Slashdot Mirror
On Luck and Randomness In Games
Gamasutra has an article analyzing random events in games, and how they can add or subtract to a player's experience. It looks at the different ways luck plays a part in games; from landing a critical strike instead of a miss to the scatter of a shotgun blast to waiting for that blasted straight piece in Tetris. "Game developers are sometimes faced with similarly challenging decisions when contemplating whether to include some kind of deliberate randomness. For example, in the video game Unreal Tournament, when a player shoots at a target with the 'enforcer' weapon, the projectile does not necessarily hit the point that is aimed at; a random deviation is added that scatters shots. This introduces a degree of realism from an observer's perspective and no doubt gives beginners a fair chance against more experienced players, but it can also potentially frustrate skilled players."
Its always good to see that people who matter are actually thinking about ways to overcome obstacles.
It also annoys me greatly when a steady handed and well aimed sniper round misses by a algorithm calculated bees proverbial.
--Question Authority--
I used to play FFXI, but one of the things that eventually drove me away from the game was the randomness of nearly everything. It was irritating to go nearly 0/10+ (and I've heard horror stories of worse) on rare item drops while Billy teh n00b would get it on first drop. Oh, and there's nothing more fun than fighting a hard fight and getting nothing as a reward. I can understand the developer's desire to keep certain items rare, but such low drop rates aren't the way to do it. I would have preferred they made the fights harder, not more random.
I recently had the opportunity to hear Sid Meier talk about random events in the Civilization series. It is unfortunate that this article doesn't mention any of his insights regarding player's psychology when it comes to "luck".
Apparently the average player expects to win regularly, even if probability allows for long strings of losses. If you lose two even fights in a row in a game of Civilization, you are literally guaranteed to win the third, IIRC. This is how their "karma" system is implemented.
Additionally, players expect a fight of 30 vs 20 to be much more of a sure thing than a fight of 3 to 2, even though the ratio is the same. Apparently you ought to get some sort of boost when the numbers are higher in order to satisfy most players. This actually makes a degree of sense to me, because I would expect the variance to be less in the first case, but he didn't address the issue and I didn't ask.
This article gives an interesting categorization of the types of randomness and luck that can exist in games, but it appears to do little to address these ideas. This is too bad, really. It might be interesting to see how these hacks affect these probabilistic features of Civilization according to their charts.
Insert self-referential sig here.
Look, game developers cannot have it both ways. Either they make it possible to have newbies have a chance against the experts and annoy the experts or they make it a game of skill and drive new users away from the experts. It sucks for the experts but they do find ways to win anyways if they have any real skill at all.
in UT 99 and 2004 and probably the rest there's also that chain gun weapon. On regular mode that thing shoots like Helen Keller with a paintball gun. Seriously at about 20 feet they can be up to about 5 feet off target. But you don't rely on luck yo hit what you're hitting from a long distance with that gun, you do what you're supposed to and only use it short range, duh! That's why it shoots so poorly, it's a short range weapon! Same the the enforcer really. Use the sniper or lightning gun or even the shock rifle to shoot long distances cuz that's what they're made for. I also can't help but mention that luck when it comes to rare drops in MMORPGs SUCKS! I know people who got 3 SOX items in one day and it took me 2 months to get one. It makes you feel like the game hates you.
Google's Super Secret Search Algorithm: SELECT @search_results FROM internet WHERE @search_results = 'good'
For example, in the video game Unreal Tournament, when a player shoots at a target with the 'enforcer' weapon, the projectile does not necessarily hit the point that is aimed at
Personally I think it does the exact opposite. I think Far Cry 2 *may* have done this. But if I line up a head shot (sniper) and put a bullet in the AIs head and he doesn't die, then this makes it seem far less realistic to me--especially when I let loose two shots to be sure and then aim down for a direct body shot and the guy still somehow manages to stand.
Randomness is good, but I don't think making bullet paths random is great. Sure, in real life there is random wind and other influences (projectile shape/smoothness, the barrel, and all that), but at the distances (and speed of projectile) I am talking about it's negligible. Two direct head shots and a just-for-fun/'cause-I-can body shot in quick succession should not fail just to add 'randomness'.
A skilled hunter can still miss a shot, and will be frustrated when the game gets away. If realism is the goal, getting the perfect shot the first time is not the way to go.
Game devs would be well advised to remember they are creating a G A M E - not an alternate reality.
They are subject to the same limitations as story tellers, song writers and actors...their imaginations.
You think that's random? Try playing Wesnoth!
I think realism is a good thing. It's scary how controllable some people think war is.
In the real world, SWAT members with 20 years experience slip on pieces of gravel and accidentally shoot their team members. Rangers walking through the forest push a branch that swings back and stabs someone else in the eye. These aren't freak occurrences - this sort of thing happens on every other mission. These guys are elite because for the rank and file, it happens on a daily basis.
A high percentage of Russian soldiers are drunk at any given point in time in combat - even tank drivers. A percentage of soldiers will stand idly by while a rookie gets himself killed, because they never liked him in the first place. Some hand grenades have bad fuses and immediately blow your hand off. A sniper will be staring at a butterfly and forget he's in a war until he's shot. While patrolling through an alley, you might find some guy getting a blowjob from a schoolboy.
Any predictability is unrealistic. The veterans are the ones who know insane ass, brain melting shit is around every corner, and are prepared to deal with it. It's better to be lucky than good.
To me as an individual i always enjoy it when when a random event occurs and it helps me. Conversely i always hate it when a random event occurs that hinders me. Im not sure why more game designers dont realize this simple fact of human nature. A real world comparison of this would be doing some auto repair. Some individuals enjoy it most individuals would be happy if they found 10 bucks under there car once they had it up on a jack. Not too many people would be too happy if some jerk randomly came over and kicked the jack out from under the car. Particularly if you were still under there
Example 4 - A player of World of Warcraft shoots accurately and delivers a Critical Strike. (Once a strike is successfully inflicted on an opponent within World of Warcraft, it has a probability-based chance of inflicting double damage; any such Critical Strike that occurs is reported to the player by an on-screen text message.)
Except that isn't true. The result of an attack is derived from a single roll. It gives rise to the property of defense being able to 'push' critical strikes off of the combat table by raising the chance to be missed, as the roll needed to score a critical cannot occur.
Yes I have no life.
So much about how to use and manipulate randomnes to fit all tastes, and so little about randomness itself.
Some people simply like randomness. Some people enjoy a game of "highest result in the dice wins". Some people hate go or chess because they lack randomness.
Then there are people who like wading through a higly random environment in games like poker, where the number of hands reduce the randomness to a homogeneous atmosphere.
It's not possible to make a single game be go and dice. You can't add both randomness and it's lack to a game, to appeal every target.
waiting for that blasted straight piece in Tetris
Random? That isn't random. It comes right after you block off the slot you were saving for it.
How can you have an article on "Luck and Randomness In Games" without at least mentioning rogue/nethack?
God damn mother fucking CRIT ROCKETS!
At least imho, apart the randomness of critical strikes, a lot of weapon/spells have a too wide range of damage. Reducing that range of damage (keeping the average damage constant but setting the max and min damage towards the average) would balance a lot the game. Then we have even critical strikes. As a rogue player the latest season I had chances to down =mail armoured opponents in arena in 10 seconds. Some other time I was so unlucky that I could only pray to be defeated fast. Cheers,
How lucky for you!
Valve likes to implement the random factor into their games a lot lately, much to my dismay. Going from Day of Defeat to DOD:Source had a large dumb-down to the gameplay, and far more random weapon spray to benefit newer/bad players.
Similar could be said about Counterstrike after the first few post-release patches. It was originally a good mix of realism and deathmatch with a Rock-Paper-Scissors balance. With the patch after 1.3 it basically became "Riflestrike" if you actually wanted to be competitive.
Then of course the random death to criticals in Team Fortress 2...
Of course the enforcer is random. The chance of the enemy surviving is pretty high.
OTOH the chance of survival of enemy surviving a hit by a volley of 6 rockets is exactly zero, no matter what his armor, health or other bonuses. That's why I don't use enforcer.
45 5F E1 04 22 CA 29 C4 93 3F 95 05 2B 79 2A B2
...the primary killers of motivation to explore.
Why should I climb the tallest tower in the furthest castle, if I get the same stuff as from the chest behind the entrance door?
Why should I conquer the strongest enemies and explore their castle if I'm better off killing millions of rats, then open a chest in the tavern cellar?
45 5F E1 04 22 CA 29 C4 93 3F 95 05 2B 79 2A B2
I can only add small tidbit from developer pov.
Actually few games are truly random at low level. On many occasions I was faced with trivial matter that "pseudo" random numbers are not really all that random.
The mentioned above problem is easily visible in aforementioned Tetris example and is direct result of poorness of random numbers.
On one occasion, analyzing one source code, I have found clever trick with premade chains of "random" numbers: applying the chains twice (two level indexing), overall the random generator had for numbers in 0..127 range loop far above 32k and all delivered numbers were well distributed: as per definition there were no same number streaks longer than 3.
As gamer though, I'd say, too much randomsness is often very distracting. Some console RPGs (e.g. mentioned above FF) are especially affected. If you look at real life, few events around us are really random...
As fan of SRPG and strategical games in general, I prefer games where random events are normally consequences of your past actions. Or even games like Heroes of Might and Magic where there are essentially no randomness in battles (or it is always in known bounds) so that outcome largely depends on player's capabilities.
All hope abandon ye who enter here.
Really?
I don't think so. The National Rifle Association in the United States has slapped their logo on various games that did a very good job of replicating the real-world problems involved in shooting accurately. IIRC, there have been games covering benchrest, hi-power, and varmint (if you're a shooter, you'll know specifically what those are).
Those games got panned by the gaming press as boring, boring, boring. They actually required people to think, account for all the variables, and then be satisfied when they merely score a correct hit. Just like real life. The gamers, however, wanted things to move faster. They wanted more Bang! Splat! Oof! to go with their game play. The notion of actually taking something close to real-life time to set up a shot was, to them, just needless tedium.
So, no, I don't think if you make the difficulty of in-game shooting more accurately mirror real life, gamers will be happy.
Then again, if you give them an infallible auto-aim, they'll complain about that, too.
Hmmm.
I'm really glad my livelihood doesn't depend on making decisions about these kinds of things.
I'm more annoyed with item drops in MMO's. Adding randomness to FPS games and using probability to do miss/critial strikes makes sense. It doesn't make the game so easy it's boring while adding a bit of realism. It is much cheaper (as in CPU time) to random a shot than to add wind and gravity into your game and compute a ballistic path.
As for item drops, when you're in a group and you have to hope the item you want drops and then hope you win it, it gets frustrating; especially when everyone else is getting stuff and you're "unlucky." I used to play Final Fantasy XI and this is one of the reasons I quit. Though, they did add a game element called Assault where after beating a level you earn Assault Points. Each area has a set of items you can purchase with Assault Points. Clear enough areas and collect enough points, you can buy whatever item you wish. No hoping it drops, no praying that you out-lot another player, no bitching and moaning from people who say they get more use out of a specific item than you do.
-SaNo
To "clock" golden eye 007 on N64 you had to pass every stage in a certain amount of time... The time given for the facility stage was 2:10... This was so low to actually pass it you needed 2 random events to happen. 1 - You had to collect a key to a door off a scientist. He would randomly appear in 4 different places, but you could only pass the stage in 2:10 if he was in 1 of the positions, (the one least out of your way) and 2 - you did not have time to stop at all or really even to turn around and shoot backwards, so you needed to get a crapload of headshots and take out as many enemys as possible so you would actually survive the run till the end, not easy with a controller. So not easy I consider it to be a random event with luck involved...If you missed a few head shots, youd have enemies chasing you shooting you from behind as you bolted through the stage... sometimes you could make it and sometimes you couldnt...even if you did make the 3/4 of the way in to meet the scientist he wasnt there most of the time to give you the key and just not enough time to go to any of the other positions he might be in...It took sooooo many attempts, but I beat that sucker!!!!!!! If i remember right that unlocked invincibility cheat lol...
Sure there is. It's called limited computing resources. Collision tests involving parabolas and volumetric effects are far more costly than simple line-primitive collision tests.
Then don't use the parabola in the collision test. Compute the parabolic trajectory 60 times a second, and then use the line segment between this frame and the next as the path for collision tests.
So, for that matter, is much military activity. Most games provide entertainment, not a simulation of life. All a game needs to do is be honest about the level of accuracy offered. It's not a question of "right" or "wrong," but a question of which target audience you're aiming for.
I piss off bigots.
waiting for that blasted straight piece in Tetris
Random? That isn't random. It comes right after you block off the slot you were saving for it.
Then why didn't you use the hold box to save an I tetromino? And why aren't you exploiting the fact that pieces come in groups of seven, one of each shape? You must be thinking of Bastet.
Funny this gets posted today because just yesterday I was playing Bejeweled (iPhone version) and thinking about the fact that if it were purely random, you'd be hitting 'No Moves Left' all the time and it would be crazy frustrating.
It seems to me that the game works this way: When you destroy some jewels, the game will give you either helpful, harmful or random jewels.
As you go up levels the balance changes - between level 1 and 3 you get mostly helpful jewels and sometimes random ones. At level 4 it starts throwing in harmful combinations. By the time you're at level 10 or so you mostly get random or harmful combination with the occasional helpful.
The helpful combinations make you feel lucky, or at least I think that's the idea.
I recall that the original Doom did random amounts of damage (since the designers were also roleplayers). This was most evident for the berserk pack, where you might do the same damage as an ordinary punch, but occasionally your fist would cause demons to explode. Also the shotgun would usually kill an imp in one shot, but not always. I loved this style of randomness, as it makes the game a little different each time, and not completely deterministic.
Meanwhile, I like the idea of adding a random direction to a shot fired. It means that a pixel perfect shooter doesn't always get his mark, but on average he'll still be more accurate than a poor shooter. I don't think I've ever heard anyone complain that their machine gun has spread, so unless it becomes too random, why worry if it affects the rest of the weapons? In real life there are plenty of factors that make guns not shoot the exact same spot every time.
Finally, (being someone who enjoys tabletop roleplaying, and also a researcher who mainly deals with stochastic simulation), randomness is a great way to allow people to play games without substituting the character's abilities for the player's. If your character is supposed to be good at shooting, and you point him at an enemy, then he'll hit more often if he's good. If you give your mook a gun, don't expect him to shoot accurately just because you can move the mouse to the right spot, because your character isn't very good at it. Conversely, he'll sometimes make a shot which is very difficult, but less often than the trained sniper (the same argument applies to other activities than shooting guns).
Dedicated to randomness. Maybe.
Yep, players suffer from Hero Syndrome, or PC-itis. If they are outnumbered the enemy 3 to 1, they will of course always win because they have superior numbers and any enemy victory is clearly impossible with those odds. But if they are outnumbered 1 to 3, they still expect to win because, hey, they're the hero.
Put another way, all victories by me are due to my superior skill and tactics. All wins by the opponent were due to luck and cheating.
This also applies to casino betting, and thus Vegas thrives.
A.
Back in the late '80s we played Gemstone 2 by Simutronics on the GEnie service; it was a great game but it had been ported to run on the GEIS (General Electric Information Service) mainframe servers and there was an "issue" with the random number generator.
In melee combat, the system rolled a D20. If it rolled a 20 it would roll again, theoretically up to three times for a max hit of 60.
I played for several years, as did a number of friends, and we kept many logs. We analyzed the combat some time later just for the fun of it. During all that time only two >20 results were ever seen; '37' being common, and one time (in our experience) a '22', which happened just before a crash so we think it was spurious.
The only number that would come up after that initial '20' was a '17'. They never fixed the bug at GEIS or in the game (which became Gemstone 3, and later Gemstone 4, which presumably had better RNGs!).
I hope that not-so-randomness is not an issue any more; it was laughable then though we always wanted to see a bigger number 'just once!'.
If a bullet fired horizontally from shoulder height takes a half of a second to fall, that's still 30 times as many calculations.
If someone's shooting a slower weapon such as propelled grenades, your simulation has to calculate the path of those 60 times a second anyway. So why not just make bullets a subclass of the same projectile class used for RPGs, hand grenades, etc.?
Random? That isn't random. It comes right after you block off the slot you were saving for it.
You mean like in this game?
That link doesn't implement Tetris properly (even ignoring the dimensions issue). When I was playing wisely, it gave me 27 Z pieces in a row (since Z pieces can't fit into each other and form lines, this is a guaranteed game-over); the only way out is to place one in a worse place and form a hole. As a link below points out, modern Tetris implementations limit the number of repeat blocks and ensure a good distribution within the random selection; you're guaranteed a straight piece at least once every thirteen pieces and no more than two of the same piece in a row.
Also, my Firefox on Linux fails to show my score; I had to count lines. I keep getting better, but I'm losing interest now that I've hit 17 lines (and I'm forced to place pieces sub-optimally to get out of loops of the same piece).
I think a better choice is the more famous one, as mentioned a few years ago here on Slashdot, Bastard Tetris Hates You. Downside: You can't play it online.
Yes, I played waaaaay too much Tetris back in the day.
Use my userscript to add story images to Slashdot. There's no going back.
Seriously, no one has mentioned Mario Kart yet? Where the random effects of items only got worse in the Wii version? (Increase to 12 racers, more powerful items like Bullet Bill, red shell no longer dodgable with sparks)
I must admit however, that the increased chaos does lead to a greater depth of gameplay and more experience players are better able to compensate for it. Still doesn't help when you're about to cross the finish line and get hit with a blue shell, red shell one-two punch and wind up finishing in 5th.
Relevant XKCD
Note to self - Bastard Tetris (bastet) has the same fallacy and will happily give you several of the same piece in a row, but it's a little smarter than Stacked Odds and less vulnerable to infinite loops.
Use my userscript to add story images to Slashdot. There's no going back.
I stopped reading on the second page of that Gamasutra article. It seemed to me that the distinction between evident and concealed luck transparency was simply a matter of perception.
Every one of the five examples given were clearly evident to me. There's not much point in reading an article when begins its explanation subjectively, presented objectively.
What about the NES, Gameboy and early PC versions, that were the craze and didn't have those features?
Those versions are no longer in print, and it appears The Tetris Company might not even let them go back into print. For instance, instead of putting Tetris for NES or N64 on Virtual Console, Nintendo had Hudson develop a new WiiWare game based on the Tetris Guideline. (In Russia, the party finds YOU!) And Tetris Worlds for Game Boy Advance has the new Guideline randomizer.
that's why the tank in UT2004 fires so poorly. Missing 3 out of 4 shots at medium range in the tank in real life would get the manufacturer killed by our own military after the mutiny. The conditions, replicated thousands of times, tank sitting still in the open and taking no fire. It took at minimum four shots just to get one on target hit. I quit shooter games because of that crap. I got sick of smashing trackballs and mice from the frustration. The guy who dreamed it up deserves a very bad end.
There is an old game called Trespasser, which somewhat pioneered realistic physics in a shooter, and you did have to use commands to manipulate your wrist and forearm to line up the sights on the guns. Needless to say, players are far better off just assuming their character is not an Olympic medalist quality marksman, and accepting some imperfection for the sake of realism.
For something like the Wii though, if you manually have to aim the accessory, and it shakes a bit between shots, or maybe the screen shakes slightly and is slightly removed from its original position to simulate recoil, there is probably no need for any randomness in the shot placement (aside from shifting the FOV slightly in the second case).
In san andreas, I was moving things with a giant crane to a dumper truck when a plane hit the top of the crane and crashed on the dumper truck. There was no reason why this plane passed here and by destroying the dumper truck, my mission failed.