Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

Tactical roguelike combat

226 views
Skip to first unread message

Andrew Doull

unread,
Feb 6, 2005, 8:20:45 PM2/6/05
to
Hi,

I have been exploring the idea of creating a more tactical roguelike combat
system, so that the player values fights with small numbers of monsters more,
as opposed to just having to hold down a keypress while running in a particular
direction.

Primarily, I'm looking at the idea of making an attack take twice as long as a
move. This should allow a player to always avoid a single monsters attack by
stepping out of the way, either by retreating (moving away from the monster) or
evading (moving to another location adjacent to a monster). The player should
only be in a position where they can be melee attacked by the monster by
backing themselves into a dead end, or fighting multiple monsters, or by
trying to attack a monster themselves. Similiarly, casting a spell should take
twice as long as a single move, in order that a player has the opportunity to
dodge out of the way, provided that they can get out of the area of effect or
line of sight of the spell.

In order to visualise it, however, I've come up with a more complicated system,
which could entail a bit too much micro-management. In order to keep things
interesting tactically, I want to display the location that the monster or
player can attack. Following this line of reasoning, it makes sense to show the
players weapon in the location adjacent to them which they are attacking.

Suppose a player takes up a single location, but their weapon ('|') and shield
('[') (or any 2 objects they happen to be holding) occupy two adjacent
locations. The player can relocate their weapon to any of 3 locations to the
right of where they are facing, and their shield to any of 3 locations to the
left of where they are facing (We assume the player is also right handed). They
can't have both in the same location. Suppose also their backpack ('&') is shown
to the rear.

Examples of this might be:

[|
@.
&.

[|.
@.
&.

[.|
@.
&.

..
[@|
&.

Similiarly long weapons, like polearms might be arranged as following:

...
.\.
[@\.
&..

.\.
.\.
[@..
&..

\..
\..
[@..
&..

\...
\..
[@..
&..

..\
.\.
[@..
&..

Now, a single action may either deliver an attack to the location the player's
weapon currently is, or relocate their weapon, or relocate their shield. You
can also step to either the 3 locations in front of you, or change facing to
any location except your rear as a single action. You can also put away a
weapon or drop your shield, or open or close your backpack, or get an item out
of an open back pack. Since you can't hold more than 2 items, you have to put
away or drop your weapon and/or shield first before getting anything out of
your backpack. Walking around with an open backpack causes stuff to fall out
occasionally, so its not a good idea.

You can put your weapon on your belt, provided its not a two handed weapon
(which is too large). You can only put a shield or two handed weapon down on
the ground or on your back, which requires you take your backpack off. You can
also only hold a torch in place or a weapon or shield and the light comes from
the torch, as opposed to yourself.

This means anytime you want to (say) quaff a potion of healing, you need to put
away your weapon, drop your shield, open your back pack, get the potion and
then drink the potion, for a total of 5 actions. If the potion happens to be on
the ground, or, say, on your belt, you can put your weapon away, get the potion
and drink the potion, for a total of 3 actions.

If you are heading through a tunnel that is only 1 square wide, you can't hold
both a weapon and shield, so you need to either put your weapon away or drop
your shield.

Now, this is a heck of a lot of micromanagement for a marginally more realistic
system. Its possible to automate a lot of this from the player's sight, but
then actions (such as quaffing a potion) take a variable amount of time.
Ideally, you'd want to record your default equipment state (weapon drawn,
shield in hand) so that everytime you got anything out of your backpack to use,
you'd automatically recover to the previous state when done (draw your weapon
again and pick up your shield), and you wouldn't have to type in 5 separate
commands.

If I ignore all the equipment micromangement side, then it becomes more
straightforward, but the tactical combat side of it loses out (if you take 1
action to drink a potion, that heals all the damage you've sustained by an
opponent carefully maneouvouring around their weapon and shield, it devalues
the combat tactics side).

I could abstract it out again to the explaination in the first paragraph, but
then it loses the advantages of being able to visualise the attack location
cleanly (e.g, if I dodge to the left, then it brings my weapon into play, so
that I can attack him straight away). It then becomes a matter of reading the
message back cleanly ('The goblin swings at you!' or 'The master mage casts a
spell!' is a hint that you need to move), or perhaps recolouring / redrawing
your icon to indicate an attack is targetted against you.

What level of detail would people be willing to tolerate in order to explore a
tactical roguelike combat system?

Andrew
--
Unangband L:C E+ T- R- P+ D-- G+(+) F:Sangband RL-- RLA-- W:F Q++
AI+(++) GFX++ SFX++ RN+++(+) PO++ Hp+++ Re--(+) S++ C- O* KG--

Ray Dillinger

unread,
Feb 6, 2005, 10:23:24 PM2/6/05
to
Andrew Doull wrote:
> Hi,
>
> I have been exploring the idea of creating a more tactical roguelike combat
> system, so that the player values fights with small numbers of monsters more,
> as opposed to just having to hold down a keypress while running in a particular
> direction.
>
> Primarily, I'm looking at the idea of making an attack take twice as long as a
> move. This should allow a player to always avoid a single monsters attack by
> stepping out of the way, either by retreating (moving away from the monster) or
> evading (moving to another location adjacent to a monster). The player should
> only be in a position where they can be melee attacked by the monster by
> backing themselves into a dead end, or fighting multiple monsters, or by
> trying to attack a monster themselves. Similiarly, casting a spell should take
> twice as long as a single move, in order that a player has the opportunity to
> dodge out of the way, provided that they can get out of the area of effect or
> line of sight of the spell.


I tend to agree that attacks should take longer than moves, but this
does not give the player the magical ability to escape a single monster.
If that monster is faster than he is, it will occasionally get an attack
in while chasing him.

I don't know how your timing details work, but in mine an actor makes
a movement (or physical attack) and it takes place immediately. Then
the actor cannot act again until the time that movement or attack
required has run out. This means, amongst other things, that a monster
who has a speed of 11 can chase a PC with speed 10 and catch up to
him at a rate of 1 square per 10 player turns. When the monster is
adjacent (one square behind) an attack will take place instantly.


> In order to visualise it, however, I've come up with a more complicated system,
> which could entail a bit too much micro-management. In order to keep things
> interesting tactically, I want to display the location that the monster or
> player can attack. Following this line of reasoning, it makes sense to show the
> players weapon in the location adjacent to them which they are attacking.


It's interesting, but I wouldn't care too much. I plan to add combat
styles to the game so that someone can set anything from "coward"
to "berserker" as a meelee style, with commensurate bonuses to dodge
or attack. But I think that's about as much complication as I expect
players to deal with.


> You can put your weapon on your belt, provided its not a two handed weapon
> (which is too large). You can only put a shield or two handed weapon down on
> the ground or on your back, which requires you take your backpack off. You can
> also only hold a torch in place or a weapon or shield and the light comes from
> the torch, as opposed to yourself.

You might want to read more about how these weapons were actually
handled and used. Most two-handed swords at any rate had scabbards
that they could be drawn from or placed into without disturbing any
other equipment. This is, of course, just an argument from realism,
and may not apply to two-handed weapons such as axes and polearms.

> This means anytime you want to (say) quaff a potion of healing, you need to put
> away your weapon, drop your shield, open your back pack, get the potion and
> then drink the potion, for a total of 5 actions. If the potion happens to be on
> the ground, or, say, on your belt, you can put your weapon away, get the potion
> and drink the potion, for a total of 3 actions.

Suggestion: have a limited number of "belt" slots. Each represents
some specialized holster for a particular type of object; bottle,
scroll, wand, blade, etc. Things in a "belt" slot can be drawn or
returned in one round. Things in the backpack may take up to
three rounds to draw or return.

Another suggestion is that having 'instant-fix' potions of healing
does devalue tactical combat. You may prefer potions of regeneration
that fix hitpoints gradually, over several dozen rounds, instead.
These provide healing capability without destroying the value of
tactical combat.

Bear

Glen Wheeler

unread,
Feb 7, 2005, 12:12:12 AM2/7/05
to

"Andrew Doull" <andre...@hotmail.com> wrote in message
news:cu6fpd$1fcm$1...@news.vol.cz...

> Hi,
>
> I have been exploring the idea of creating a more tactical roguelike
> combat
> system, so that the player values fights with small numbers of monsters
> more,
> as opposed to just having to hold down a keypress while running in a
> particular
> direction.
>
> [..]

>
> What level of detail would people be willing to tolerate in order to
> explore a
> tactical roguelike combat system?
>

I'm a bit strapped for time now, but thanks for the good read. I would
certainly be willing to tolerate that level of detail. It is my belief that
a good interface can overcome most of the cumbersome details and still
retain the desired gameplay effects. At the very least I would give it a
go.

--
Glen
L:Pyt E+++ T-- R+ P+++ D+ G+ F:*band !RL RLA-
W:AF Q+++ AI++ GFX++ SFX-- RN++++ PO--- !Hp Re-- S+


ABCGi

unread,
Feb 7, 2005, 12:17:06 AM2/7/05
to
Andrew Doull wrote:

*SNIP*

*SNIP*

Interesting approach, you may want to take a look at my BRL idea (Big
RogueLike) which toys with using more than one square for a character...

http://codemonkey.sunsite.dk/projects/beyond/newstuff2.html#brl

Fixed Font...

"Or perhaps less ASCII arty and more tactical;"

############################|r |########################################
#######|----------|######|---==---------------|----------------|########
#######| |######| | #########
#######| |---||-| r + ##########
#######| || r + #########
-------| || | ###~###
+ -- | ~~~~~#
+ G+- ^^ | ~~~~~~~
+ GG: <! @} @> ------\ ~~~~~#
+ G+) <@ *> |######\ ~###
#######| |#######\ ######
#######| @@ |########\ #########
#######| vv |#########\ ########
#######| |# ##\ |#######
#######| |# $$$$$ ###\-----|#######
#######| |# $BRL$ #################
#######| |# $$$$$ #################
#######| |# #################
#######| |#########################
#######| |#########################
#######| |#########################
#######| |#########################
#######---------------------++------------------------------------------


| P
+##\
db - R.Alan Monroe

--
ABCGi (geek) http://codemonkey.sunsite.dk S14 D15 C13 I17 W9 c12
GCS/IT$/L/B$ d+(-) s: a? C++ ULUSU-- P+ L+>++ E- W++$ N+ o+ K--
w+++(--)$ O- !M- V PS++(+) PE-@ Y+(++) PGP>++ t++ 5+ X R(+++) tv
b++(+) DI++++ D+++ G e++>+++ h++(home office!) r++ y++* BAS-----

Ilya Bely

unread,
Feb 7, 2005, 1:24:01 AM2/7/05
to
Hi, there!

Ray Dillinger wrote in message <MaBNd.5684$m31....@typhoon.sonic.net>:

> It's interesting, but I wouldn't care too much. I plan to add combat
> styles to the game so that someone can set anything from "coward"
> to "berserker" as a meelee style, with commensurate bonuses to dodge
> or attack. But I think that's about as much complication as I expect
> players to deal with.
>

I never understood these tactical switches. It seems that they only add
micromanagement. Usually, either one of stances have a significient
advantage or it does not make much difference. Can't you just assume
that player always use most optimal stance?


--
May your code work forever and never have a bug.
At your service, Ilya Bely

Brent Ross

unread,
Feb 7, 2005, 2:34:00 AM2/7/05
to
In article <36ofs1F...@individual.net>,
Ilya Bely <gm...@ilyabely.swapGMAILandILYABELY.com> wrote:
// Hi, there!
//
// Ray Dillinger wrote in message <MaBNd.5684$m31....@typhoon.sonic.net>:
//
// > It's interesting, but I wouldn't care too much. I plan to add combat
// > styles to the game so that someone can set anything from "coward"
// > to "berserker" as a meelee style, with commensurate bonuses to dodge
// > or attack. But I think that's about as much complication as I expect
// > players to deal with.
// >
//
// I never understood these tactical switches. It seems that they only add
// micromanagement. Usually, either one of stances have a significient
// advantage or it does not make much difference. Can't you just assume
// that player always use most optimal stance?

That's why I think there's not much use in having much granularity in
such a system. It's simply better to assume the PC makes the correct
adjustments (giving some things the benefit of the doubt) than it is to
bother the player with it.

If there is any micromanaging, it should be transparent... the internal
melee state machine machine should just say "this monster is practically
harmless to us" and thus do things like prioritize shield block attempts
to more dangerous things. Similarly, if the player has chosen to use a
magic item this turn, it's probably safe to say that the PC would
probably be as evasive as possible if it was in the middle of melee.

As far as user control goes, I think it's better to not design for
"modes" so much as "attacks". For example, attacking overly
aggressively is something which could be useful in desperation (we can
ignore aggressive attacks from positions of superiority can easily and
safely be done transparently). So if we take this do-or-die concept as
a design decision we can try to design a special form of attack which is
only desirable for use when the PC is likely to die... it shouldn't be
something which would be reasonable for common use (after all, that's
the typical problem with tactic modes... they just get set as high as is
comfortable and left there). The end result would ideally be another
option in panic situations... the player would have to weight their
current odds against the success of a last ditch all-or-nothing attack.

Brent Ross

Andrew Doull

unread,
Feb 7, 2005, 6:26:11 AM2/7/05
to

The general idea is that any attack is pre-announced some how so that the player
is given the opportunity to choose whether they should move to avoid the attack
or counterattack. In the 'graphical' outline I gave, the fact that the monster
has a weapon sticking out in a particular direction allows the player to
attempt to avoid the weapon - using a message queue, it would simply say at the
top of the screen 'The goblin swings at you!'.

Considerations of speed differential are important, but basically boil down to
allowing a player to exploit the movement system to beat a monster in ways I
consider a little more artificial. e.g. by drinking a potion of speed that
increases the player's speed above that of the monster's, it is possible to
consistently avoid attack 100% of the time.


>
> > In order to visualise it, however, I've come up with a more complicated system,
> > which could entail a bit too much micro-management. In order to keep things
> > interesting tactically, I want to display the location that the monster or
> > player can attack. Following this line of reasoning, it makes sense to show the
> > players weapon in the location adjacent to them which they are attacking.
>
>
> It's interesting, but I wouldn't care too much. I plan to add combat
> styles to the game so that someone can set anything from "coward"
> to "berserker" as a meelee style, with commensurate bonuses to dodge
> or attack. But I think that's about as much complication as I expect
> players to deal with.
>
>
> > You can put your weapon on your belt, provided its not a two handed weapon
> > (which is too large). You can only put a shield or two handed weapon down on
> > the ground or on your back, which requires you take your backpack off. You can
> > also only hold a torch in place or a weapon or shield and the light comes from
> > the torch, as opposed to yourself.
>
> You might want to read more about how these weapons were actually
> handled and used. Most two-handed swords at any rate had scabbards
> that they could be drawn from or placed into without disturbing any
> other equipment. This is, of course, just an argument from realism,
> and may not apply to two-handed weapons such as axes and polearms.
>

Agreed. I was just setting some parameters for how such a system might work.

> > This means anytime you want to (say) quaff a potion of healing, you need to put
> > away your weapon, drop your shield, open your back pack, get the potion and
> > then drink the potion, for a total of 5 actions. If the potion happens to be on
> > the ground, or, say, on your belt, you can put your weapon away, get the potion
> > and drink the potion, for a total of 3 actions.
>
> Suggestion: have a limited number of "belt" slots. Each represents
> some specialized holster for a particular type of object; bottle,
> scroll, wand, blade, etc. Things in a "belt" slot can be drawn or
> returned in one round. Things in the backpack may take up to
> three rounds to draw or return.
>

Also agreed. Unangband, the Angband variant I currently maintain does exactly
this.

> Another suggestion is that having 'instant-fix' potions of healing
> does devalue tactical combat. You may prefer potions of regeneration
> that fix hitpoints gradually, over several dozen rounds, instead.
> These provide healing capability without destroying the value of
> tactical combat.
>

Also agreed. Instant-fix heals are unbalancing: high-level priests in Angband
are basically invulnerible to any damage that does not kill them within a
single round because of this.

Antoine

unread,
Feb 7, 2005, 9:13:18 AM2/7/05
to

> As far as user control goes, I think it's better to not design for
> "modes" so much as "attacks". For example, attacking overly
> aggressively is something which could be useful in desperation (we
can
> ignore aggressive attacks from positions of superiority can easily
and
> safely be done transparently). So if we take this do-or-die concept
as
> a design decision we can try to design a special form of attack which
is
> only desirable for use when the PC is likely to die... it shouldn't
be
> something which would be reasonable for common use (after all, that's
> the typical problem with tactic modes... they just get set as high as
is
> comfortable and left there). The end result would ideally be another
> option in panic situations... the player would have to weight their
> current odds against the success of a last ditch all-or-nothing
attack.

In a way, you don't even need a separate type of attack to allow a
"do-or-die option". Roguelikes provide this kind of choice all the time
- the do-or-die option is simply to keep fighting another round when
you're on low hp, rather than (eg) teleporting out.

To put it another way, if a player wants to be "in coward mode" they
can run away from the monsters, if they want to be "in aggressive mode"
they can run up and attack them.

Or then again you can model this sort of thing through equipment. The
"coward" option is to wield a shield, the "aggressive" option a
two-handed weapon.

A.

Brent Ross

unread,
Feb 7, 2005, 3:40:58 PM2/7/05
to
In article <1107785598.9...@z14g2000cwz.googlegroups.com>,
Antoine <ma...@guildgame.com> wrote:
// In a way, you don't even need a separate type of attack to allow a
// "do-or-die option". Roguelikes provide this kind of choice all the time
// - the do-or-die option is simply to keep fighting another round when
// you're on low hp, rather than (eg) teleporting out.

I'm talking about a bit more than the regular attack here though. With
a regular attack there's the idea that something is held back for
defense... this special attack would throw all of that away. The PC
would charge in with no intention of ever defending or maintaining
position... if they fail to take the opponent down, they'd be in for one
doozy of a beating (ie multiple critical attacks). This would be a true
do-or-die option, not just the standard "attack and see if I die" option
that's regular fair... this puts the entire game on the line on a single
attack.

// To put it another way, if a player wants to be "in coward mode" they
// can run away from the monsters, if they want to be "in aggressive mode"
// they can run up and attack them.
//
// Or then again you can model this sort of thing through equipment. The
// "coward" option is to wield a shield, the "aggressive" option a
// two-handed weapon.

See the other part of my post... these are proper modes, which I said
should be transparent (just like you). The do-or-die attack is not
about adding a mode (those can be done internally... by simply
interpreting the commands and equipment and adjusting the state)... this
is about adding a tactical melee option outside of that. As such, it
needs to be something extreme and irregular, and beyond the standard
assuptions... the type of thing that's only used occasionally for a
single attack and is a real decision (in this case a strong life or
death gamble). More of a special attack or melee feat than any sort
of tactical mode at all.

Brent Ross


R. Dan Henry

unread,
Feb 24, 2005, 1:26:43 AM2/24/05
to
On Mon, 7 Feb 2005 01:20:45 +0000 (UTC), Andrew Doull
<andre...@hotmail.com> wrote:

>What level of detail would people be willing to tolerate in order to explore a
>tactical roguelike combat system?

Create it and we'll see if it justifies tolerating the level of
detail!

R. Dan Henry
danh...@inreach.com

ABCGi

unread,
Feb 24, 2005, 2:12:29 AM2/24/05
to

To my mind ADoM implements a pretty useful and good example (but not
perfect) of different combat modes. And I have suggested something
similar for GREEKIE.

The main thing it gets right is that is not just - well in this
situation this setting in always right - which means the computer could
just pick that setting automatically for you.

I normally have to choose between;
* Do I rely on my natural stats to get me through this?
* Do I try to get through this quickly and give a boost to my luck
(which I am now relying a little on)?
* Do I be patient and fight defensively until it is safer to take them out?

Other times when it doesn't work so well, and as has been pointed out
about my THUMP attack idea, is when the enemy is weak you just always go
through on berserk and whack them all. Which does lead to the inevitable
- opps did I leave that on... which is an arguement for the temporary
berserk mode with the red screen and recharge time etc

Am I boring yet? Or was I always? ...

Brent Ross

unread,
Feb 24, 2005, 4:14:17 AM2/24/05
to
In article <421d7e67$0$29276$1472...@news.sunsite.dk>,
ABCGi <ab...@yahoo.com> wrote:
//
// To my mind ADoM implements a pretty useful and good example (but not
// perfect) of different combat modes. And I have suggested something
// similar for GREEKIE.

To my mind ADoM shows that the idea of tactical modes isn't so great.
They essentially provide unnecessary weight for the player with
micromanagement which could largely be covered with a transparent and
automatically with a melee state machine. So players often ignore
them... pushing them to progressively higher "optimal" levels as the
game progresses, with only occasional tweaks for special cases. All of
this can be done without the added weight of micromanagement, or even
forcing the player to calculate the optimal value... simply do that work
in the code itself and present the player with only the strong short
term choices (ie the options that would never be desirable as a long
term mode, the extreme exceptions).

They also aren't very management efficient. Balancing the penalties and
benefits of tactical modes over the entire course of a typical RL (ie
keeping them truly meaningful) is extremely difficult. It's not just
adding a simply variable to the equation, it's adding a whole set of
differential equations. For that reason alone, I'd recomend staying
away from them in development... they'll make it too difficult to
see how your game is truly playing (or more precisely, how and what to
fix when the inevitable balance problems start cropping up after the
initial release). Again, giving the player only solid choices is much
better here... it's much easier to balance something that's guaranteed
extreme.

// The main thing it gets right is that is not just - well in this
// situation this setting in always right - which means the computer could
// just pick that setting automatically for you.
//
// I normally have to choose between;
// * Do I rely on my natural stats to get me through this?
// * Do I try to get through this quickly and give a boost to my luck
// (which I am now relying a little on)?
// * Do I be patient and fight defensively until it is safer to take them out?

The catch is that it's really hard to balance this and, in my experience,
ADoM doesn't really do a great job of that except in the early game. This
is typical of things that are difficult to balance, as the early game is
easy to set, but the rapid character development of most RLs makes it hard
to keep balanced. If you really want to do this from early development,
I'd suggest dramatically reducing the scale of character development to
a minimul and the range of monster difficulty (ie don't go newts to dragons,
start well into "middle" level monsters)... it will make things much easier
to tinker with once things are up and running.

// Other times when it doesn't work so well, and as has been pointed out
// about my THUMP attack idea, is when the enemy is weak you just always go
// through on berserk and whack them all. Which does lead to the inevitable
// - opps did I leave that on... which is an arguement for the temporary
// berserk mode with the red screen and recharge time etc

Well, I'd simply have the internal melee code look at such things and
say "mook" and automatically assume maximum offense/minimum defense for
such situations[1]. The PC should automatically know not to waste
things like dodges and shield blocks (assuming both these cost
something), on some rat who'd damage is insignificant to our current HPs
and probably won't even get through the platemail anyways. It should,
however, be saving and using such things for the dire troll that's also
beside you or the wizard across the room who's likely going to throw
some nasty spell your way (thus, essentially you have mixed tactics...
something that's very awkward to get the player to specify).

It's not even that you really have to evaluate what the PC is going to
attempt immediately after the player has entered a command either...
when it comes to these decisions, you can get a good approximation by
simply defering the decisions by making the assumption that the PC is
always optimal when things happen (this is sort of true in most RLs
already... the lack of facing essentially is defering the question of
which way the character is turned until it becomes important, at which
point the character is always facing the correct direction (even if it
is unlikely that the character could turn around that fast)).

But if you really, really want modes... I'd suggest keeping them
abstract and not as direct stat tradeoffs (ie "defensive", "normal",
"aggressive"... no specifics and no inbetween points to complicate your
development further). Then use whatever mode is set, not to tweak
values, but as to guide the state transitions and assumptions that are
made with the melee machine. Design that mode so that the modes "work"
and capture the "feel" of the style... without any considerations for
what any of the specific numbers are. Use the tweaking of stats and
values as an absolute last resort[3], and try and capture the fighting
"styles" you want completely without them... this should produce
something which is far more natural, elegant, and less prone to tricky
balance issues.

// Am I boring yet? Or was I always? ...

I'd say no, but I don't think there's really much left to be said
on the subject anymore.

Brent Ross


[1] This, of course, is refering to the RPG Feng Shui. In it there are
two types of NPC the PCs will run into: unnamed mooks (typically in
groups) and named characters. Since unnamed mooks are insignificant
they can be trivially disposed of by the PCs if desired (it mostly comes
down to declaring a cool way to kill them and then they're typically
assumed to automatically die). In a simlar way, a monster which is no
longer a threat to a RL PC and thus insignificant (especially for
any form of PC benefit like XP) could be made easily removable by the
player simply with the declaration of an attack (ie a hit of sufficent
skill and value autokills monsters considered beneith it... without
tempting a 95% negative binomial distribution [2] to produce a 5).

[2] For the non-statistical people: The negative binomial distribution
is the number of trials (at a specific percentage) before a success
(well, generally it's actually a certain number of successes... in this
case just one (after four misses, thus "5")). It's very useful in RLs.

[3] Even then, it's wise to abstract the effects of the tweaking.
Instead of saying +5 strength, for example, specify it in words like
"GOOD STRENGTH BOOST". Then, inside the melee machine, throw that token
(enum) to a special function to interpret it (which looks at the current
character and decides what "GOOD" currently really means, as opposed to
"SMALL" or "LARGE" boosts). In this way, you isolate your complicated
to balance data into something far more robust (and your intentions will
be far more transparent as well).

Jukka Kuusisto

unread,
Feb 24, 2005, 4:41:26 AM2/24/05
to
bwr...@csclub.uwaterloo.ca (Brent Ross) writes:

>To my mind ADoM shows that the idea of tactical modes isn't so great.

I think the different tactical modes in Adom are good, but there are too
many of them. Three would suffice: berserk, normal and coward.

-Jukka
--
Jukka Kuusisto

Brent Ross

unread,
Feb 24, 2005, 5:00:34 AM2/24/05
to
In article <jkuusist....@vipunen.hut.fi>,
Jukka Kuusisto <jkuu...@cc.hut.fi> wrote:
// bwr...@csclub.uwaterloo.ca (Brent Ross) writes:
//
// >To my mind ADoM shows that the idea of tactical modes isn't so great.
//
// I think the different tactical modes in Adom are good, but there are too
// many of them. Three would suffice: berserk, normal and coward.

Yes, that would certainly make it a lot better.

Brent Ross

ABCGi

unread,
Feb 24, 2005, 8:01:48 AM2/24/05
to
On 2005-02-24 11:00:34, bwr...@csclub.uwaterloo.ca (Brent Ross) wrote:

> In article ,


> Jukka Kuusisto wrote:
> // bwr...@csclub.uwaterloo.ca (Brent Ross) writes:
> //
> // >To my mind ADoM shows that the idea of tactical modes isn't so great.
> //
> // I think the different tactical modes in Adom are good, but there are too
> // many of them. Three would suffice: berserk, normal and coward.
>
> Yes, that would certainly make it a lot better.
>
> Brent Ross

I'd agree with that too, nice post too Brent. Of course I'm loathe to have the
PC automatically change tactics, I would more desire to focus on the tips you
had to make tactics meaningful, but I see your point that the PC is essentially
already assuming some decision making anyway. I've been a fan of adding facing
into RLs for a while...

--
ABCGi ab...@yahoo.com http://abcgi.fly.to S14 D15 I17 W12 C9

Erik Piper

unread,
Feb 24, 2005, 10:44:24 AM2/24/05
to
bork bork bork Brent Ross bork 10:14:17 AM bork 2/24/2005 bork bork:

> In article <421d7e67$0$29276$1472...@news.sunsite.dk>,
> ABCGi <ab...@yahoo.com> wrote:
> //
> // To my mind ADoM implements a pretty useful and good example (but not
> // perfect) of different combat modes. And I have suggested something
> // similar for GREEKIE.
>
> To my mind ADoM shows that the idea of tactical modes isn't so great.

While ADOM has many troubles, I wouldn't say tactical modes are one of them.
They're a function key, an F1 through F7, that you press once in a while on
top of what else you're doing. Or don't; I get the feeling a lot of people
don't. I am a relentless minimaxer and twiddle around minimaxing them. I
never found them to get in the way, though I will admit that they don't
really add much either.

[...]

> ...players often ignore [the intricacies of melee modes]


> ... pushing them to progressively higher "optimal" levels as the
> game progresses,

Not in ADOM; basically from a minimaxing standpoint you either want to
maximize shield marks and minimize boredom at the cost of some lost weapon
marks and some gained risk (so you use the second most aggressive setting) or
you want to minimize boredom at the cost of gaining no shield marks,
increasing risk, and gaining less weapon marks (so you use the most
aggressive settings), or you want to maximize weapon mark...

...but this is getting boring. In any case, these needs don't reach
"progressively higher "optimal" levels as the game progresses" in ADOM. What
fits a player's style at the beginning of the game will generally fit their
style at the end.

> with only occasional tweaks for special cases. All of
> this can be done without the added weight of micromanagement,

"Tapping a key once in a while," which is what the aggression levels system
amounts to, is not micromanagement... there's plenty of micromanagement in
ADOM, but this is not it.

> or even
> forcing the player to calculate the optimal value... simply do that work
> in the code itself and present the player with only the strong short
> term choices (ie the options that would never be desirable as a long
> term mode, the extreme exceptions).

Like Crawl's berserking? ;-)

[...]

> The catch is that it's really hard to balance this and, in my experience,
> ADoM doesn't really do a great job of that except in the early game.

I never found it unbalanced; none of the modes made you insanely hard to
dodge or insanely hard to hit if you weren't already. The real problem with
the attack modes system is that it's an extra learning load in a game that
already has a steep learning curve, yet really does not bring much extra fun
or interest for the average player.

[...]

Erik

Ray Dillinger

unread,
Feb 24, 2005, 12:20:52 PM2/24/05
to
I have an opinion about the game mechanic of combat modes.

In a normal combat, when somebody takes a whack at you, you
spend a fraction of a second parrying, dodging, or getting
your shield up to deflect the blow.

The tactic "berserk" means don't bother. Don't waste any
time at all on self-defense, just hit the opponent as hard
and as often as you can. Likewise the tactic "coward" means
use as much time as you can consistent with blocking all blows.

The greater the time penalty you take per hit, the less likely
the blows are to land.

That said, I haven't implemented it yet, and don't have meelee
attacks currently costing the victim *any* time.

Bear

Glen Wheeler

unread,
Feb 25, 2005, 1:05:52 AM2/25/05
to

"Jukka Kuusisto" <jkuu...@cc.hut.fi> wrote in message
news:jkuusist....@vipunen.hut.fi...

I think tactical modes are pretty bunt.
Much better would be actual *tactics* chosen by the player. If they want
to be a coward, then they can run away. Otherwise you end up with stupid
things like the ``cowards'' being awesome spellcasters raining death from
afar.

Antoine

unread,
Feb 25, 2005, 2:01:20 AM2/25/05
to

Glen Wheeler wrote:
> "Jukka Kuusisto" <jkuu...@cc.hut.fi> wrote in message
> news:jkuusist....@vipunen.hut.fi...
> > bwr...@csclub.uwaterloo.ca (Brent Ross) writes:
> >
> >>To my mind ADoM shows that the idea of tactical modes isn't so
great.
> >
> > I think the different tactical modes in Adom are good, but there
are too
> > many of them. Three would suffice: berserk, normal and coward.
> >
>
> I think tactical modes are pretty bunt.
> Much better would be actual *tactics* chosen by the player. If
they want
> to be a coward, then they can run away. Otherwise you end up with
stupid
> things like the ``cowards'' being awesome spellcasters raining death
from
> afar.

I agree, in as much as possible without knowing what 'bunt' means...

Tactical modes really come into their own in Guild however, when you
want to tell the characters you're not currently controlling to use
tactics such as "attack any monsters you meet" or "stay out of trouble"
or "run for your life!". It doesn't affect their stats but it does
influence what they do.

A.

Jukka Kuusisto

unread,
Feb 25, 2005, 2:11:08 AM2/25/05
to
"Glen Wheeler" <ge...@uow.edu.au> writes:

> Much better would be actual *tactics* chosen by the player. If they want
>to be a coward, then they can run away. Otherwise you end up with stupid
>things like the ``cowards'' being awesome spellcasters raining death from
>afar.

This can also be solved by making spellcasting harder (or even impossible) when
on coward mode. Same goes for other actions that would potentially abuse the
coward mode.

-Jukka
--
Jukka Kuusisto

Glen Wheeler

unread,
Feb 25, 2005, 7:31:53 AM2/25/05
to

"Antoine" <ma...@guildgame.com> wrote in message
news:1109314880.0...@z14g2000cwz.googlegroups.com...

>
> Glen Wheeler wrote:
>> "Jukka Kuusisto" <jkuu...@cc.hut.fi> wrote in message
>> news:jkuusist....@vipunen.hut.fi...
>> > bwr...@csclub.uwaterloo.ca (Brent Ross) writes:
>> >
>> >>To my mind ADoM shows that the idea of tactical modes isn't so
> great.
>> >
>> > I think the different tactical modes in Adom are good, but there
> are too
>> > many of them. Three would suffice: berserk, normal and coward.
>> >
>>
>> I think tactical modes are pretty bunt.
>> Much better would be actual *tactics* chosen by the player. If
> they want
>> to be a coward, then they can run away. Otherwise you end up with
> stupid
>> things like the ``cowards'' being awesome spellcasters raining death
> from
>> afar.
>
> I agree, in as much as possible without knowing what 'bunt' means...
>

Good old localised language.

> Tactical modes really come into their own in Guild however, when you
> want to tell the characters you're not currently controlling to use
> tactics such as "attack any monsters you meet" or "stay out of trouble"
> or "run for your life!". It doesn't affect their stats but it does
> influence what they do.
>

Which is exactly what tactics are! Seems ridiculous to set AI for the
player...unless of course we are talking about making a borg.

Glen Wheeler

unread,
Feb 25, 2005, 7:32:32 AM2/25/05
to

"Jukka Kuusisto" <jkuu...@cc.hut.fi> wrote in message
news:jkuusist....@vipunen.hut.fi...

Isn't that retrofitting a fix to a bad solution in the first place?

Jukka Kuusisto

unread,
Feb 25, 2005, 7:45:48 AM2/25/05
to
"Glen Wheeler" <ge...@uow.edu.au> writes:


>"Jukka Kuusisto" <jkuu...@cc.hut.fi> wrote in message

>> This can also be solved by making spellcasting harder (or even impossible)
>> when on coward mode. Same goes for other actions that would potentially abuse
>> the coward mode.

> Isn't that retrofitting a fix to a bad solution in the first place?

Nothing wrong with that, if it makes the bad solution a good one. After all, if
you're on coward mode, you put all your concentration in trying to evade incoming
attacks. There's no time for waving hands for spellcasting etc.

I'm not saying that tactical modes are necessary, most games do fine without them.
But I think in Adom it's a nice tactical decision when facing certain monsters and
the berserk and coward modes are justified.

-Jukka
--
Jukka Kuusisto

jasonn...@yahoo.com

unread,
Feb 25, 2005, 9:36:02 AM2/25/05
to

Brent Ross wrote:
> To my mind ADoM shows that the idea of tactical modes isn't so great.
> They essentially provide unnecessary weight for the player with
> micromanagement which could largely be covered with a transparent and
> automatically with a melee state machine. So players often ignore
> them... pushing them to progressively higher "optimal" levels as the
> game progresses, with only occasional tweaks for special cases.

Yes, a continuum from offensive to defensive is fairly boring, and
could be automated, but one could be more innovative in the dynamics of
tactics than simply offense is inversely proportional to defense.

e.g. tactic to distribute damage and defense among multiple adjacent
enemies (a la matrix reloaded).
or the converse-disadvantage against monsters back is turned to.
stunning blows- may not do as much damage, but create better
opportunity/roleplaying;
grappling/hostages ( a la goldeneye)
these may not be quite as simple to automate.

R. Dan Henry

unread,
Feb 26, 2005, 6:07:05 AM2/26/05
to
On Thu, 24 Feb 2005 09:14:17 +0000 (UTC), bwr...@csclub.uwaterloo.ca
(Brent Ross) wrote:

>In article <421d7e67$0$29276$1472...@news.sunsite.dk>,
>ABCGi <ab...@yahoo.com> wrote:
>//
>// To my mind ADoM implements a pretty useful and good example (but not
>// perfect) of different combat modes. And I have suggested something
>// similar for GREEKIE.
>
>To my mind ADoM shows that the idea of tactical modes isn't so great.

Oh, I agree. ADOM's tactical modes are the finest in the roguelike
world and they still aren't that great. Certainly 7 modes is too many.

>The catch is that it's really hard to balance this and, in my experience,
>ADoM doesn't really do a great job of that except in the early game. This
>is typical of things that are difficult to balance, as the early game is
>easy to set, but the rapid character development of most RLs makes it hard
>to keep balanced. If you really want to do this from early development,
>I'd suggest dramatically reducing the scale of character development to
>a minimul and the range of monster difficulty (ie don't go newts to dragons,
>start well into "middle" level monsters)... it will make things much easier
>to tinker with once things are up and running.

Actually, for GREEKIE, this would be a great idea anyway. Heroes
should be up against tough opponents like bands of brigands and wild
boars from the start. No fluffy bunnies or eleven types of mice. Not
that weaklings shouldn't be present, just that they should be minor
inconveniences (or innocents to protect) from the beginning.

But I still would advise against tactical modes. A "status" type
effect like Crawl berserking or Angband's fear would be acceptable
(and a PANic effect fear status would be very appropriate), but I
would reflect different fighting techniques by having different
Fighting Styles and you would normally stick with just one (although
if someone wanted to spend skill points on multiple Fighting Styles
and allow switching, I guess that could work, too). These would cover
more than dull +/- Attack and +/- Defense, but various special attacks
and defenses, preferred weapon types, etc. Different available
Fighting Styles would be a way to distinguish different races and
cultures.

>But if you really, really want modes... I'd suggest keeping them
>abstract and not as direct stat tradeoffs (ie "defensive", "normal",
>"aggressive"... no specifics and no inbetween points to complicate your
>development further).

I might even limit to two if I had modes: aggressive and defensive.
The lack of a so-called "normal" mode will at least get everyone to
make a decision and think about what style of play they want, even if
few bother to switch back and forth. I don't really believe there is
some perfectly balanced style; that you either think as an attack or
as a defender.

R. Dan Henry
danh...@inreach.com

Raymond Martineau

unread,
Feb 26, 2005, 9:50:10 PM2/26/05
to
On Sat, 26 Feb 2005 03:07:05 -0800, R. Dan Henry <danh...@inreach.com>
wrote:

>On Thu, 24 Feb 2005 09:14:17 +0000 (UTC), bwr...@csclub.uwaterloo.ca
>(Brent Ross) wrote:
>
>>In article <421d7e67$0$29276$1472...@news.sunsite.dk>,
>>ABCGi <ab...@yahoo.com> wrote:
>>//
>>// To my mind ADoM implements a pretty useful and good example (but not
>>// perfect) of different combat modes. And I have suggested something
>>// similar for GREEKIE.
>>
>>To my mind ADoM shows that the idea of tactical modes isn't so great.
>
>Oh, I agree. ADOM's tactical modes are the finest in the roguelike
>world and they still aren't that great. Certainly 7 modes is too many.

Adom simply uses two tactical modes, along with five intermediate
gradiations. There's nothing inherintly wrong or complex with this, as
it's just an assignment on how aggressive or defensive you wish to be.

Of course, there is a flaw with Adom's implementation: changes in tactics
take effect instantly - I wouldn't be suprised if a person writes a script
that flips back and forth between "Coward" and between attacking and moving
normally - it gives the benefits of the +DV bonus without really earning
them.

>>But if you really, really want modes... I'd suggest keeping them
>>abstract and not as direct stat tradeoffs (ie "defensive", "normal",
>>"aggressive"... no specifics and no inbetween points to complicate your
>>development further).
>
>I might even limit to two if I had modes: aggressive and defensive.
>The lack of a so-called "normal" mode will at least get everyone to
>make a decision and think about what style of play they want, even if
>few bother to switch back and forth.

This looks okay - two tactics is good enough for almost everything.
Personally, I would also include a stance that provides limited combat
capability but slightly increase movement speed, such as running or
sprinting. (This might not work in all cases, but shouldn't be a problem in
the Angband or Adom engines.)

> I don't really believe there is
>some perfectly balanced style; that you either think as an attack or
>as a defender.

Likewise, there shoudn't be a perfect attack or defensive style (unless you
are playing with missile weapons.) Thus, I'd suggest that advancements in
one of the two styles should affect the other by a small amount.

ABCGi

unread,
Feb 28, 2005, 9:52:47 PM2/28/05
to

Nothing's broken... to sit from afar raining death on the enemy and
doing everything to avoid being hit is cowardly.

Brent Ross

unread,
Mar 1, 2005, 12:39:06 AM3/1/05
to
In article <386b2mF...@individual.net>,
Erik Piper <efrn...@sdky.cz> wrote:
// bork bork bork Brent Ross bork 10:14:17 AM bork 2/24/2005 bork bork:
//
// > with only occasional tweaks for special cases. All of
// > this can be done without the added weight of micromanagement,
//
// "Tapping a key once in a while," which is what the aggression levels system
// amounts to, is not micromanagement... there's plenty of micromanagement in
// ADOM, but this is not it.

It's not "in your face" micromanagement (because it can be avoided for
the most part), but it's still ever present and far too high
granularity. It's always taunting you with the fact that maybe you
should probably have tweaked up or down for that last slightly OOD
monster. I just always seemed to me that that level of miniscule
repetitive task is exactly the sort of thing that makes me write Perl
scripts (when dealing with files, data, or computation), so why
shouldn't the game just do all that work for me and present me with only
the /real/ options.

// [...]
//
// > The catch is that it's really hard to balance this and, in my experience,
// > ADoM doesn't really do a great job of that except in the early game.
//
// I never found it unbalanced; none of the modes made you insanely hard to
// dodge or insanely hard to hit if you weren't already.

That's only one side of the balance (the over-weight, break the game
side). In order to balance something you also need to look at the other
side (the under-weight, pointless factor side). These for a window
(since it's typically got at least a bit of leeway), in which you want
to be. With melee you typically have a moving target here... the PC
typically gains in ability and the monsters need to match that in order
to keep the game from sliding one way or the other. That's tricky
enough without adding a sliding window on top of that, but that window
may have to be additionally balanced for magnitude of it's effect
(since, as the scale of the game grows, so does what the player will
consider a valid factor). The added variable factors can simply be too
much of a pain if you want to be able to guarantee and say things about
the PC's capabilities at any point in the game. It's simply not worth
the effort to try and create some fungible stat relationship that stays
meaningful, while not disrupting you design intents. It's better to
take your design intents and try to make them come out in other ways...
like using them to help make important internal choices that would be
painful to inflict upon the player. For example, if shield blocks have
a cost, do you really want to bother the player with prompting on each
incoming attack or having them toggle the monsters, monster classes, and
attacks to block? This gives a very limited and controlled access to
the internal machine... and, at least, it doesn't have the min-max feel
of looking through a list of point tradeoffs.

Brent Ross

Brent Ross

unread,
Mar 1, 2005, 12:52:45 AM3/1/05
to
In article <U1oTd.8070$m31.1...@typhoon.sonic.net>,
Ray Dillinger <be...@sonic.net> wrote:
// I have an opinion about the game mechanic of combat modes.
//
// In a normal combat, when somebody takes a whack at you, you
// spend a fraction of a second parrying, dodging, or getting
// your shield up to deflect the blow.
//
// The tactic "berserk" means don't bother. Don't waste any
// time at all on self-defense, just hit the opponent as hard
// and as often as you can. Likewise the tactic "coward" means
// use as much time as you can consistent with blocking all blows.

This is pretty much what I was talking about. You have a real cost
factor, with the decision being mostly in the hands of the PC, not
the player. Exactly the sort of thing you want to give the player
a bit of control over (but not so much that it becomes a burden).

One comment I'd like to make about this implimentation is that using
speed makes it extra tricky to work with. It's a big cost, but you end
up needing to worry about the PC getting into a semi-starvation mode
when it comes to player inputs. Simply limiting the number of blocks
(perhaps by having a maximum amount of time allowed between player
commands) is one small solution.

Another solution is to capture the same idea in a different way. For
example, by tracking how far the PC is "out of position" and, instead of
making that cost extra time, make that apply penalties on the next
round. This should at least be easier to work with balance wise as
time tends to be a tricky factor to work with in RLs since a little can
have a very large effect.

Brent Ross

0 new messages