Rocco Privetera
United States New York

Doesn't make sense? Read on, fellow designers!
Here's the scenario: You have a game where you have worked out some basic balancing of resources. For example, you've playtested a 6 turn game, and on each turn you want people to get specific resources (call them a and b), to wit:
turn 1: res A 6 units, res B 2 units turn 2: A 5, B 3 turn 3: A 5, B 3 turn 4: A 4, B 3 turn 5: A 4, B 3 turn 6: A 3, B 4
In the game's theme, it makes sense for dwindling A and increasing B. A is the basic money, and B is the thing you need the money to deal with. I want players to feel the crunch as we go.
So far, this is working ok, but I'd like players to not be able to plan quite as well. How can I make something like an event card draw where players draw slightly randomized resources, but the scale doesn't change radically than what I have? That is: so that the slope is about the same, and the starting/ending values are the same (or very close), but the variance in the game is not always the same.
For example: If on turn 2 you get A 4, then on a subsequent turn you get one more resource someplace to help balance it out (hopefully not too far down the line).
Any ideas? I was thinking you could just swap adjacent cards but I forsee the setup for that being weird.
I'm thinking of combining this as cards with my regular events in the game, so I still only have one "Event" draw per turn.

Graham Muller
South Africa Cape Town

A resource bag will work. First turn place out 6A tokens and 2B tokens.
Before turn 2 place the tokens in a bag with an extra B token. then draw 8 tokens. so could be 6/2, 5/3 or 4/4, adjust the resources going into the bag so that there is always some variance. Either place an extra A or B resource in or draw fewer tokens.
EDIT: for the bag, alternatively seed the bag with only Bs, no change and remove 1 A tokens. Players draw one a turn and apply the effect.
Same thing could be achieved with a dice with those faces, but wont have the fixed variance specified number of cubes could provide

Martin Windischer
Austria Innsbruck

Two card decks:
Deck A has cards with the values: 0,0,0,1,1,1 Deck B has cards with the values: 0,0,0,0,1,1 (exact number of the cards has to be tested, probably you want to use more cards than rounds to keep it random even in the last round)
At the start of each turn, reveal the top card of each deck (they stay revealed until the game ends). The value of A is 6 plus the value of all cards revealed from deck A. Value of B is 2 plus the value of all cards revealed from deck B.

Sturv Tafvherd
United States North Carolina

I'd use something inspired by Zombie in my Pocket: Recycling Clock cards.
.... Divide the Overall game into X "seasons"... let's say that's 4 seasons.
Each season has Y "weeks" ... let's say it's 13 weeks.
Overall, the season has an average set of statistics (like: temperature, precipitation) that you'd expect to get. And you can set those averages such that over the course of the 4 seasons, you'll see the overall trend.
However, each week can vary.
I'd prepare 13 cards. Each card has 4 sections, each section corresponds to one of the 4 seasons. Over the course of the game, you track which season you are in, and use the corresponding section on the card. The seasonsection on the card would have the statistics that would, on an overall basis, match a seasonal average.
So... going back to why I said this is ZimP inspired. At the start of each season, you'd shuffle up the 13 cards, and then draw a card for each of the 13 weeks in the season. When you're out of cards, shuffle up the discards and a new season begins.
...
You can even choose to just use a subset of those 13 cards. If you're doing months, then you just need 3 cards to re/cycle.

Sturv Tafvherd
United States North Carolina

MartinWin wrote: Two card decks: Deck A has cards with the values: 0,0,0,1,1,1 Deck B has cards with the values: 0,0,0,0,1,1 (exact number of the cards has to be tested, probably you want to use more cards than rounds to keep it random even in the last round)
At the start of each turn, reveal the top card of each deck (they stay revealed until the game ends). The value of A is 6 plus the value of all cards revealed from deck A. Value of B is 2 plus the value of all cards revealed from deck B.
I like this solution. Only two problems I have with it ... but then again, I don't know what Rocco is designing:
(1) it assumes that the resource value only moves in one direction (always upward or always downward), and the change from one turn to the next is always going to be 0 or 1.
(2) at some point, there's no randomness (as you mentioned). And it can happen much earlier than the last round. Let's say you drew the two 1's from the B deck on the first 2 or 3 rounds ... then you already know that resource B is not going to change from that point.
I guess one problem is that Rocco wants the endpoint(s) to be fixed (or as close as possible). Granted, my suggestion completely ignores that!
I'd make one tweak to this suggestion:
Deck A has cards with the values: 1,0,0,0,1,1,1,2 Deck B has cards with the values: 1,0,0,0,0,1,1,2
So you'd only draw 6 out of each deck's 8 cards... and you never know if the "extreme/unusual movements" might occur.

Martin Windischer
Austria Innsbruck

Another idea:
A deck with 8 cards (exact number can vary) Each card has a value for A and B dependend on the round.
So one card can be: turn 1: A 6, B 2 turn 2: A 5, B 3 turn 3: A 5, B 3 turn 4: A 4, B 3 turn 5: A 4, B 3 turn 6: A 3, B 4
another card can be: turn 1: A 7, B 2 turn 2: A 7, B 2 turn 3: A 6, B 2 turn 4: A 5, B 3 turn 5: A 4, B 4 turn 6: A 3, B 4
and so on...
At the start of each turn reveal one of those cards. Look at the line with the current turn to get the values for A and B.
So for example if you reveal these cards in turn 1 and 2, the values in turn 1 are A 6, B 2 and in turn 2 you have A 7, B 2.

Zack Boatman
United States Tesuque New Mexico

Modified die roll (probably use the same number for both A and B , but you could roll two dice)
Turn one: A gets a +2 B gets a 2 Turn two: A +2 B 1 Three: +1 0 Etc

Nyles Breecher
United States Milwaukee WI

If you're worried about things being too random you can use the deck idea to randomize the order, and then just reveal it all at the beginning. Thus players would know the values per round and could plan around that all game. If you wanted specific distributions you could even make a set of cards for various distributions, so you know that players would always get one of a few specific combinations each game.
This adds setup randomness, but doesn't add ingame randomness, so it may not be the solution you're looking for. The other card solutions proposed so far are what I would come up with. I like best the suggestion with a couple "extreme events" that Sturv posted.

Peter Strait
United States Sacramento California

There are a few ways to have mitigated or controlled randomness. You can lean on the law of large numbers, or you can allow some sort of trading at some ratio. You can also allow rerolls or repeats of the test, though that's essentially just invoking the law of large numbers again.
What you've requested specifically is for the average to follow a specific trend (i.e., for one to grow more scarce over the game and the other to grow more abundant), and for chance to not wander too far away from the average (i.e., some randomness, but not too swingy). Going directly from your example, you could have two slim decks of cards, A and B, and a player could draw a certain number from each deck on each of their turns (more As than Bs drawn for early turns, more Bs than As in later turns). These cards could in turn offer a number within some range (say 1, 2, or 3 of the associated resource), and these could have as finely tuned of a distribution as you felt to be appropriate (say, twenty "1"s, ten "2"s, and five "3"s, versus an even distribution of ten of each). Thus, for example, if a player draws five A cards and three B cards on turn one, and the reverse on turn six, you can tune how likely they are to have a given amount of either resource on a turn, and the multiple tests (both perturn and over the game as a whole) ensures some likely orbiting of the mean.

Peter Strait
United States Sacramento California

Alternatively, you can mitigate through allowing player choice. Consider a similar setup, but with dice: in round one, each player rolls five Amber dice and three Blue dice, and in the last round, three Amber and five Blue. Now imagine all rolled dice get added to a pot in the center that players draft a single dice at a time from.
Pooling the tests and allowing choice via drafting means that players are more likely to see an average set of numbers and to then evenly distribute said numbers among themselves (players will take "3"s until they're gone, then take "2"s, then "1"s, making it much less likely that a single player will either get all "3"s or all "1"s). Note that without some other factor this might be too effective at mitigating randomness and defeat the purpose of rolling dice at all.

Rocco Privetera
United States New York

These are great ideas guys, thanks. I'm trying to find something not too onerous for players  my guess is this is an idea that could be the core of an interesting game all by itself.
For my game, all the players get the same # of resources each turn, so solutions that vary from player to player won't really work. But your suggestions have given me a few places to work on!


