To anyone who has written a game...

Question: Did you enjoy playing it yourself? Games I wrote, I knew too much 
about to enjoy playing them myself, other than testing they were right. 
Indeed, sometimes the more I knew about a game, the less fun it was. (Like, 
once I learned the production rules in detail for M.U.L.E., it was less 
exciting.)

Premise: Infinite spawning of enemies on a level leads to less replayability 
in a complex shooter game with lots of ways to win. This is because the 
player cannot set arbitrary goals such as "clear the level using nothing but 
the wrench/crowbar/whatever".   How fun would Thief be if the loot spawned 
randomly when you weren't in the room?

Minor curiosity: Bioshock - is it qualitatively different on different 
difficult levels, like Thief is, or is it just a rebalancing of hit-points 
and damage levels?

-- 
   Darren New, San Diego CA, USA (PST)
   Insanity is a small city on the western
   border of the State of Mind.

Post a reply to this message

Darren New <dne### [at] sanrrcom> wrote:
> Minor curiosity: Bioshock - is it qualitatively different on different 
> difficult levels, like Thief is, or is it just a rebalancing of hit-points 
> and damage levels?

  I think the vast majority of games with different difficulty settings use
just "fake difficulty", ie. some really cheap tricks such as giving the
enemies more HP and/or reducing the damage of your weapons (some games
simply spawn more enemies at higher difficulty levels, Doom being the most
obvious example), rather than going so far as adjusting the enemy AI.

  The most blatant fake difficulty I have seen is in Doom 3. (Or was it
in Quake 4? I always confuse the two because I have both and they are
so similar in so many ways. Well, it's not all that important.) In the
hardest difficulty level your health is always drained to half (or was
it even to one third?), even if you get it temporarily higher with health
packs. This feels like a really cheap trick.

  I suppose that the reason for this is rather simple: Adjusting the game AI
to consistently correspond to some kind of difficulty level is probably a
really hard problem, and the game developers have their hands full of work
already with everything else that needs to be done by the deadline and under
the given budget to worry about such irrelevant things.

  About the only game type where difficulty level directly correlates to AI
strength are board game programs, chess programs being the most prominent
example. (There might be other examples in some sports and strategy games.
I don't play many of those, so I really don't know.)

  Of course difficulty can sometimes be adjusted by other means, such as
in the Thief games: In harder difficulty levels you have to achieve more
(eg. collect more loot), and you are given rational limitations (eg. don't
kill anybody, which is a very logical limitation in a stealth-based game).
While these feel less like "cheating" from the part of the game designers,
it still doesn't mean it's not a cheap trick.

-- 
                                                          - Warp

Post a reply to this message

"Warp" <war### [at] tagpovrayorg> wrote in message
news:4a44a1ae@news.povray.org...
> Darren New <dne### [at] sanrrcom> wrote:

>   I suppose that the reason for this is rather simple: Adjusting the game
AI
> to consistently correspond to some kind of difficulty level is probably a
> really hard problem, and the game developers have their hands full of work
> already with everything else that needs to be done by the deadline and
under
> the given budget to worry about such irrelevant things.
>
>   About the only game type where difficulty level directly correlates to
AI
> strength are board game programs, chess programs being the most prominent
> example.

I might be wrong but all chess games I used to play used analogous fake
difficulty / cheap tricks too, in that instead of using a different,
"smarter" engine, they use the same engine but allowed it longer time to
calculate ahead (alternatively, they artifically limit the calculation
time/level for lower difficulty levels). Coding multiple chess engines for
different difficulty levels doesn't make any more economic sense than coding
multiple AI's in action games. Of course the parameters to tweak are
different in action games than in board games, but I don't think it's fair
to say that longer times (or more properly, more processor cycles) allocated
for a given task makes the AI stronger.

Post a reply to this message

somebody <x### [at] ycom> wrote:
> I might be wrong but all chess games I used to play used analogous fake
> difficulty / cheap tricks too, in that instead of using a different,
> "smarter" engine, they use the same engine but allowed it longer time to
> calculate ahead (alternatively, they artifically limit the calculation
> time/level for lower difficulty levels).

  That's precisely fine-tuning the AI parameters to match the difficulty
level.

  There's not really such a thing as a "smart" engine. There's only number
crunching: Crunch more numbers and the program will be a stronger opponent.
Crunch less numbers and it will start making more mistakes and poorer moves
for the simple reason that it's not seeing far enough. Not completely unlike
human players: A beginner will see only one or maybe two moves forward, and
this for only a few pieces. A pro will see a dozen or more moves forward
(more for the relevant pieces, less for the irrelevant ones).

  There are a few chess engines out there which at easiest difficulty
levels will make deliberate mistakes. In other words, rather than blindly
choosing the best move it has seen so far, it deliberately at times chooses
a slightly worse one, just to lower the difficulty level. This is also
adjusting the AI (in a way that it sometimes ignores better moves as if it
had never even seen them).

> Coding multiple chess engines for
> different difficulty levels doesn't make any more economic sense than coding
> multiple AI's in action games.

  I never talked about multiple AIs. I talked about tuning the AI strength
according to the difficulty level.

-- 
                                                          - Warp

Post a reply to this message

Warp wrote:

>   I think the vast majority of games with different difficulty settings use
> just "fake difficulty", ie. some really cheap tricks such as giving the
> enemies more HP and/or reducing the damage of your weapons (some games
> simply spawn more enemies at higher difficulty levels, Doom being the most
> obvious example), rather than going so far as adjusting the enemy AI.

In most cases, yeah.

(It's kind of crazy playing Quake II on the highest difficulty level. 
You run up to some fleshy cyborg on put 4 depleted uranium slugs into 
him and he's not dead. WTF?)

I think the original HalfLife did this fairly well. On the lower 
difficulty levels, the enemies tend to just run at you, while on higher 
difficulty levels they seem to behave more tactically.

For example, the headcrabs. On Easy, they just wander towards you until 
you're in jumping range. On Hard, they tend to scuttle into corners and 
wait for you to walk past, and then ambush you.

Similarly, the marines on Easy just shoot at you a lot. On Hard, then 
throw grenades, duck behind things, and try to run round and flank you. 
(I must say, I didn't find anything "revolutionary" about the game AI - 
one of the things that was highly touted. It was good, but not fantastic.)

Of course, even in HalfLife, they still used the main principle of 
increasing the hitpoints of all the enemies, which is dull and annoying. 
But at least where they add extra enemies, they put them in strategic 
places. Indeed, walking round the level on Easy, you can tell where the 
extra enemies are probably hiding. "Oh look, a useless corner that 
doesn't go anywhere. I wander what *that* might be for?"

Having struggled to complete HalfLife 2: Episodes 1 and 2 on the very 
lowest difficulty setting available, I'm not sure I want to know what 
it's like on Hard. :-/

In particular, most of the hard things about HalfLife 2 and the two 
episodes are overwhelming forces. It started with Ravenholm, an entire 
level teaming with zombies and almost no ammo. Sure, killing zombies 
with only a crowbar is pretty damned hard. But it's not exactly 
"enjoyable" replaying the same 15 seconds of game hundreds of times over 
until you manage to get past.

In short, a game that's too easy is boring, but so is a game that's too 
hard, and Ravenholm was absurdly hard.

In HL2:EP1 we find Lowlife, a similar premise really. Except somebody 
thought it would be funny to turn off all the lights. It's not big, it's 
not clever, and it's not even remotely enjoyable. It's repetative and 
tedious, actually.

And of course, HL2:EP1 rounds off with a battle with an almost 
unkillable Strider. I mean, seriously. How many rocket hits does it 
take?? What the hell is it MADE OF?!

This isn't a challenge, it's silly.

HL2:EP2 is somewhat similar with the two antlion guards. It's basically 
a big soft squidgy thing which you never the less have to dump ludicrous 
amounts of ammo onto before it will freakin' DIE. Several times I played 
that section and actually ran completely out of ammo. As in, the only 
thing I had left was the crowbar. I had to go run outside like a crazy 
thing hoping there might be some ammo around somewhere...

Still, at least HL2:EP2 managed to be challenging in more interesting 
ways. Right up until the battle for White Forest... *sigh*

Post a reply to this message

Invisible <voi### [at] devnull> wrote:
> Having struggled to complete HalfLife 2: Episodes 1 and 2 on the very 
> lowest difficulty setting available, I'm not sure I want to know what 
> it's like on Hard. :-/

  The final battle of Episode 2 is HARD. Really HARD. Even when you get
the idea of what you have to do, it's still really hard.

  But it can be beaten.

-- 
                                                          - Warp

Post a reply to this message

>> Having struggled to complete HalfLife 2: Episodes 1 and 2 on the very 
>> lowest difficulty setting available, I'm not sure I want to know what 
>> it's like on Hard. :-/
> 
>   The final battle of Episode 2 is HARD. Really HARD. Even when you get
> the idea of what you have to do, it's still really hard.
> 
>   But it can be beaten.

Oh, I beat it eventually. It took about 2 months though.

It wouldn't surprise me if they released a game update to turn down the 
difficulty, and that's why I beat it. :-P

(I actually sent an email to Gab### [at] valvecom about it. Not that I 
suppose Gabe actually reads those...)

The car is... uncontrollable. It accelerates very slowly, yet hitting a 
twig or a pebble is sufficient to bring it to a complete halt. At top 
speed, it handles exactly the same way as in first gear (i.e., it will 
still make 90 degree turns without hesitation). In short, the physics 
are completely fschk'd up. The stearing controls are of course almost 
unusable

And then, we have hunters. They're wickedly manuverable, and every time 
you meet one earlier in the game, you have to almost exhaust your entire 
ammo supply to kill just ONE of them. But out here, you suddenly have 
dozens of the running around, AND NO AMMO ANYWHERE. (Sure, a few of the 
huts have tiny quantities of ammo. But once they've been blown to hell, 
that won't help you too much.)

The number of times I've tried to ram a hunter with the car, only for it 
to nimbly sidestep me at the last second is nobody's business. And then 
of course the car hits a pebble and comes to a total halt, and it takes 
10 minutes to execute a 2,000-point turn and get back on the move. 
Assuming you're not dead by this point from the multiple hunters and 
strikers attacking you.

On top of that, it's almost impossible to figure out where the hell the 
enemy is. And on top of THAT, to kill a strider you've got to select the 
gravity gun, pick up a bomb from somewhere, throw it, go through your 
inventory, select the pistol, aim at the strider... oh, wait, too slow. 
Sorry. Try again.

When I *finally* destroyed the last strider (which, by the way, was 
inches from the base at this point), everybody started cheering. I was 
so certain that there was going to be another wave... but, as it turns 
out, there wasn't.

Similarly, as they launched the missible, I was certain that a bunch of 
combine soldiers would suddenly jump out or something, or the G-Man pop 
up and say "I'm s-ssorrry Dave, I c-can't llet you dooo that..."

But no, it' launched, and it seemed to work. (Again I was half-expecting 
it to go wrong and teleport me and Alyx back to the Xen homeworld or 
something weird.)

And then... well, you've all seen it, right? ;-)

Post a reply to this message

Warp wrote:
> While these feel less like "cheating" from the part of the game designers,
> it still doesn't mean it's not a cheap trick.

Hmmm. Well, I liked that you could not meet all of the goals in the hard 
level of thief by playing on the easy level. There was loot you had to 
collect on the hard levels that just weren't *there* in the easy level. The 
door to the hall that led to the part of the castle with loot #3 was a blank 
wall on difficulty #1 or #2. You literally couldn't play the easy mode and 
see the whole level.

That doesn't seem like too much of a "cheap trick" to me. But then the AI 
wasn't really what you'd call stunning.

Plus, of course, there were things like not needing to get back out again 
once you got the loot, on the easy levels. And you had less health, and the 
bad guys had better hearing, and so on, on the hard levels, as expected.

-- 
   Darren New, San Diego CA, USA (PST)
   Insanity is a small city on the western
   border of the State of Mind.

Post a reply to this message

Warp wrote:
>   I never talked about multiple AIs. I talked about tuning the AI strength
> according to the difficulty level.

Actually, I saw a research paper where they did focus group tests, and when 
all they did is turn up the number of HP the computer opponents had, players 
thought the AI was smarter, even tho it was exactly the same code shrugging 
off more bullets.

-- 
   Darren New, San Diego CA, USA (PST)
   Insanity is a small city on the western
   border of the State of Mind.

Post a reply to this message

Darren New escreveu:
> To anyone who has written a game...
> 
> Question: Did you enjoy playing it yourself?

I never did, but I guess being Sid Meier, Kojima, Miyamoto or Peter 
Molyneaux would suck too much wasn't it for the vast amounts of money 
they get.  I mean, not being able to enjoy such masterpieces because you 
are too spoiled already??!

> Premise: Infinite spawning of enemies on a level leads to less 
> replayability in a complex shooter game with lots of ways to win. This 
> is because the player cannot set arbitrary goals such as "clear the 
> level using nothing but the wrench/crowbar/whatever".   How fun would 
> Thief be if the loot spawned randomly when you weren't in the room?

How about people who just enjoy mindless shooting carnage for the sake 
of it rather than specific goals? :P

-- 
a game sig: http://tinyurl.com/d3rxz9

Post a reply to this message

I mean, I never did MAKE a game... :P

nemesis escreveu:
> Darren New escreveu:
>> To anyone who has written a game...
>>
>> Question: Did you enjoy playing it yourself?
> 
> I never did, but I guess being Sid Meier, Kojima, Miyamoto or Peter 
> Molyneaux would suck too much wasn't it for the vast amounts of money 
> they get.  I mean, not being able to enjoy such masterpieces because you 
> are too spoiled already??!
> 
>> Premise: Infinite spawning of enemies on a level leads to less 
>> replayability in a complex shooter game with lots of ways to win. This 
>> is because the player cannot set arbitrary goals such as "clear the 
>> level using nothing but the wrench/crowbar/whatever".   How fun would 
>> Thief be if the loot spawned randomly when you weren't in the room?
> 
> How about people who just enjoy mindless shooting carnage for the sake 
> of it rather than specific goals? :P
> 


-- 
a game sig: http://tinyurl.com/d3rxz9

Post a reply to this message

nemesis wrote:
> How about people who just enjoy mindless shooting carnage for the sake 
> of it rather than specific goals? :P

Well, sure. And Serious Sam is still one of my favorite speed-runs to 
exercise to. :-)

-- 
   Darren New, San Diego CA, USA (PST)
   Insanity is a small city on the western
   border of the State of Mind.

Post a reply to this message

Warp wrote:
>   About the only game type where difficulty level directly correlates to
>   AI
> strength are board game programs, chess programs being the most prominent
> example. (There might be other examples in some sports and strategy games.
> I don't play many of those, so I really don't know.)

In chess, adjusting the difficulty may mean adjusting the depth used in
the 'tree' of possible decisions. That is, not looking too far into the
future.

I have also seen algorithms that take exponential time or similar, and the
program sets a time limit. When the time limit is reached,
the 'best-so-far' move is performed. Increasing the time limit may let the
computer make better moves.

(Disclaimer: I don't know how to play chess. I know the rules, I know how
each piece can be moved, but that's not the same as "knowing how to play")

Post a reply to this message

Warp <war### [at] tagpovrayorg> wrote:
>   There are a few chess engines out there which at easiest difficulty
> levels will make deliberate mistakes. In other words, rather than blindly
> choosing the best move it has seen so far, it deliberately at times chooses
> a slightly worse one, just to lower the difficulty level. This is also
> adjusting the AI (in a way that it sometimes ignores better moves as if it
> had never even seen them).

When I was working on my triangular chess game, I experimented with different
methods of "rating" a given board, and using those ratings to come up with
compound ratings for the different possible moves.

At that point, though, I used fuzzy logic to pick the move the computer would
use.  Adjusting the "fuzziness" made for a similar effect, where sometimes you
would see the PC make a move that just made you go, "Wha?"

Overall, it wasn't as effective as traditional brute-forcing (I only went about
ten levels deep on the highest difficulty setting), but it also felt more
satisfying to program.

....Chambers

Post a reply to this message

Nicolas Alvarez wrote:

> (Disclaimer: I don't know how to play chess. I know the rules, I know how
> each piece can be moved, but that's not the same as "knowing how to play")

Knowing the rules doesn't really help. Trust me.

Exhibit A: Go.

The rules are trivial, and yet it is impossible to win.

(Disclaimer: By "impossible" I obviously mean "so absurdly difficult 
that the difference between this and geniunely impossible is so minute 
as to be insignificant".)

-- 
http://blog.orphi.me.uk/
http://www.zazzle.com/MathematicalOrchid*

Post a reply to this message

Orchid XP v8 <voi### [at] devnull> wrote:
> Nicolas Alvarez wrote:
>
> > (Disclaimer: I don't know how to play chess. I know the rules, I know how
> > each piece can be moved, but that's not the same as "knowing how to play")
>
> Knowing the rules doesn't really help. Trust me.
>
> Exhibit A: Go.
>
> The rules are trivial, and yet it is impossible to win.
>
> (Disclaimer: By "impossible" I obviously mean "so absurdly difficult
> that the difference between this and geniunely impossible is so minute
> as to be insignificant".)

You should play me.  You'd probably win ;)

....Chambers

Post a reply to this message

> Question: Did you enjoy playing it yourself? Games I wrote, I knew too 
> much about to enjoy playing them myself, other than testing they were 
> right. Indeed, sometimes the more I knew about a game, the less fun it 
> was. (Like, once I learned the production rules in detail for M.U.L.E., it 
> was less exciting.)

I spent quite a long time enjoying playing against some basic AI I wrote for 
a racing game once.  The AI was quite basic, essentially following a set 
path "roughly", and having other input from the location of other people 
nearby.  After I had tweaked the power that the AI had, it was quite fun to 
race against them.  In the same game, usually when I was meant to be testing 
out some other feature, I usually ended up doing lap after lap trying to 
beat my best lap time :-)

I also made a very simple 2D game a long time ago where you had to maneuver 
an articulated lorry (semi truck) to make deliveries at various stops 
without hitting the scenery.  That was pretty fun to play, and taught me 
that it's far easier to reverse with longer trailers than shorter ones :-)

Post a reply to this message

>  The most blatant fake difficulty I have seen is in Doom 3.

In FarCry it's pretty poor too.  I usually played it on one of the easier 
levels, but then one day decided to do it on the hardest.  I gave up after 
about 1 minute, because I shot a guy in the head from about 1 metre away 3 
or 4 times and he just stood there preparing to shoot me.  Eventually he got 
round to shooting me and I died instantly.  Firstly, if I shoot someone in 
the head I at least expect them to move a bit, and secondly if a "hard" AI 
is standing right in front of me, I expect him to shoot me almost instantly, 
not take 2 seconds to realise I am there.

>  About the only game type where difficulty level directly correlates to AI
> strength are board game programs, chess programs being the most prominent
> example. (There might be other examples in some sports and strategy games.
> I don't play many of those, so I really don't know.)

In most car racing games it's really hard to get the AI correct.  In most 
games they just give the cars lower power to simulate worse drivers, but 
that just makes them really easy to overtake on the straight parts of the 
track.  Some games make the easy AI brake earlier than the hard AI, but 
again once you know this they are easy to beat.  AFAIK no racing game 
actually simulates very well how real drivers are better than others.  There 
is a quote from some racing driver, which states that anybody can drive a 
car fast in a straight line, and anybody can drive a car fast in a curve, 
the skilful bit is the transition between the two.

Post a reply to this message