above "this is not an immediate concern", and that is what I had in mind.
(Yet, I've still all options in the game.)

Frame 1210: You should be praying here, not eating your only guaranteed source
of nutrition. There's nothing in the immediate future to suggest you'll need
your prayer, and food management can be key.

In *retrospection* I may agree that this decision might have evolved
better in my game; but it's certainly no clear decision which choice
would have been undoubtly better. Rather; my expectation was that now
we're comming into the DLVL/XL range where there will be enough food
from upcoming jackals and the like available soon. And since this bat
did hit me so hard it seemed a wiser choice to keep the praying option
intact for real emergencies.

Frame 1257: ...You really need to switch away from the sling.

The intention was to get better proficiency in my ranged weapon skill.
(And I don't like meleeing with clubs; I've made bad experiences with
that choice - not that Basic sling were any better.)

Frame 1285: Quick examination shows a normal rabid rat, not a wererat...

Frame 1293: ...Why are you Elberething against a rabid rat? Just kill it.

Probably psychological; incidentally, one of my played NAO Cavemen died
due to a rabid rat with similar data (HP:39, Dlvl:4) not long ago.

Frame 1307: Regardless, Elbereth succeeds. (5/6, 83%)

Frame 1349: Another Elbereth succeeds. (6/7, 86%)

Frame 1375: An Elbereth fails. (6/8, 75%) Why aren't you just beating this thing to death with your club? You have 50 hit points (43, now); it's not going to
be able to kill you unless you manage to make yourself helpless.

Frame 1393: An Elbereth succeeds. (7/9, 78%)

Frame 1407: That took a lot longer than the club would have; when you're low
on food, speed is imperative.

I've explained above the various aspekts that I considered.

Still uncritical (no hunger, full HPs) so that a couple turns ranged
proficiency training can be spent. And I have the pray option still
available, if necessary, that I saved for emergencies of any kind
(praying is, as a joker that fixes many things, very valuable).

Frame 1423: Not much point in picking up the spellbook, though your Caveman is surprisingly smart at IN:10.

Yes, I could have left that behind for the moment, or take it with my.

Frame 1466: Bashed a grid bug to death with the sling instead of your club.

I either missed that, or didn't find it worthwhile to switch weapons
just for a gridbug; I often do that in meaningless cases. (Training
the club was not my primary choice in this case, as explained.)

Frame 1488: Apparently thinking better of carting the useless spellbook all the way back to the DL1 chest, you drop it here. Could've saved some time by just not picking it up in the first place...

That's standard play; drop items that shall not get lost besides the
down-staircase of the previous level.

Frame 1490: And hunger sets in.

Frame 1541: Dark mines. Less than ideal, but dwarves have infravision and the most hazardous inhabitants (the dwarves) will at least be peaceful. Also,
lots of gnomes here for cheap XP to boost your XL, and while gnomes aren't a
lot of nutrition they are at least better than nothing.

Frame 1549: Back up to continue exploring DL4 (a good plan).

Frame 1590: Hunger status shifts to weak. Time to pray.

Frame 1592: You apparently don't notice you're Weak...

I did. Between weak and fainting you have some more time to act. I delayed
the prayer because I was able to reach some more adjacent, yet unexplored,
rooms for some possible random food on the floor which could save my prayer
for other reasons. Immediately after I entered the third room and saw there
wasn't any food I prayed. (You may have noticed that seemingly inattentive
moves may be intentionally choosen. Whether you didn't think of that option
or whether you consider this option as bad or irrelevant must be considered
individually.) And...

Frame 1626: ...until now, and you do #pray. Remember, the earlier you pray, the earlier you get your prayer back.

....and the expected jackal (two actually) turned up only ~30 turns later.

Frame 1653: You opt to descend further into the DoD rather than the Mines. I don't like this choice at all; the Mines will bring you Minetown with a possible delicatessen, and no shortage of gnomes to kill on the way to collect XP. Depending on luck, you could get a general store or hardware store, both of which you could use -- and, there's always the chance of a magic lamp and coaligned altar. Plus, gnomes aren't fabulous nutrition but are better
than nothing -- which is what you have right now.

I already explained why I think that the mines are IMO the less attractive
choice. The stated reasons included that the problem lies before reaching
minetown with all it's potential or rare chances. I see what you say, and
I understand it. While I understand that one might prefer that route, my
choice would be different in this case.

There is little reason to fear the Mines if lawful with infravision; this is definitely not the right path to be taking given your circumstances.

Plus, you still haven't switched back to the club.

Yes. For the reasons mentioned. (Regularily looking at the #enhance menu.)

Frame 1660: Hmm, ring. I'd be tempted to go back to the altar, but rings do
cost hunger and you probably can't afford to test it now anyway.

Yes, I always ponder exactly about the same questions. Sometimes in favour
of the one decision, sometimes the other. It's often nuances in the game
situation that tips the balance for an actual decision.

Frame 1681: Why are you running away from a 'd' while you have no idea what it is? If you're worried about werejackal, examine it. Then, when you see it's just a normal jackal, beat it to death with your club.

Frame 1737: It's worth noting here that you've picked up two different types of gems so far and haven't made any effort to test them for hardness; while engraving E in the floor with a hard gem does take 8 turns, it can serve as a 'safe' point to return to in case of danger. It does require some prep to use,
yes.

I agree in the necessary preparation and I agree in the usefulness, and I
can assert you that I use all available types of engravers according their
specific applicability in situations. Where I disagree is; it really isn't
worth to check every gem so early in the game. I am sure that I've never
ever found a hard gem so early in that shallow a level. But anyway; you can
always defer (since that check does not cost any turns) that check whenever
actually required or whenever you need to know whether you have such an
option or not. No need to check any single stone. Later you can even tell
all the glass apart by their large stacks without any explicit checks.

Frame 1750: The sling's useful in the early game because of the flint stones. Carrying around rocks is just not worth it, though; and definitely don't worry about enhancing it. If you had been beating things to death with your club, on the other hand, you might be able to enhance _that_ and get some actual benefit. Skill slots aren't really at a premium for a Caveman, fortunately, so even
though you'll never use the club any more after you get something better, you
aren't really costing yourself anything.

Frame 1795: Kill the dwarf and check for mithril. You can't pray anyway, and you'll have plenty of time to fix your alignment while your prayer timeout is regenerating. Plus, he just wielded a pickaxe, so you know he doesn't have a mattock. Decent expectation: you should get a hat and boots to get to AC:3 instead of AC:7.

Hmm.. - I am reluctant to try to kill that single dwarf for the chance of
a piece of armor. I prefer backing my own armor first up by pieces that
are easier obtainable, e.g. typically lying around on the floor. But YMMV;
it's an option one may choose to consider.

Frame 1815: Trapdoor surprise!

Frame 1816: ...er, where are the soldier ants again?

The frame recoding information is misleading; the turns are relevant...!!

It's turn T:1523, I saw and met them at T:1526, being immediately attacked
without own turn. IOW; at the moment I first saw them they reduced my HPs
*immediately* from 50->17 in one turn. (The frame protocol below and how
you presented it is very misleading. I assume inadvertently.)

Frame 1822: Ah, THERE they are. You can see two of them: major threat. Elbereth ASAP. Once you get one down you can drop a rock and back away down the corridor, and look for the upstairs in the other direction.

But what you propose is academic theory and was not possible, because the
Elbereth's didn't succeed, as you also noted below.

Frame 1826: Bite. 50->46. Sting. 46->37. Bite. 37->31. Sting/poison. 31->17. Soldier ants are speed 18, of course, which is why they're so dangerous.

Frame 1850: Elbereth fails (7/10, 70%, expected 70%) and the soldier ant bites and stings; 17->9.

Frame 1867: Elbereth fails again, dead. (7/11, 63%, expected 70% -- well within normal limits of variance for finger Elbereth).

First encounter: HP:50->17 (without own turn - and unburdened)
Then two consecutive unsuccessful Elbereths and DYWYPI.

Thanks again for the protocol.


It's not possible to predict trapdoors, no. But you made a lot of mistakes leading up to this, not the least of which being picking the wrong branch _and_
letting your pet die early on.

As so often we disagree in the resume.

DoD is not preferable to the Mines in this case, and wasn't relevant WRT
the death anyway. This had also been discussed in another thread and I
won't expand on it again. We have apparently differing opinions on that.
How the pet was best involved is disputable; I think I made my point clear
with what I said above, partly supported by your own stated facts, and I
think the decision was not unreasonable.

To sum up; I am sure I do not play flawless like a perfect machine, but the
choices taken should be comprehensible and mostly consider reasonable, even
if you'd have preferred in certain cases an other decision. There are points
you pesented as a bit absurd, at least you didn't seem to have seen the
motivation behind. I think I could make some motivations clear. That does
not mean you should adopt that views, or that it fits to your optimal way
to play. But what you gave as alternatives was really not convincing to me;
possible alternatives, yes, but generally not superior alternatives, no.
Some seem (especially in retrospective) to be a bit better, some worse.
The game gave many options where to choose from - yeah, that's Nethack,
interesting as ever - unavoidable was only the final death in three turns.

I'm inclined to say that this was not at all "unavoidable"; if you had been picking safer
choices, you would've been in better shape (trapdoors in the Mines often drop you into
Minetown, for example). Plus, you _know_ there is stock dwarvish armor in the mines,
and a hat, boots, and roundshield would've put you at AC:0 or AC:1 -- while the XP from
the gnomes would've given you a bigger HP cushion with each level (CON:20!), and that
would've been a great help in at least giving you a turn or three more in bad spots.

I deliberately put my resume before yours so that anyone who has read thus
far - thanks, tough crowd! - will read your opinon as the "last word" thus
reducing the risk of an repeated endless debate. :-)

My intention was to give more insights in my playing and thinking than your
analysis alone could give.

Thanks again for your analysis and for your interesting views on the topic.

Janis
.

Relevant Pages

  • Re: Requesting tips, comments for an EDT thread-safe game architecture
    ... the amount of time that has passed between each frame of animation on those it doesn't), and then I use that elapsed time to calculate whether to skip ahead a couple of frames (on computers that are a bit slow for this game) or to interpolate between two frames of animations. ... take the character 2 seconds to walk from point A to point B", and the engine will smoothly animate a sprite moving from point A to point B at the highest frame rate that the computer can handle. ... But I *do* want a decorate frame, so that the player can move the window around and resize it. ... public static void mainthrows InterruptedException, ...
    (comp.lang.java.gui)
  • Re: timer not being called when OnPaint calls Invalidate
    ... You don't say what kind of game you have, but it may or may not be the case ... that you need the highest frame rate. ... requires a high frame rate will not use the normal window messaging paradigm ... rate is to invalidate the entire window. ...
    (microsoft.public.dotnet.languages.csharp)
  • Re: Requesting tips, comments for an EDT thread-safe game architecture
    ... For active rendering I would take your rendering out of the EDT ... a bit fast for this game). ... the highest frame rate that the computer can handle. ... public static void mainthrows InterruptedException, ...
    (comp.lang.java.gui)
  • Re: Acorn Invaders (Machine Speed Checks)
    ... "I'm writing a game. ... So you have two "clocks" to worry about - the frame rate, ... The presence of the VSync wait in the loop locks the loop's rate to the ...
    (comp.sys.acorn.programmer)
  • Re: Question for those that have a job
    ... are going to fast food restaurants, where the food is almost as cheap ... If you want to know what went wrong with the game .... ... Then fuck all you money changers. ... You need fish to feed a shark, but the fish need to be fed also. ...
    (rec.gambling.poker)