How we developed the game, where there are no fixed rules

One of our tester ogreb child shovel for sand on the face. The reason is very simple - agreed to play with the small child in the kulichiki. And then he began to defend the originally voiced version of the rules. But it’s impossible to do with children - they change the rules on the fly as they realize that they can lose. Many adults do the same.

The result of the dispute is logical. Well at least it did not come to bytovuhi.
')
Thus began the history of the development of the game without rules. On her example, I will illustrate how mechanics are tested and tested, and which loops can be cut during the search.

Set the vector space

Since the game is without rules, the first question is with what it will be played. More precisely, with what gaming tools to start development. In board games, there are several pretty basic things: a means of hiding information (maps), random number generators (coins, dice, all sorts of hands and roulettes), timers (eg, hourglasses), open statuses (playing fields, for example, strategy cards), different drives (counters, resource markers, money), components for agility (tower blocks), and so on. Classic games are developed, as a rule, on information hiding tools and random number generators - this is, for example, a deck of cards. The second type of classic games is fully deterministic games, but with very high computational complexity (go, chess).

A deck of cards as a base set fit perfectly. Therefore, to begin with, we installed a test copy of the game in a predictable state and used the classical mechanics of “Drunkards” - more precisely, rather familiar to you from “Uno”, “Svintus”, “Pirate cards” and even dominoes.

So, in a basic state, players do very simple things:

1. Take 5 cards
2. Spread cards one by one in the middle of the table. If there is a green five - you can put on it either any green card, or a five of any color.
3. If there is no card, you need to pull it further from the deck.
4. The winner is the one who first get rid of their cards.

This mechanic is pure degenerate case where the skill of the game almost does not solve anything (which is why many people like Uno so much, for example). But it is great for the base.

Version with closed information

Now we try the first hypothesis on how the game should go. We take and write a few closed rules (20 pieces for a test) like "you can put only triples on a two", "when you put the seven, you have to hit the table", "you can put only even numbers on red numbers". Then we plant four players and give everyone a closed rule.

The first player walks. The other three are checking whether he did the right thing in terms of the rules he knows. If yes - move to the second player. If not, they roll back his turn (give the laid out card back to him) and inform him that he needs to draw the card.

After 10-12 moves, players begin to get hypotheses about what can and cannot be done. This part of the gameplay is extremely interesting and requires a lot of mental stress, the players like it very much.

A tester recalls a stroke with a spatula in the face and reports that he has similar sensations. Immediately appears the name of the game - "Trump Montana" on the appropriate anecdote. Reduced to "Montana".

The following are quite unpleasant side effects:

1. A player who knows all 20 rules, has a decisive advantage in the game with beginners. On the other hand, a company of experienced players does not feel a fan - it simply goes through the rules mathematically, the benefit of which it has. The game turns into a business-specific XLS-etude, and not the discovery of something new. In theory, this can be solved by introducing hundreds more rules, but ...
2. Players catch razvilya, from time to time forgetting about their rule or incorrectly applying it, which introduces numerous distortions in the hypotheses of other players. In addition, the game has no protection against cheater. From the very first move, the party can “skew” without the possibility of clearly rolling back or rebalancing the state.

Thus, due to (2), we now need to check why it is impossible to make such a move when you are returned a card. This can be done in two ways:

• Trust information transfer.
• The arbitrator.

In our case, either the player who found the mistake of another, shows him his rule, or there is a moderator who is following everything.

Board game with the lead is immediately an unconditional commercial failure. In our development there is a simple rule - all players must play on equal terms. Company games are not “indulgent” games of parents with children, when the goal is to educate a child. Here the goal is to get everyone high. Lead falls out of this process. Being leading is boring, and therefore games with him are immediately swept aside by the invisible hand of the market.

It is impossible to transfer information - all the rules will be revealed too quickly, and the game will degenerate. Ok, we are trying to finish this thread - add rules to the game, giving each one not one rule, but two or three.

Here we run into another classic limitation - the performance of human subsystems. Players can not keep in mind so many entities (their three rules, about 10-15 hypotheses and already "explored" the rules of other players, plus watch the cards on the field).

The result - we refuse this branch of closed rules. But we note the mechanics as suitable for testing in other conditions (I can tell about it separately).

Please note that we have not even tried to use any mathematical or intuitive balancing methods yet. Just the game in this version did not go.

Version with open information

We continue the experiments. We take the same cards that we had with the rules, and lay them in the deck of ordinary ones. Now they can be played right on top of the table, so that they change the rules of the game every turn.

That is, you can lay out a card with a number, but you can - a new rule.

The first batch looks very funny and wildly happy. The rules change very often, but players know what they are doing. We change players - since these have already tested the first iteration and are used to the rules of the prototype. New ones read the rules with great difficulty and cannot maintain the dynamics of the game.

Conclusion - the rules are too difficult to understand. They cannot be read quickly (as a map icon, for example) due to their diversity. They are uncomfortable to memorize - because as soon as you master a large piece of information, it will immediately be replaced by a new one that needs to be mastered and applied.

When players learn the rules, the strategy comes up - they change too often, and you can’t count your actions. This reduces the "playability", that is, the player's sense of control in this case. The game turns into Russian roulette.

So, there are problems:

• Too complicated rules.
• Too often change.
• Heavy load in the head.

We try to work out a branch of constancy of rules. Igromechanically, they are still taken from the general deck along with regular cards, but they are laid out side by side on the field and are valid until the end of the game. Fun, but withstand 10 rules - very difficult. We leave for now as a controversial hypothesis and look for a better way.

There is a problem with the fact that in the mechanics of the rules in the main deck there is often a player with two rules on his hands - and he will inevitably win because his next two moves are predictable. This reduces the pleasure of other players and creates an unnecessary element of "karma", that is, inevitability in the game. We work on the version by putting the numbers on the rule cards. That is, we look, whether it is possible to lay them out not always, but only at certain moments, when the field allows. It turns out very interesting, the players like it. But this immediately puts the rules into the game - that is, they must be on the stack and overlap with new ones almost immediately. In the game, it turns out, only one last rule is valid.

In parallel, we solve the problem of the complexity of the rules. We try to limit the mechanics of gameplay changes to a minimum. For example, “an action on a certain card” (“tap a foot when you put the seven”, “you can put any card on a deuce”), victory conditions (“type three odd cards”) and a move action (“turn transfer goes in the opposite direction” ). All sorts of great meta-opportunities like “take a card if you spoke during the turn”, “pull the card if you are facing north” are leaving - in general, everything we have dragged from the idea of ​​asprin's dragon poker.

With a limited set of rule types (5 types, each with 3 actions), it is possible to make icons that are easily remembered and readable. For example, the number and icon of the clap - it is necessary to clap during the course of the card of this nominal. And so on. From the hypothesis of the icons it follows directly that the cards should be placed next to the field, and not on the field itself - so that they are always visible. We try. The result is, it seems, good, but it’s painful to fix the rules. It is difficult to memorize.

We apply the solution from the previous dead-end branch - we limit the number of rules. In this case, the mechanics mutates to "there are three slots for the rules, you can put your own in any." Seems alright. At this moment we crawl off the draft prototype of the game (alpha) and begin to do the beta. Beta is different in that the deck will consist of a real number of cards, all the rules will be spelled out exactly and more or less balanced by the non-domination method (I wrote more about it here ), the cards themselves will be of factory quality (simple mixing) and with which something more or less distinct design, rather than pencil scribbles.

On this beta, you can evaluate the quality of the game "in the collection." The process of its manufacture, in fact, is not very trivial, and requires serious work. But this work, in general, is predictable, so the entire development team does not participate in it - this is done by one specialist.

Go to the tests again. More precisely, we are already giving the beta to the testing team. These cute people "swing" the game for quite a long time, with three different compositions of players (the last iteration is people from the street or real game stores, where the game is enclosed in a pile like a real serial).

The report from the beta test makes us change a bit of mechanics (split cards into three types and push each type of rules into its slot), make the icons clearer, even slightly shorten the list of rule modifiers, change a lot in the wording of the rules (this is a common situation).

We are balancing the game according to the basic resource method (in Montana it is really very simple and does not require any mathematical knowledge other than intuitive). Do another beta. We drive for combat tests (long, the goal is to identify the vulnerabilities of rules and balance inconsistencies to search for the dominant strategy). Another iteration of minor edits (by this time the path of the game took about a year). On the way, the art and layout appeared just in time, so we are assembling the final prototype. It looks like a serial game, only slightly worse in print quality. We look, we twist in hands, we bring in a couple of little things regarding the layout and architecture of the box.

Is done. Printing, assembly, retail.



And here we are.