Twelve people in a living room is the awkward size for games. It is too many for one board game and too few to split into separate rooms, and the first twenty minutes usually go on deciding what to play and who is with whom. This is a plan for a full game night at that size, about an hour of play, using three randomizers in a row: one to make teams, one to deal charades cards with a timer, and one to run a category race. Every number below is worked out for twelve guests, so you can see where the luck lands before anyone sits down.
The running order
| Block | Tool | Teams | Play time |
|---|---|---|---|
| 1. Split the room | Random Team / Group Divider | 2 teams of 6 | 2 minutes |
| 2. Charades | Random Charades / Pictionary Word | Same 2 teams | about 20 minutes |
| 3. Re-split | Team divider again, or the Balanced Team Generator | 3 teams of 4 | 2 minutes |
| 4. Category sprint | Category Sprint | Same 3 teams | about 25 minutes |
| 5. Final | Category sprint | Top two teams | 5 minutes |
The re-split in the middle is deliberate. Charades works best with two big teams, because more guessers means faster guesses. A category race works best with three small teams, because everyone has to talk. Changing teams halfway also stops a bad first draw from souring the whole night.
Block 1: two teams of six
Type the twelve names into the team divider, one per line. Under Mode, leave it on By number of teams and set Number to 2. Put something fun in Team labels, say Owls, Foxes, and the result cards use those names instead of Team 1 and Team 2. Click Make teams.
“By number of teams” shuffles the list and deals it round-robin, so every team is within one person of the others. With twelve and two, that is exactly six each. The other mode, “By size per team”, cuts the shuffled list into chunks of the size you give and leaves any remainder in a smaller last team, which is fine for 12 in fours but would leave a team of two if a thirteenth guest walks in. Use “number of teams” when you care about even sides.
The question every host gets asked is whether couples are together. On a straight random split into two teams of six, any two named people end up on the same team with probability 5 in 11, about 45%. That is because once one of them is placed, there are 5 remaining spots on their team out of 11 remaining spots. If there are four couples among the twelve, the chance that all four are split up by luck is only about 10% (96 of the 924 possible ways to fill the first team). So if your group has a rule about couples, randomness alone will break it nine times out of ten. Section 3 has the fix.
If someone complains the split was rigged, click Make teams again in front of them. There is also a Seed box: type any word and the same list with the same seed always gives the same teams. That is mainly useful if you want to send the teams out in a group chat before the party and have them match what you show on the night.
Block 2: charades with a clock
The charades tool needs no list. Pick a Category (Movies, Actions, Animals, Songs, Kids or Occupations), set Timer seconds (60 by default), and click Draw a card. The word appears in large type with the category above it and a countdown below. The last five seconds tick if sound is on, and a ding marks zero. Setting the timer to 0 turns the clock off.
Turn order: alternate teams, and each player acts once, so twelve turns. At sixty seconds of acting plus about forty-five seconds to hand the phone over and settle, that is about 21 minutes.
Two things about how the cards are dealt matter for a whole evening.
Each draw is independent. The tool picks one card at random from the chosen category every time, with no memory of earlier cards. Each category holds 32 cards. If you leave it on Movies for all twelve turns, the chance that at least one card comes up twice is about 91%. If you rotate through all six categories, two turns each, the chance of any repeat drops to about 17%, because a repeat can only happen between the two draws in the same category. Rotating categories is better for variety anyway: Songs and Occupations play very differently from Movies.
The card is on the screen for anyone to see. The actor should hold the phone, draw, read, and turn it face down. The Recent results panel under the tool also lists the last ten cards, and it stays collapsed unless someone opens it, so it is safe on a phone but worth closing if you are mirroring to a TV. If a repeat does come up, just draw again. Drawing a new card resets the clock, so draw before the actor starts, not halfway through.
Scoring: one point for a correct guess before zero. If you want a tie-breaker, note the seconds left on each correct guess and add them up.
If children are playing, the Kids category is made of simple actions like “Brush teeth” and “Build a sandcastle”, which are easier to act than film titles.
Block 3: three teams of four, and the couples question
Now re-split for the race. If nobody minds who they are with, go back to the team divider, set Number to 3, and click Make teams. Any two people now share a team with probability 3 in 11, about 27%, and the chance a given pair of charades teammates ends up together again is the same 27%. So most people get new teammates.
If couples should be split, the team divider cannot do it, but the Balanced Team Generator can, with a small trick. It reads a tag in square brackets after each name and deals each tag’s members out to the teams one after another. Give each couple its own tag:
Priya [c1]
Tom [c1]
Hana [c2]
Leo [c2]
Sam [c3]
Jo [c3]
Ade [c4]
Mira [c4]
Chris
Dev
Nina
OwenSet Number of teams to 3 and click Build balanced teams. The two members of a tag are always dealt to different teams, so no couple lands together, and the four untagged guests fill the gaps. The teams still come out at four each. How lucky would you need to be to get that from a plain random split? With four couples and three teams of four, only about 29.5% of random splits keep every couple apart, so you would expect to click two or three times before getting one. The tags make it certain.
One honest limitation: the balancer goes through the tags in the order they first appear in your list, so the pattern of which teams share couples is fixed by your typing order. Which partner goes to which team is random. For a game night this does not matter, but it is not a fully random split.
Block 4: the category sprint
The category race is the loudest part of the night. One team at a time gets a category, and they shout out as many valid answers as they can before the clock runs out. Another team keeps count.
Open the category sprint. It deals a first category as soon as the page loads. In the options, set Timer seconds (anything from 10 to 300; 60 is the classic length) and a Category pack: General, Kids, Food & drink, Travel, Entertainment or Holiday. The tool remembers those two settings in your browser. Click Start to begin the countdown, Stop timer to pause it if a dispute breaks out, and New category to deal the next one. When time runs out the screen shows “Time’s up!” and Start stays disabled until you deal a new category, which stops a round accidentally starting before the next team has read their prompt.
Run four rounds per team, so twelve sprints. At a minute each plus a minute for counting and arguing, that is about 25 minutes.
Again, each category is drawn independently, so repeats depend on pack size. The packs are not equal:
| Pack | Categories | Chance of at least one repeat in 12 rounds |
|---|---|---|
| General | 168 | about 33% |
| Entertainment | 45 | about 80% |
| Food & drink | 40 | about 84% |
| Holiday | 30 | about 92% |
For a game night in late October, the Holiday pack is tempting because it has Halloween costumes, Halloween candies, Halloween monsters and “Things at a Halloween party”. But with only 30 categories, a twelve-round game will almost certainly repeat one. Use General for the main rounds and keep Holiday for the final. If a repeat comes up, click New category before starting. Nobody has seen the prompt long enough for it to count.
Judging works best with two simple rules decided before round one: an answer has to be specific enough that the counting team agrees it fits, and no answer can be used twice in the same round. The earlier category sprint hosting guide has more on settling disputes.
Block 5: a final that feels earned
Add up the sprint scores and take the top two teams into a final. Set the pack to Holiday, set the timer to 45 seconds to raise the pressure, and give each finalist two categories. Highest combined total wins.
If two teams tie for second place, settle it with a single sprint at 30 seconds, same category for both, with the other team out of the room for the first run. That is fairer than a coin flip, because the decider is still a test of the game you were playing.
Timings at a glance
| Block | Turns | Length per turn | Total |
|---|---|---|---|
| Charades | 12 | 60 s + 45 s changeover | about 21 min |
| Sprint rounds | 12 | 60 s + 60 s counting | about 24 min |
| Final | 4 | 45 s + 60 s | about 7 min |
| Team splits | 2 | about 4 min |
That comes to just under an hour, which is about the attention span of a room of twelve adults after dinner. If you have more time, add a second charades block before the sprint, using the categories you skipped the first time.
What the tools leave to you
None of these tools keep score. The charades and sprint tools deliberately stop at a prompt and a clock, so you need a sheet of paper or a notes app for points. The team divider does not remember previous splits either; if you play every month and want people to meet new teammates, the No-Repeat Group Mixer remembers past pairings and steers away from them.
What the tools do take off your hands is the part that usually stalls a party: deciding teams in a way nobody can argue with, having a card ready the moment the actor stands up, and having a clock that everyone can see. Set the options before the guests arrive, put the phone on a stand, and the evening runs itself.