mattbell: (Default)
[personal profile] mattbell
Angry Birds is an iphone game where you use a slingshot to propel exploding birds (yes, really) at various structures containing pigs, which you're trying to kill by hitting directly or by causing the structures to collapse on them.  

The funny thing about Angry Birds is that it's actually not a great game in one sense -- while playing the game takes some skill, skill is no guarantee of success.  

The trouble is that your only control method,  the angle at which the bird is launched from the slingshot, is the input to a chaotic system with sensitive dependence on initial conditions.  If you launch the bird at an angle of 72 degrees, it hits a block off to one side and then stops.  If you launch the bird at 73 degrees, it hits the same block off to the side a little bit harder, which starts a crazy chain reaction that causes much of the structure to collapse.  If you launch the bird at 74 degrees, it hits the block even harder, and the force whacks another block way out of place, but it does so in such a way that the structure remains standing, making it even harder to hit the pigs.  Basically, it's the butterfly effect, the favorite metaphor of chaos theorists to describe phenomena like the weather where tiny changes in initial conditions make a big difference.  

Since you can't precisely aim the birds, it's difficult to control the angle -- you just know that you sometimes get good results from hitting a particular area.  

If you graphed the score you got from the full range of launch angles, it probably would have fairly chaotic properties in high-scoring regions.  It would probably look like a one-dimensional version of this, which is a map of which magnet a pendulum ends up attracted to when started from different angles:

See those messy regions?  Those are areas where your predictive power is diminished, where the score isn't up to you but instead up to something effectively outside your control.  Psychologists have a name for this -- partial reinforcement.  This basically means you only sometimes get rewarded for a behavior.  The ironic thing is that this is more effective at entraining that behavior than full reinforcement.  

There are other sports, such as bowling, that also exhibit these chaotic characteristics, but good bowlers have found a stable region in the chaos that they can hit consistently.  

Angry Birds, once you reach a basic level of prowess, is less like a game of skill and much more like something else entirely -- a slot machine.  

Date: 2011-02-02 08:37 am (UTC)
From: [identity profile]
I could agree with this if I didn't observe improvement, but I do observe improvement. It's true that some of the maps take longer to clear and getting just the right outcome is harder on some than others, but in general I find that I can control the birds' flights predictably enough to get relatively consistent results. And moreso over time. There is a lot of learning... figuring out what the optimal use is of each bird, and then getting each of them to just the place you mean to put them in the right order. It's not easy, but it's predictable enough that learning and practice allow you to get through faster over time.

At first it felt like scientific experiment without the benefit of informed hypothesis. Just try stuff, see what happens, then create and (imperfectly) test theories after amassing a lot of results. But over time I started seeing that my intuition about structural weaknesses and capabilities and ranges of different birds started leading to better outcomes sooner on unfamiliar maps. I could see where the most exploitable weaknesses are likely to be, then probe those points first. Of course, as this happens, the maps get more difficult, so the time spent working each of them can take just the same time, or even longer. But on the other hand, the feeling of satisfaction goes up, too, because you conquered something that previously would have seemed impossible.

I agree that your ability to control the birds is somewhat limited, and that limitation leads to a lot of restarts not because you didn't know what you wanted to do, but because you weren't able to perfectly execute. And in that respect they've certainly increased the number of swipes of my finger they can get me to try within their game (thus tying me to it for longer periods). But to me that's just another way of saying they've designed a good game. Make it too easy, and I get bored. Too difficult, and I give up in frustration. Give me just enough difficulty, and I feel that I can get it if you just let me try it again. Retries are quick and unlimited, so it doesn't bother me to quickly do it again when I don't get quite the pull I meant to.

With a slot machine, you'd never see improvement. For every time you got it done quicker, you'd have just as many times getting it done slower. That's not what I think I see with Angry Birds. I get better as the challenge becomes greater.

Has that not been your experience? Does it really seem that difficult to control to you?

It seems pretty similar to me to shooting a basketball, or aiming a reticle in a first person shooter. No one can do it perfectly every time. Not even the most skilled. But some are better than others, and everyone can get better with practice.

Date: 2011-02-02 05:47 pm (UTC)
From: [identity profile]
Thanks for the thoughtful reply. It's possible I don't give the game enough credit - I only played through the levels in the free version.

I could see some potential longer term improvements in terms of learning the peculiarities of the in-game physics engine (eg the yellow bird will go through N wooden walls before stopping, or an explosion from a black bird will destroy pigs within X radius).

I could also see the game as an effort to find islands of more stable, predictable results in an ocean of chaos, and there's likely some skill in realizing what will have predictable vs chaotic effects.

However, my experience was that, beyond the first few levels, the best solutions required sensitive dependence on initial launch angle, too sensitive for me to eyeball consistently. I was reduced to counting pixels between the bird and the base of the slingshot to get the most accurate results. It's interesting you see that as part of the challenge. It just frustrated me. Maybe that's because I was thinking of it as more of a puzzle game. It is already sort of a puzzle game, with other elements thrown in.

Date: 2011-02-02 06:24 pm (UTC)
From: [identity profile]
Just as there are areas of the pendulum map with reasonably wide stretches of a single color, there are angles in Angry Birds that are rewarding even with a reasonable margin of error. Further, as the player learns the chaotic mechanism of success and failure, he is more willing to try the same shot repeatedly to determine whether there is a high-reward angle within his margin.

With practice, the player's margin of error shrinks. It won't shrink so that he can avoid the chaos altogether, but I discovered in the later levels of the holiday version (which was much more demanding) that I could aim a bird with much greater precision than was ever called for in the original version.

Finally, with the time-sensitive special abilities of many of the birds, the graph *is* two-dimensional. The boomerang bird in particular has an *extremely* chaotic special ability, and a difference of a few milliseconds can change the outcome of the shot from a catastrophic chain-reaction to a clean miss.

Date: 2011-02-02 09:10 am (UTC)
From: [identity profile]
Apologies if the feedback isn't interesting or wanted. This topic is very interesting to me I guess because I've spent a fair amount of time playing Angry Birds lately and find it enjoyable (as does my wife)... so, consider my bias fully disclosed. ;-)

Thinking about this more, I guess my essential response is this: to say Angry Birds is more like a slot machine than a game of skill is hard for me to accept. Most games involve skills that are not perfectly predictable even for the best players.

No bowler bowls 300's every game, for example. There is certainly a gradient for typical variations among sports and games. E.g., Steve Nash hits 90% of his free throws, and Shaq about 50%. Good baseball players get hits only about 30% of the time, the best near 40%.

The best Asteroids players could basically play until they fell asleep. And the best Rubik's Cube solvers can always solve in under a minute.

At the other end of the gradient, the slot machine, I guess there is theoretically no difference between the "best" at pulling the arm and the "worst". Everyone's chances are completely random. Are you really saying the ability to aim the birds is more like that than more predictable (yet still imperfectly repeatable) skills like the others I mentioned?

Date: 2011-02-02 05:10 pm (UTC)
From: [identity profile]
At the other end of the gradient, the slot machine, I guess there is theoretically no difference between the "best" at pulling the arm and the "worst". Everyone's chances are completely random.

Actually if you know what you're doing, you can do a bit better than random.
1) With really old slot machines it was possible to know when the next pull would have a much higher chance of winning based on historical data. The timing of when the spinners would stop was not random and actually fairly predictable.(These are no longer in use for obvious reasons)
2) But it's more in picking the machine than how to play it. If you see a machine right near the front door that seems to be paying out more than the others, play it for a couple minutes and leave. The casinos randomly turn the payout on machine near the doors to more than 100% (ie they pay more than the take in). They do this to draw people into the casino. Typically one machine never does this for more than 5 minutes at a time.

Date: 2011-02-02 05:50 pm (UTC)
From: [identity profile]
Is this known for sure? This sounds like the kind of story casinos would come up with to get people who dismiss slot machines as money-losing propositions to play.

Date: 2011-02-02 06:19 pm (UTC)
From: [identity profile]
Some mechanical slot machines had predictable behavior. In some cases, it was designed into the machine, like an increasing chance of payout over time in order to satisfy gaming regulations. In other cases, gamblers would exploit real or perceived flaws in the design. Casinos were very careful to avoid addressing the real flaws, because the perceived flaws were very profitable, if not effective for the gambler. :)

Modern digital slot machines, even those with moving parts, are carefully and specifically programmed to be just random enough to replace "partial" reinforcement with true "intermittent" reinforcement, which can consistently satisfy the regulations on payout without giving any gamblers an advantage over others (or the house). This doesn't remove the perceived flaws. There are tales of slot players who swear that pressing the lower left corner of the "spin" button gives them a better chance of winning.

Date: 2011-02-02 06:45 pm (UTC)
From: [identity profile]
I heard about the high payout thing when at the USITT (US Institute of Theater Technology) conference in Vegas back in 1995. Give we were all back stage types, we got a lot of tours of the inner workings of the theaters and the casinos they were in. Note that the Wikipedia page on slot machines does talk about the eproms used to set the payout rate can be switch between several different rates.

One of the tidbits I pick up that week was the 'payout' for Treasure Island Casino's free sidewalk pirate show. At that point the show was scheduled to run 5 times a night every night. Unfortunately for safety reasons, they can not run the show if there is too much wind (and Vegas gets lots of wind). So the Casino has pretty good data on the difference in the casino when the show run or doesn't run. At that point the show's unit cost (the cost to run the show one more time) was about $5000. The extra profit in the casino in the 5 minutes after the show was over was $8000. Note that it took me about 15 minutes to even make it into the casino after the show was over.

Date: 2011-02-02 05:53 pm (UTC)
From: [identity profile]
It's certainly a continuum. Angry Birds isn't a slot machine, and there is some skill involved, but I felt like it had enough of the elements of a slot machine that I got turned off.

Date: 2011-02-02 11:33 am (UTC)
From: [identity profile]
thanks for the connection :)

Date: 2011-02-02 02:05 pm (UTC)
From: [identity profile]
It's called intermittent reinforcement. :)

Date: 2011-02-02 05:48 pm (UTC)
From: [identity profile]
Both names are in common use, according to a couple of online sources I checked.

Date: 2011-02-02 03:54 pm (UTC)
From: [identity profile]
I just couldn't get into Angry Birds. Your last post on it pretty much perfectly summed up the reason why. It got deleted from my iPad and iPhone.

Now Cut the Rope. Brilliantly done and makes great social use of the touch screen medium. And feeding an adorable monster candy is so rewarding!! :)

Date: 2011-02-02 05:49 pm (UTC)
From: [identity profile]
I played Cut That Rope in the Vegas Apple Store while waiting for a show to start.

Yeah, that game is really well designed. The puzzles are very good and in a problem domain that we're not used to. And that monster might just be the cutest game character I've seen in quite a while.

Date: 2011-02-02 06:17 pm (UTC)
From: [identity profile]
Wait? There is a game based on Animal Farm?

(no, I really don't play many games, so have been oblivious to this whole "Angry Bird" craze)

Date: 2011-02-02 06:27 pm (UTC)
From: [identity profile]
Four legs good, two legs baa-aa-aad!

Date: 2011-02-02 06:55 pm (UTC)
From: [identity profile]
Right? So, I'm not the only one that thought that?

Date: 2011-02-02 07:18 pm (UTC)
From: [identity profile]
Angry Birds could be considered a sequel to Animal Farm, where the birds realize the pigs are stealing everything they worked so hard to make, and topple the regime... only to see it replaced by an endless series of porcine kleptocracies. ;)

Date: 2011-02-02 07:55 pm (UTC)
From: [identity profile]
I like it. Still won't play it, but I like it :)

Date: 2011-02-02 08:31 pm (UTC)
From: [identity profile]
Right. The pigs "stole" the eggs because, as the most intelligent beings, they were the proper ones to decide how to allocate the eggs for the benefit of the Animal Collective.


mattbell: (Default)

February 2011

   123 45
67 89101112

Most Popular Tags

Style Credit

Expand Cut Tags

No cut tags
Page generated Sep. 25th, 2017 01:26 pm
Powered by Dreamwidth Studios