Chat Transcript: Analysis Paralysis

The topic for the December 8th, 2003, Monday Night Chat was Analysis Paralysis and working to reduce its likelihood in your designs.

I edited the chat to clump conversations somewhat, and removed the pleasantries (greetings and such).

Please feel free to add additional thoughts!

This is the first hour or so of the chat. I'll add the second portion in the near future.

FastLearner: I define Analysis Paralysis as a player finding that he has so many possible options and/or consequences to analyze before making a move that he becomes nearly "paralyzed," grinding the game to a stop.

FastLearner: Is that what you guys think of when you hear the term?

Scurra: Sounds about right to me

tjgames: That's what I thought it meant

FastLearner: Some players are particularly susceptible to AP, and there's not a lot we can do about the differences in players. We can, though, work to create games that aren't terribly prone to it.

Scurra: I find Chess to be one of the most interesting examples of this.

Scurra: When you start, there are loads of options

FastLearner: So true. Chess is nearly defined by AP.

FastLearner: Hence chess clocks.

Scurra: As you get better, the options decrease, and then suddenly you get good enough that the options open up again

FastLearner: I see two core causes: too many choices to make, and too many consequences to consider. Some AP-prone games suffer from both, but more commonly it's one or the other.

FastLearner: Chess has both, I'd say.

Scurra: Which is why I think AP games suffer from AP

FastLearner: Precisely.

Scurra: (action point vs an.para)

DarkDream: I think you identified well, FL the causes of AP

FastLearner: Which brings me to:

FastLearner: Let's look at the first one, too many choices to make. Are there any games or mechanics that come to mind where this is a problem?

Scurra: One of the worst of these is "Citadels"

Scurra: Although that's a little unfair

DarkDream: Can you elaborate?

DarkDream: Where do you think there is too many choices?

Scurra: Well, one of the central ideas of the game is the "psychology" aspect

DarkDream: agreed

Scurra: You try to guess what others think you think they think you are going to take, which leads to AP

Scurra:

FastLearner: I think Citadels particularly suffers from the "too many consequences" issue.

Scurra: Oh, that too - I was going to mention it again later

DarkDream: FL, where do you see too many consequences?

FastLearner: I'll hit the consequences of Citadels after we talk about choices problems a bit more.

Scurra: But the choices are quite complex, beyond the consequences of those choices

FastLearner: Action Point games in general suffer from AP much of the time.

Scurra: It's one of the reasons I hate Tikal (et al)

FastLearner: It's interesting, though.

Scurra: You have (almost) perfect information at certain points

DarkDream: You think players spend too much time figuring whether or not to purse a certain avenue to get more action points?

Scurra: and that's one of the great AP generators

FastLearner: With those three games -- Tikal, Mexica, and Java -- Mexica doesn't have nearly the same problem, and I think it's a design issue.

Darkehorse: Interesting..

Scurra: Well, I couldn't bring myself to play Mexica - but perhaps I should

FastLearner: In Mexica you still have a wide range of choices and plenty of Action points, but the things you'd likely want to do are generally much more limited each turn.

FastLearner: In Java, on the other hand, nearly all of the actions could well be good.

Darkehorse: Torres doesn't have that problem as well?

FastLearner: Aye, Torres has some of it, too.

Scurra: I'm not so sure about Torres. It certainly has AP issues but the limitations are often clearer

FastLearner: Perfect information, as you mentioned Scurra, seems to open up the "consequences" side of things a lot, too.

Darkehorse: It's a paradox really.. We want multiple paths to victory which means we need more choices which in turn leads to AP

FastLearner: It seems like the best way to allow for mp2v while not having much AP is to help players get on "tracks".

DarkDream: It sounds like a balance of having multiple paths but not too many

Scurra: And the more "perfect" the game information is, the more AP is generated too

Darkehorse: tracks? Like Politics

Scurra: I really really don't like to do that

Scurra: I find that I'd rather let the players figure their own "tracks" out

FastLearner: That is, if every single turn you have multiple paths to victory then AP becomes rampant, but each path requires a bit of "investment" then they're less likely to wander.

FastLearner: I agree that I don't want to shepherd them too much, but I do want to help create... channels, let's say, where for two or three turns you can focus on making a particular thing work.

Scurra: I can see what you mean, but it can backfire sometimes

FastLearner: I certainly agree with that. In fact if you're not careful you'll end up creating a game where there's only one real path to victory.

DarkDream: So position and resources in a game for a particular player can limit the multiple paths of victory and hence AP

FastLearner: That's certainly what I mean, yeah. Great way to put it, DD.

FastLearner: Are there games, Scurra, where you've seen it as a problem? That is, railroading?

DarkDream: Would Settler's be a decent example here

DarkDream: in terms of somewhat limiting the paths of victory based on position and resources?

FastLearner: I think so. I like that if you decide you're going to go for some points via, say, longest road, then you can't realistically also go for largest army, too, at least most of the time.

FastLearner: Aye, too: if you're not producing brick then you're not going to be the longest road guy, probably.

DarkDream: So in this case you are not bombarded every turn with multiple paths only certain turns

FastLearner: That's what I think is a good solution.

Darkehorse: Imperfect information can be just as paralyzing for players who like to calculate odds

FastLearner: So anyway, Darkehorse you mentioned that imperfect info can lead to AP: any examples?

Darkehorse: Well... Let's see

FastLearner: While you ponder, how about the second one everyone, too many consequences to consider. Are there any games or mechanics that come to mind where this is a problem?

FastLearner: Chess is obviously a great example of a game with an incredible number of consequences.

Scurra: Carcassonne

Scurra: when you play it at a "high" level, anyway...

FastLearner: Ah, Carc. Good point. Would you mind elaborating?

sedjtroll: what about this fix for AP... introduce a turn timer

sedjtroll: like a 20 sec hourglass

Darkehorse: Ok can't think of a good example... Usually this happens when a player knows the game backwards and forwards and he knows exactly what can come up (for instance in a card draw). The player could spend a great deal of time weighing every possible conseq.

DarkDream: You have to analyze various moves five moves deep and evaluate the consequences

sedjtroll: happens in Magic...

Scurra: In Carc, when everyone knows all the pieces, the placement options multiply hugely

Scurra: since you start calculating likelihoods and so on

FastLearner: Excellent point.

Scurra: but, of course, played socially, Carc is quite fast and fun

DarkDream: I have only played it socially myself

Scurra: You should try playing Carc "seriously" - it's an education!

Scurra: Good players won't give an inch away, and will complete wreck you...

FastLearner: Does randomizing decrease the likelihood of AP?

Darkehorse: Again, it depends on how well the players know the game... I.E. they know what to expect

Scurra: Randomizing what?

FastLearner: Randomzing anything.

FastLearner: Carcassonne is a good example where the random draws don't solve the problem for players who have memorized the tileset.

Scurra: A good example of removing AP by adding a slight random factor is "Alhambra" if you've seen my review in here

Scurra: Stimmt So had some bad AP because there wasn't anything else you did with the shares

Darkehorse: But again that goes against the point, good/experienced players are SUPPOSED to know what's going to happen

Scurra: And that's why some games just don't work if you mix newbies with experienced players

FastLearner: Anything that makes it less likely that you can truly know how something will come out and so you can't "see" past a certain future point with enough certainty to make it worthwhile.

DarkDream: I think it can help eliminate AP, as with randomizing certain elements there is no way to guess what is coming next. Thus, you cannot calculate and experience AP.

sedjtroll: There was some complaint of A/P in 8/7c

sedjtroll: that may be a good example

sedjtroll: esp. if you know the cards

FastLearner: What was causing it, do you think?

sedjtroll: well, in part the "maximizing $ for ads

FastLearner: Seth: perhaps you could describe that more.

sedjtroll: with the "ad math", people would try and search for the "best" place to put an ad. Also, deciding on the best play based on what cards you might draw... that was also a candidate

sedjtroll: for A/P

sedjtroll: and A/P might be a little bit of a problem in 8/7c

Darkehorse: Seriously though, if you want to have a strategy game, AP is pretty much going to be a fact of life. If you randomize it too much then you risk ruining your game

DarkDream: Agreed, it is really a matter of degree. Randomizing should be done parsimoniously in key places where AP seems to exist.

FastLearner: Well... maybe. Union Pacific is an example of a game imo where there's enough control to make it a strategic game but enough randomness and hidden info to have very little AP.

tjgames: I would agree with DH Too random less strategy

Darkehorse: Balance is the key, as with anything..

Scurra: Actually, I find that having a "methodical" player in my current playtest group helps me spot AP issues early

Scurra: Since he makes the game last a very long time if there is a problem!

FastLearner: That sounds like a very useful playtester!

Scurra: He's a very good player, but sometimes we get a little exasperated!

Darkehorse: LOL.. Doesn't every group have that 'slow' player? In our group, his name is Steve. If you take too long on your turn, you are said to be 'pulling a Steve'

FastLearner: Let's look at the solution side a bit. What can be done to ensure that the number of choices is reasonable?

Scurra: Well the rule is 7 to 9 options at any time, isn't it?

DarkDream: Reduce overall complexity of the game, and try to keep the choices a maximum of 4-5.

Scurra: But they don't have to be the same options each time of course

Darkehorse: Another one of your 'rules'... Where do you derive these rules from?

DarkDream: I read that psychologically we have difficulty keeping more than 7 pieces of information at one time in our head.

tjgames: One Solution is to give players the info the need before their turn

Darkehorse: Excellent point!

FastLearner: Aye, excellent point.

sedjtroll: games like that and Carc are like- you're almost not supposed to consider that stuff too much. but you can't really tell people that

sedjtroll: the only way I see to ELIMINATE it would be to put people on a clock

DarkDream: I just don't think people would put up with that

Darkehorse: Clocks tend to alienate certain personality types, and in truthfulness they change the feel of the game...

FastLearner: Agreed.

Scurra: Have you seen Magna Grecia?

Scurra: That has a nice "here are the options you will have on this turn" device

FastLearner: Yeah, I've played it.

Scurra: which lets you consider what to do

FastLearner: I agree that MG has a clever mechanism for that.

Scurra: I was really impressed with the MG solution

Scurra: (mind you, I like the game a lot too, which surprised me based on a description)

sedjtroll: with Robo Rally we have an interesting solution (talk about AP!)

FastLearner: What's that, Seth?

Scurra: It's the "Ricochet Robot" solution, I would guess

sedjtroll: any player who has finished programming their robot can start the 2-minute timer (if it hasn't already strarted)

sedjtroll: when time is up, you are done programming- period

sedjtroll: so it's not a strict timer

Scurra: Hey, I was right

FastLearner: Interesting. That would speed it up.

Scurra: Because it's pretty similar to the one we use too

sedjtroll: you always get at least as much time as the fastest guy, plus 2 minutes

sedjtroll: is that like an official version? I don't know where it came from

Scurra: Seth, no, I think "great minds think alike"

FastLearner: I've not heard of it, but I like it. Though sometimes the fastest guy might only have two cards.

Darkehorse: seth: still it will tend to alienate certain personality types.. Some people just don't think well under pressure

Darkehorse: This again, is where a computer hybrid board game would come in handy. If the computer could present you with possible actions, it would considerably speed up the process.

FastLearner: True, as long as it didn't give too much of the game away.

Scurra: Oh yeah - like a puzzle game which gave you multiple-choice options!

Scurra: I think that's a bit like what Knizia's King Arthur game does

FastLearner: Do you know anyone that's tried it, Scurra?

Scurra: I haven't spoken to any of them about it, but I know some of the playtesters. They could probably talk now!

FastLearner: I'd like to know more about King Arthur.

Scurra: I'll drop you a PM tomorrow about it.

DarkDream: I think there is no way to eliminate it, the goal is to mitigate it

FastLearner: I think TJ may have been referring to things like "draw a card at the END of your turn instead of the beginning," for example.

DarkDream: I think the important idea presented is that you can reduce AP, by having the player able to think on the opposing players turn

FastLearner: I think that's a big one, DD, agreed.

Darkehorse: Well I think players are supposed to do that anyway.

Scurra: it's not always possible though

Darkehorse: But often times its difficult because the game changes so drastically depending upon the # of players

Scurra: take Tigris & Euphrates for instance

tjgames: At least if you are thinking on other player turns to things are solved

DarkDream: That the key, it only effectively works if the gaming field does not drastically change.

Darkehorse: I can't think of any examples, but I remember playing games where it was IMPOSSIBLE to plan more than 1 turn in advance, and even then it sometimes didn't work out

FastLearner: Sometimes players really can only do so much of it, though. For example in games where you draw a card at the beginning of your turn and the card can dramatically change what you're capable of that turn, it's very difficult to think ahead too much.

FastLearner: In a bad way.

Darkehorse: exactly...

tjgames: 1) you are not bored

tjgames: 2) You have already covered some of the options,

FastLearner: In other games the board (situation) changes so much that anything you figure out before your turn is likely moot.

tjgames: That's true Fl but that just has to happen sometimes

FastLearner: True. It seems to vary by game a lot, though.

tjgames: Example FL

FastLearner: TJ: I'm thinking about it... I remember several games that were like that...

Darkehorse: In my game poker face... there are two general phases to each turn... You draw your cards at the beginning of the turn, but they don't actually affect the first phase.. So you have a good time to think about what you are going to do with your cards

DonovanLoucks: Drawing cards at the end of your turn also indicates to the other players that you're done.

Scurra: I've changed a lot of designs to "draw at end of turn" precisely because of that

DarkDream: great idea

FastLearner: The first game I played that had that mechanism really threw me for a loop, but now I like it.

DonovanLoucks: That's a mechanism I've used in several of my (unfinished) designs.

tjgames: That also a Good reason to draw at the end of turn

Scurra: Oh yes, and now I find it odd to do things the other way around

Scurra: although having been doing a bunch of Rummy-variants lately, "discard at end of turn" is a good signal too!

FastLearner: Anyone have anything else in particular about Analysis Paralysis? Any other good tips or solutions?

tjgames: One that I used is in a current game is player being a sort of timer

FastLearner: Explain, if you will.

Scurra: I have the "do nothing" option (a la PR Prospector) which is a quick get out if they genuinely can't think what to do

Scurra: the pay-back is tricky

Scurra: tho'

FastLearner: That's a good point. One game I have in design now uses that, too (you can just take 2 gold instead of one of the other actions).

FastLearner: What do you mean by "pay back," Scurra?

Scurra: You have to make the return on the "do nothing" option worth taking, but not so much better than taking a "real" action

FastLearner: I agree... not so bad that there's no way you'd consider it, but never so good that it's clearly the best choice.

DonovanLoucks: Ah. I thought of one. In "Apples to Apples", the last player to play doesn't actually get to play.

FastLearner: Great example.

Scurra: We don't like that rule in A2A - it encourages random card throws

FastLearner: We usually play the "slow baked apples" version that eliminates that rule, too, as many players just throw junk. I'm a fast thinker, though, and like the rule.

Scurra: It's not as though A2A is prone to AP though, is it?

FastLearner: It sure is for some people.

DonovanLoucks: Scurra: No, but my point is that some sort of penalty could be foisted on the last player.

FastLearner: I played a nearly-3-hour game of Frank's Zoo the other day, due to two new players with heavy duty AP

Scurra: Urgh!

DonovanLoucks: RoboRally's a good example. The last player makes no moves this turn.

DonovanLoucks: Of course, this penalty mechanism is best when you have a lot of players. It obviously doesn't help much when there are only 2.

FastLearner: Donovan: that's an optional RR rule, right?

DonovanLoucks: Matthew: Not to my knowledge, though it might be.

FastLearner: Donovan: I don't remember that as a rule, which is why I asked.

tjgames: I am working on a game where player take turns being the farmer.

tjgames: It is a real time game but can sometime drag

tjgames: The farmer job is to keep it going, by acting as a count down mech

tjgames: If he fell that game is dragging he counts down. Player can stop him simply by playing a card and then he must start the count down again.

FastLearner: Interesting. Nice idea.

Scurra: TJ, that's rather neat

DonovanLoucks: (Every time I see "AP" I think "Action Points" instead of "Analysis Paralysis"...)

FastLearner: (Donovan: Yeah, we joked about that earlier)

Scurra: BTW, talking of A2A, have you ever played the Christmas Card game?

FastLearner: Christmas card game?

Scurra: Keep the front pictures of your Christmas cards. Deal them out to the players at random

Scurra: Then someone chooses a category - "fattest Santa" or "campest Angel"

Scurra: and you all put a card in, a la A2A

FastLearner: LOL, that sounds like a hoot!

Scurra: It is, I promise you.

Scurra: You have to put some rules in - people can't just throw random cards in like in A2A

Scurra: and if two people put the same picture in (for some reason) they are both out

Scurra: But it's a great New Year's party game

FastLearner: Sounds perfect for New Years, Scurra.

Scurra: We were playing it way before A2A appeared too

tjgames: I think sometimes your stuck with AP

FastLearner: Any other Analysis Paralysis thoughts, y'all?

tjgames: Play with fast players

FastLearner: Heh, good point.