News for hackers. Stuff that matters.
Re:How do you stop a randomized game from randomly being bor... by pjc50 on Monday May 23, 2016 @05:38AM (#11752774)
XCOM2 has a slightly controversial "randomness compensator": repeatedly missing increases your chance to hit, and being hit repeatedly by the aliens lowers their hit chance.

Various games (L4D is the first prominent example, there may be others before it) include "AI directors" that try to keep the player involved through balancing exiting events and periods of tension. This makes it a lot harder to have a "boring" game, or one where you get "unfairly" hammered by the randomness.

In fiction I'd trace this to Asimov's Foundation series, which is premised on "psychohistory" being a sort of fully deterministic fusion of economics and psychology enabling the development of future society to be predicted. Until a few books into the series where the "mule" is introduced, who is an individual powerful and unpredictable enough to throw off the determinism of historical inevitability.

Re:How do you stop a randomized game from randomly being bor... by Agentlien on Monday May 23, 2016 @09:32AM (#11753703)
This reminds me of the AI rubber banding implemented in many racing games, including Need For Speed.

Full disclosure: I work as a Software Engineer for Ghost Games and was involved in the development of Need For Speed (2015). I am not an official spokesperson for Ghost or EA and I am not authorized to speak on their behalf.

In Need For Speed, AI racers will slow down when you're behind and speed up when you're ahead. The idea was to always make the player feel like they were having a fun challenge, regardless of skill level.

In the end, the AI rubber banding was such a strong effect that, when the game first came out, it was one of the main complaints. We ended up fixing it in the very first patch.

IGN reporting on it: http://www.ign.com/articles/2015/11/23/need-for-speeds-most-... Our announcement which first mentioned it: http://www.needforspeed.com/en_AU/news/future-plans

Re:How do you stop a randomized game from randomly being bor... by stepvhen on Monday May 23, 2016 @11:03AM (#11754308)
This was the worst part of F-Zero GX, to the point where the best strategy for some of the story missions is to stay just behind the lead until the end and boost at the finish. I had a truly hellish time playing through that story mode because of rubber band AI. I'm surprised it was still a standard game mechanic in 2015.
Re:How do you stop a randomized game from randomly being bor... by hvm on Monday May 23, 2016 @10:35AM (#11754097)
I've seen this in an older version of NFS on PC. I can't remember which one it was (I think Underground?). The way they always caught up with me when they were behind was insane.

I actually tried a bit to look back and see it happen beacuse, I was wondering, do they just teleport or actually drive at like 800km/h.

Re:How do you stop a randomized game from randomly being bor... by rasz_pl on Monday May 23, 2016 @11:28AM (#11754465)
this feature is literary called "Catch Up" https://en.wikipedia.org/wiki/Game_mechanics#Catch-up and has been a thing in racing games since before 2000. Diddy Kong Racing (1997) had it, so did motorhead in 1998, then EA got familiar with it and we got Hot Pursuit 2/Most wanted and burnout 3.
Re:How do you stop a randomized game from randomly being bor... by softawre on Monday May 23, 2016 @12:15PM (#11754794)
The original Mario Kart did this with the strength of Powerups (and maybe speed as well?)
Re:How do you stop a randomized game from randomly being bor... by btgeekboy on Monday May 23, 2016 @01:04PM (#11755140)
This sounds very similar to what I've noticed from the Mario Kart series. It's a hard game to consistently win, as the AI is always compensating. It seems to do this with powerups for multi-player as well - it wants you to get ahead, but when you start to do well, it gets a lot harder to stay ahead.
Re:How do you stop a randomized game from randomly being bor... by Macuyiko on Monday May 23, 2016 @08:45AM (#11753475)
> XCOM2 has a slightly controversial "randomness compensator": repeatedly missing increases your chance to hit, and being hit repeatedly by the aliens lowers their hit chance.

Many games do, actually, and it's an interesting topic in of itself. In case someone wants some further reading:

- http://www.redblobgames.com/articles/probability/damage-roll...

- http://sinisterdesign.net/unpredictability-and-control-in-tu...

- http://www.gamasutra.com/blogs/RogerHgensen/20101123/88511/C...

- http://techland.time.com/2012/10/12/why-xcoms-random-results...

- http://forum.zdoom.org/viewtopic.php?t=50643

Another funny anecdote is that Wesnoth had to include a FAQ entry about the game being unfair because it relies too much on "luck": http://wiki.wesnoth.org/FAQ#There.27s_too_much_luck_in_this_...

Re:How do you stop a randomized game from randomly being bor... by munchbunny on Monday May 23, 2016 @08:32PM (#11758070)
Dota 2 does this as well for chance based skills. For certain skills, each subsequent failure will increase the chance that the next test will pass. When the test passes, the bias resets. Patch notes will sometimes talk about switching a skill from "true random" to "pseudo-random" to indicate this change. It's generally seen as a light to moderate boost because it ups the reliability of the skill noticeably.

XCOM 2 had a few interesting implementation details. You can actually see it all in the source code provided in their modding tools, which is a really awesome thing IMO, to be able to see a huge chunk of XCOM 2's actual code.

On the two lower difficulties (equivalent to "easy" and "normal") there is a relatively minor, invisible bonus to your next shot's chance to hit if you have a chain of misses. On the "hard" difficulty, this correction factor actually increases to compensate for several other changes that make the game much less forgiving of tactical errors. Then when you get to the "legendary" difficulty, the bonus goes away entirely and you're fighting a completely fair RNG along with all of its streakiness. I think this puts an interesting perspective on how "RNG fairness" works.

Re:How do you stop a randomized game from randomly being bor... by sevensor on Monday May 23, 2016 @10:59AM (#11754260)
It's funny, when I see XCOM2, I think Terror From the Deep, into which I sunk so many hundreds of hours during my misspent youth. Although various guides online indicate that the game is actually winnable, I never managed to beat it. And it certainly didn't try to compensate you for, say, not detecting any alien subs until July and then having your crew slaughtered by giant lobsters.
Re:How do you stop a randomized game from randomly being bor... by pjc50 on Monday May 23, 2016 @11:35AM (#11754524)
I can't remember beating it either! Although I did beat the original original XCOM.

The story behind TFTD's difficultly (you may have heard this one) is that people complained that XCOM was too easy, even on the hardest difficulty. But this was due to a bug: save/load did not persist the difficulty level properly, so most people played one or two missions on ultra-hard and the rest on easy, without ever realising. So when the developers turned up the difficulty and fixed the bug, suddenly the sequel was incredibly hard.

Re:How do you stop a randomized game from randomly being bor... by ralfd on Monday May 23, 2016 @11:38AM (#11754540)
Saved
Re:How do you stop a randomized game from randomly being bor... by evincarofautumn on Monday May 23, 2016 @05:53AM (#11752821)
> repeatedly missing increases your chance to hit

That’s amazing—a logical fallacy as a feature!

Re:How do you stop a randomized game from randomly being bor... by pjc50 on Monday May 23, 2016 @07:31AM (#11753146)
So here's the thing. We (the playing public) expect different types and standards of "fairness" from games of skill and games of chance. We expect that games of skill are beatable by playing better and that they should generally be beatable. People get very upset if they've sunk hours into something to discover that it's unwinnable, especially if it was unwinnable all along as soon as the random seed was chosen.

In games of chance on the other hand, people understand that you can work at making better bets but you can't make yourself "luckier" by force of will. So it's easier for people to blame the RNG gods than berate one another for lack of skill.

Edit: remember in game design, especially human vs. AI, it's not about beating the player but about maximising their enjoyment of the process. You have to make the defeats fun as well.

Re:How do you stop a randomized game from randomly being bor... by eterm on Monday May 23, 2016 @06:59AM (#11753029)
It's interesting, because it means that on average you'll do substantially better than your expected hit rate. For poisson processes a lot of the expectation of the time taken "comes from" long tails.

If you're rolling a fair dice, if you cap at 12 (i.e. the 12th roll will land a 6 no matter what) then your expected time to roll a 6 goes down from 6 to 5.4.

So by applying a "correction" on strings of misses without applying a similar correction to a string of hits it skews the mean down which actually changes what your average hit-rate is. It would be interesting however if the tested hit-rate was adjusted lower to account for this effect, that would make it "fair" again by the hit-rate stat reporting the average hit-rate, but would make the randomness generate "appear more random" to humans.

Re:How do you stop a randomized game from randomly being bor... by Kurtz79 on Monday May 23, 2016 @06:46AM (#11752984)
How is this a logical fallacy ?

Technically you would see the effect of your shot (i.e. too much to the left) and adjust your aim (more to the right).

Re:How do you stop a randomized game from randomly being bor... by MereInterest on Monday May 23, 2016 @07:17AM (#11753094)
He is referring to the Gambler's Fallacy (https://en.wikipedia.org/wiki/Gambler%27s_fallacy). That is, the belief that in a series of uncorrelated events, after a long string of losses, you are more likely to have a win.
Re:How do you stop a randomized game from randomly being bor... by Kurtz79 on Monday May 23, 2016 @10:05AM (#11753900)
Thanks, I knew about the Gambler's Fallacy, but I really didn't make the connection (also since they are not really uncorrelated... :) )
Re:How do you stop a randomized game from randomly being bor... by pjc50 on Monday May 23, 2016 @07:34AM (#11753165)
Yes, although in XCOM it's applied across all shooters and all targets, so a miss by trooper A against alien B will improve the chance of trooper C to hit alien D.
Re:How do you stop a randomized game from randomly being bor... by tremon on Monday May 23, 2016 @07:24AM (#11753116)
https://en.wikipedia.org/wiki/Gambler%27s_fallacy

the mistaken belief that, if something happens more frequently than normal during some period, it will happen less frequently in the future, or that, if something happens less frequently than normal during some period, it will happen more frequently in the future

Re:How do you stop a randomized game from randomly being bor... by vintermann on Monday May 23, 2016 @06:42AM (#11752969)
It's not so unusual. Sid Meier has talked about a feature like this, and I think even Civilization (the first) had it.
Re:How do you stop a randomized game from randomly being bor... by minikomi on Monday May 23, 2016 @07:13AM (#11753079)
Mario Kart certainly has a "the more you're losing, the faster you go" .. or at least the inverse implemented.

Along with "the further behind you are, the more likely to get a blue shell".

Re:How do you stop a randomized game from randomly being bor... by vintermann on Tuesday May 24, 2016 @02:21AM (#11759329)
That's a catch-up mechanism, that's different than merely smoothing out outlier streaks.

Though I think that the reason this feels necessary, isn't that people are irrational and believe in streaks - it's that they used a poor random model in the first place. Maybe a spearman should simply never beat a battleship.

Re:How do you stop a randomized game from randomly being bor... by bluesign on Monday May 23, 2016 @08:12AM (#11753320)
also many racing games etc, have speed bonus if you are behind , ex: wipeout allows speed up, if you are tailing an opponent.
Re:How do you stop a randomized game from randomly being bor... by ymse on Monday May 23, 2016 @08:38AM (#11753438)
> XCOM2 has a slightly controversial "randomness compensator": repeatedly missing increases your chance to hit, and being hit repeatedly by the aliens lowers their hit chance.

Do you have a source for this? I'm not doubting you, but I've played copious amounts of XCOM2 (until it broke on NixOS..) and didn't notice. But now that I think about it, the hit chances are less predictable than in the prior games.

Re:How do you stop a randomized game from randomly being bor... by pjc50 on Monday May 23, 2016 @09:50AM (#11753803)
Re:How do you stop a randomized game from randomly being bor... by stefs on Monday May 23, 2016 @10:28AM (#11754039)
wow is another good example of something like this for the "collect 10 wolf pelts" types of quests. in the beginning you had a drop chance of, say, 30% and a dice roll on every kill. so after ~30 wolves you should have your ten pelts. but if you were really unlucky you could 100 wolves and still not have enough.

--- note: seems i remembered this wrong, and i can't find the source for this anymore.

iirc, they fixed this by creating a "random queue", i.e. a fixed size queue (say, twice the average) with the amount of needed drops randomly distributed. this way the players bad luck streak would have an upper bound.

--- note: what they actually did was increase the drop chance progressively, but it might be that this was only for single item drops?

Linear Regression of Progressive Percentage in World of Warcraft

http://math.arizona.edu/~ercolani/pyzdek.pdf

Re:How do you stop a randomized game from randomly being bor... by rasz_pl on Monday May 23, 2016 @11:32AM (#11754501)
>XCOM2 has a slightly controversial "randomness compensator"

This is because human brain is Terrible at statistics. Many games curate random events into Normal distribution, otherwise you get complaints about all those missed 99% accuracy shots.