Skip to Content

Further Analysis Paralysis Discussion

8 replies [Last post]
jwarrend
Offline
Joined: 08/03/2008

Hey, how was the chat last night? This is a topic I'm particularly interested in, as some of my games seem to take way longer to play than I think they ought to. My guess is that this is in general because I design my games with a "phased" turn structure in which players each take one or two short, punchy actions. This dramatically reduces downtime, but unfortunately, increases the number of decisions a player must make.

Perhaps the best "solution" I've seen is that of a game I've mentioned several times, Wallenstein, where players select their actions all at once, and then the turn is spent just carrying out those actions, most of which are deterministic (the only exceptions being combats). I like this because it puts all the decisions in one place, and has all the players making their decisions simultaneously. So, it's probably as close to perfect as you can get in terms of reducing Analysis Paralysis.

But, I'd be very interested in hearing the results of the chat. Could there be a forum specifically for chat transcripts? Is it difficult to provide them in the archives? I really like the topics you choose, but can't be available to actually participate in the chats themselves...

-Jeff

zaiga
Offline
Joined: 12/31/1969
Re: Chat transcript -- Analysis Paralysis?

jwarrend wrote:
Perhaps the best "solution" I've seen is that of a game I've mentioned several times, Wallenstein, where players select their actions all at once, and then the turn is spent just carrying out those actions, most of which are deterministic (the only exceptions being combats). I like this because it puts all the decisions in one place, and has all the players making their decisions simultaneously. So, it's probably as close to perfect as you can get in terms of reducing Analysis Paralysis.

That doesn't solve analysis paralysis. Players just suffer from analysis paralysis simultaneously, which reduces downtime, but it doesn't necessarily solve analysis paralysis although it does make it a non-problem, because the real problem is downtime not analysis paralysis.

Analysis paralysis often introduces downtime (except in the cases where players make decisions simultaneously such as in Wallenstein or RoboRally). Downtime is not necessarily always the result of analysis paralysis. Downtime is generally bad, so if the source of downtime is analysis paralysis then you have to tackle that.

Jeff, I wonder if AP is the source of your problem? Do players take a long time making their decisions? Or is the problem the fact that playingtime is too long? It seems to me that by dividing the game into small phases with only a few choices per phase, you already tackled the most important causes of downtime (AP and long time between turns).

Wolfgang Kramer mentioned that AP is a problem of the players. I do not agree with that. Certainly, it is sometimes true. If the game is simple and decisions straightforward, there is no reason for people to take a long time making a decision. In that case it IS the fault of the player(s).

However, I do believe there are many tricks that a designer can employ to prevent AP.

- Reduce the amount of things to choose from in a player's turn. If a player can choose between 2 things he will reach a decision faster than when he has to evaluate 7 different choices.

- Reduce the amount of permutations of choices in a single turn. For example, if a player can take 2 actions and can choose from 3 different actions than he can perform AA, AB, AC, BA, BB, BC, CA, CB or CC -- 9 different permutations. Imagine how many permutations a player has to evaluate when he has a total of 4 actions and 7 possible choices for each action! I think this is the most important point. Action point system often suffer from analysis paralysis by having a large amount of permutations of choices.

- Introduce "hidden" information or a luck component to make the outcome of a choice non-deterministic. This forces a player to approach the problem with a "more-or-less" attitude, instead of trying to come up with the perfect solution to a deterministic problem.

- Introduce the "Human Element" which is similar to hidden information or luck, except that in this case it is the unpredictability of opponents which makes the outcome of a choice non-deterministic.

There are probably more solutions to AP and I love to see the chat transcript!

- Rene Wiersma

SVan
Offline
Joined: 10/02/2008
Further Analysis Paralysis Discussion

AP to me is when a player needs to make a choice but freezes because the outcome for each decision is too hard to tell which one is better or which one is worse. When there are 3 or more decisions like this, the downtime can be horrific.

Sometimes players can do it themselves. My last game of Settlers I played lasted over 2 hrs because we had someone that analyzed every move he made (and because of that he won, also.)

I like for players to have plenty of choices, but have found that Roborally type turns are good for the game, especially if the game is a longer game. When each player can make decisions at the same time then it makes the game feel like it has a lot less downtime.

I've heard it said that Puerto Rico has very little downtime. I haven't played that yet, but really want to. What makes this game have very little downtime?

phpbbadmin
Offline
Joined: 04/23/2013
Transcript

The transcript will come. Fastlearner has part 1 and I have part 2.

-Darke

jwarrend
Offline
Joined: 08/03/2008
Re: Chat transcript -- Analysis Paralysis?

zaiga wrote:
jwarrend wrote:
Perhaps the best "solution" I've seen is that of a game I've mentioned several times, Wallenstein, where players select their actions all at once, and then the turn is spent just carrying out those actions, most of which are deterministic (the only exceptions being combats). I like this because it puts all the decisions in one place, and has all the players making their decisions simultaneously. So, it's probably as close to perfect as you can get in terms of reducing Analysis Paralysis.

That doesn't solve analysis paralysis. Players just suffer from analysis paralysis simultaneously, which reduces downtime, but it doesn't necessarily solve analysis paralysis although it does make it a non-problem, because the real problem is downtime not analysis paralysis.

Ok, fair enough, then it's possible I'm using the term incorrectly. My impression was that AP was, by definition, a player problem moreso than a game problem -- that it means, basically, a situation where a player, by virtue of having a lot to think about, spends a lot of time making his decision. However, it seems you're using it slightly differently, more like what SVan is saying, that it's related to no clear way to decide which choice is "better". What definition are you working off of?

Quote:

Jeff, I wonder if AP is the source of your problem? Do players take a long time making their decisions? Or is the problem the fact that playingtime is too long? It seems to me that by dividing the game into small phases with only a few choices per phase, you already tackled the most important causes of downtime (AP and long time between turns).

I think in some cases, it's both. In my Civ game for example, which lasts >3 hours, there's just a lot to do, yet in each phase, there's only one action being required of players. I just find that in my group, every time a decision is required, some players will spend a lot of time pontificating (myself included). And that this can bog down even the shortest of games.

The obvious solution is to make games that are more abstract, or have less things going on. Yet, I tend to like, as a designer, games that have different things to think about. I think I just haven't always found the happy medium of how many choices you can give someone and still have it be calculable in a short period of time.

Quote:

Wolfgang Kramer mentioned that AP is a problem of the players. I do not agree with that. Certainly, it is sometimes true. If the game is simple and decisions straightforward, there is no reason for people to take a long time making a decision. In that case it IS the fault of the player(s).

However, I do believe there are many tricks that a designer can employ to prevent AP.

- Reduce the amount of things to choose from in a player's turn. If a player can choose between 2 things he will reach a decision faster than when he has to evaluate 7 different choices.

- Reduce the amount of permutations of choices in a single turn. For example, if a player can take 2 actions and can choose from 3 different actions than he can perform AA, AB, AC, BA, BB, BC, CA, CB or CC -- 9 different permutations. Imagine how many permutations a player has to evaluate when he has a total of 4 actions and 7 possible choices for each action! I think this is the most important point. Action point system often suffer from analysis paralysis by having a large amount of permutations of choices.

These are definitely good points to keep in mind! I am working on an action point based game now and this is a good bit of wisdom for me to be thinking about. I agree that too many decisions can lead to too much to think about. But too few decisions is also bad. It definitely has increased my respect for the "great designers" as I've tried to start designing games myself, that a "simple" game engine that still has "agonizing" decisions is really a tough thing to create. Yet, the games themselves (Web of Power, New England), seem so simple and "effortless" that I wonder why it's hard to create them myself!

Quote:

- Introduce "hidden" information or a luck component to make the outcome of a choice non-deterministic. This forces a player to approach the problem with a "more-or-less" attitude, instead of trying to come up with the perfect solution to a deterministic problem.

This is a good point. One thing I'm getting away from is "hidden information that was once public", like "closed holdings in Acquire" for example. I personally like playing by impression and trying to remember what everyone has, yet I think that "memory" based stuff can overload people's minds. So, I am trying to have public information stay public in general, although that too can give people too much to think about. And, I think it gives way too easily to "hit the leader", but that's another matter...

Quote:

- Introduce the "Human Element" which is similar to hidden information or luck, except that in this case it is the unpredictability of opponents which makes the outcome of a choice non-deterministic.

Speaking for myself, this makes my decisions take longer; I often spend a lot of effort trying to guess "what's this guy going to do?" and then playing accordingly based on all possible eventualities. I am a horribly slow chess player for this very reason. Obviously the human element introduces variability and unpredictableness and that's good, but it doesn't, in my experience, always reduce AP.

Thanks for some great thoughts!

-Jeff

zaiga
Offline
Joined: 12/31/1969
Re: Chat transcript -- Analysis Paralysis?

jwarrend wrote:
Ok, fair enough, then it's possible I'm using the term incorrectly. My impression was that AP was, by definition, a player problem moreso than a game problem -- that it means, basically, a situation where a player, by virtue of having a lot to think about, spends a lot of time making his decision. However, it seems you're using it slightly differently, more like what SVan is saying, that it's related to no clear way to decide which choice is "better". What definition are you working off of?

My definition of analysis paralysis is: the inability of a player to make a choice within a reasonable amount of time, caused by the player wanting to analyze each move and each possible outcome of each move. So, it is a player problem in a sense, but I think a good design can try and salvage it somewhat.

Quote:

I think in some cases, it's both. In my Civ game for example, which lasts >3 hours, there's just a lot to do, yet in each phase, there's only one action being required of players. I just find that in my group, every time a decision is required, some players will spend a lot of time pontificating (myself included). And that this can bog down even the shortest of games.

I would say that is AP, yes.

Quote:

These are definitely good points to keep in mind! I am working on an action point based game now and this is a good bit of wisdom for me to be thinking about. I agree that too many decisions can lead to too much to think about. But too few decisions is also bad. It definitely has increased my respect for the "great designers" as I've tried to start designing games myself, that a "simple" game engine that still has "agonizing" decisions is really a tough thing to create. Yet, the games themselves (Web of Power, New England), seem so simple and "effortless" that I wonder why it's hard to create them myself!

Well, those are the things that distinguish a mediocre design from a good design and there is a fine thread between "too many choices"and "not enough choices". This is stuff that should be tweaked during playtesting.

Quote:

This is a good point. One thing I'm getting away from is "hidden information that was once public", like "closed holdings in Acquire" for example. I personally like playing by impression and trying to remember what everyone has, yet I think that "memory" based stuff can overload people's minds. So, I am trying to have public information stay public in general, although that too can give people too much to think about. And, I think it gives way too easily to "hit the leader", but that's another matter...

I have been thinking about this a lot. In principle I agree with you, information that was once public should stay public, otherwise you are creating a memory aspect that might overload people's mind and if a game is not about memorizing things, it seems like a nuisance.

On the other hand, mane great games like Euphrat & Tigris and Puerto Rico make use of hidden VP's and in Taj Mahal or Settlers the cards in one's hand may not be shown even though everyone could technically keep track of what every one has in his hand (as the card draws are open).

Fact is that 99% of the people do not possess a perfect memory. This means that often people will have some idea of who is in the lead in E&T and Puerto Rico, but not exactly by how much because there are so much VP transactions that a player cannot keep track of them all. This means that in the endgame it is not possible to exactly calculate everything, you have to go by feel and in a sense this prevents AP. I think this is a good thing.

Quote:

Speaking for myself, this makes my decisions take longer; I often spend a lot of effort trying to guess "what's this guy going to do?" and then playing accordingly based on all possible eventualities. I am a horribly slow chess player for this very reason. Obviously the human element introduces variability and unpredictableness and that's good, but it doesn't, in my experience, always reduce AP.

I don't think this is much different from the "hidden" aspect. After all, one could also keep in mind what die rolls or card draws are possible and start calculating all the permutations with that in mind. I agree that it doesn't always reduce AP, it depends on how it is implemented.

For example, the blind bidding phase in Amun-Re, you know if you want to bid high or low, but you cannot exactly caluclate if you should bid 6 or 7, so here you have to rely on intuition rather than on calculating skills. However, I could see how this element could introduce AP into a game like Puerto Rico: "If I take the Settler, then John will probably take the Mayor and Mary the Craftsman, which means I can take the Trader, unless Mary takes the Trader herself, in which case I may be better off taking the Prospector, otherwise John will... ", etc.

Maybe Wolfgang Kramer was right after all ;)

One last tip. Often you'll see AP during the endgame, because players are trying to squeeze te most out of their last moves. A good way to prevent this is by having the game end in a somewhat unpredictable way. Example: Union Pacific end when the 4th dividend card is drawn, but when this card is drawn is not exaclty known. Another example is Ra, which ends when the Nth (depending on number of players) Ra token is drawn. Incidentally this not only prevents AP (somewhat) it also makes endgames often quite tense and makes for some exciting push-your-luck scenarios.

- Rene Wiersma

Torrent
Torrent's picture
Offline
Joined: 08/03/2008
Further Analysis Paralysis Discussion

I think one of the really good examples of Analysis Paralysis is Chess. Except in Chess it almost isn't even a problem, but part of the culture of the game. The AP in Chess is due to the large number of posible combinations backed up by the reletive painfulness of making a mistake. You only have so many pieces, each is fairly unique and is painful to lose. This goes into what was mentioned earlier in the thread about EndGames causing more AP. At that point people are more aware of points and winning, thus decisions seem to take on more 'worth'. IE a wrong decision feels to hurt more.

On the other hand, games like Puerto Rico and Tigris (disclaimer, I haven't played either) seem like any AP that they cause is due to the number of players times number of variables to watch. A choice that affects a lot of different variables, even if there are only a few choices possible, will tend to spark more AP than fewer variables.

The thing that comes to me in writing this is that decisicions seen as more important will cause more AP. So in theory if you could make a game with lots of little decisions that each didn't affect much, it would have less problem with AP. Games where making a decision can give perceptively large gains or loses will tend to spark more AP in people that are prone to it. I think everyone has a little bit of it, and is affected differently. Note the use of Perceptively Large Gains/Loses. Even if the actual gains/loses in a choice aren't big, if they seem to be big, it will cause more AP.

Thus I claim that if you can make the decisions less perceptively large in scope, you should be able to reduce the amount of pondering going on. Fewer variables affected by each decision, and few choices helps to accomplish this, limiting the scale of any decision. THe downside of this of course is for a game to have some deal of 'heft' to it, these little decisions would have to make up the same amount of 'gaming ground' in change that the big ones would, which invariably means more of the little decisions. This could add play time, but you have to decide if the increased play time is balanced by the reduced 'downtime'.

Andy

Anonymous
This just worked for me...

I used action points in my RTS/ Wargame hybrid Tyrant and it involves having up to four actions. Each action could be one of seven different things, so dealing with AP was a real concern, in fact the biggest. So I used a time honored tool: FEAR.

I introduced a certain politcal element into the game, players behind the active player would control a powerful neutral force, after the active player's turn. The effect was that the player's that took too long got crushed by the neutral forces penalizing them. Also, I introduced cards that could be played during the course of anyone's turn (like Magic Instants), so that while one player took their time the others blasted each other with cards, or blasted the slow player. So effectively the players dictated the pace of the game. This is surprisingly effective, because different groups or ages will want to play the game at different speeds, so it was built into the game mechanics. If players are allowed to interact with each other in very direct ways, AP tends to be less problematic.

I also introduced resource cost to cut down on the question of what to do, or rather provide weight to making decisions by assigning them a value in resources.

For example to create a superbeast, players must spend most if not all their resources under their control. So if they chose to do that they must spend the rest of their actions doing things that cost no further resources. So some of the choices have just been filtered out by making the first one, and players don't sit there wondering.

Of course the famous egg timer can reduce AP forcibly to the dismay of some players. I do remember an older post on the forum regardling this very subject.

IngredientX
IngredientX's picture
Offline
Joined: 07/26/2008
Further Analysis Paralysis Discussion

SVan wrote:
I've heard it said that Puerto Rico has very little downtime. I haven't played that yet, but really want to. What makes this game have very little downtime?

PR has its share of downtime, but not as much as one would expect from a game of its heft. The game works by simulating the production of goods. Each step in the chain of production is represented by a "role." Each player selects a role on his turn. When that role is selected, everybody performs the step represented by the role. The player who selected the role gets a bonus, depending on the role.

This reduces downtime, because a given player will be doing something on everyone else's turn. You can't really leave the table after your turn, because you'll be doing something when the next player chooses his role.

It's important to note that while the mechanic reduces downtime, it does not reduce Analysis Paralysis. If anything, there should theoretically be more AP, because players have to make so many decisions! Fortunately, most of the choices presented in the game do not have a very wide range of possibilities, so min/maxing is not much of a factor. Every once in awhile, AP does bog the game down. This usually happens late in the game, when a player tries to mentally juggle the effects of the selections of different roles, and how they would benefit/hurt other players. It seems to happen only a few times each game, though; again, surprisingly few times from a game of its depth.

If you haven't played it yet, try to! It's an awfully well-designed game.

Syndicate content


forum | by Dr. Radut