Judge a game jam on three things, weighted so the game itself decides the result: game quality 50%, creativity and theme 30%, presentation 20%. That is the rubric the OYU GameDev Bootcamp 2026 used to rank 13 teams competing for a $16,000 prize pool at Odlar Yurdu University in Baku, and it held up because every team knew it before the jam began.
Most judging advice is written for hackathons, where a project is pitched and the question is whether the idea has impact. A game jam is different. The entry is something you play, and the best evidence of quality is what happens in the first two minutes with a controller in your hands. The criteria below are built for that.
The rubric at a glance
| Criterion | Weight | What judges are measuring |
|---|---|---|
| Game quality | 50% | Is it fun, does it play well, does it feel finished? |
| Creativity and theme | 30% | Did the team do something of its own with the theme? |
| Presentation and pitch | 20% | Could the team show the game and explain its decisions? |
Three criteria is deliberately few. Past four or five, judges stop telling them apart and every score drifts toward the middle. OYU's rubric broke each category into sub-criteria so judges had something concrete to score. The breakdowns below are our recommendation for the same weights, not a copy of OYU's own.
Game quality: 50%
Half the score goes to the game because the game is what the jam produced. A team that builds something genuinely fun to play should not lose to a team with a better slide deck, and a 50% weight makes that close to impossible.
Split it into parts a judge can observe while playing:
| Sub-criterion | Share of the category | Share of the final score |
|---|---|---|
| Fun and core loop | 35% | 17.5% |
| Controls and playability | 25% | 12.5% |
| Art, audio and feel | 25% | 12.5% |
| Stability and completeness | 15% | 7.5% |
Fun and core loop carries the most because it is the hardest thing to fake in 48 hours. Ask whether the judge wanted to keep playing once they understood the game.
Controls and playability covers whether a first-time player can pick it up without the team explaining it. A great idea behind unreadable controls scores low here, and that is fair: players outside the jam will not have the team standing next to them.
Art, audio and feel is polish, but judge cohesion rather than production value. A consistent style drawn in an afternoon beats mismatched store-bought assets.
Stability and completeness is weighted lowest on purpose. Jam games crash, and a rubric that punishes every bug rewards teams who played it safe. Score whether the judge could reach the end, or at least the point of the game.
Creativity and theme: 30%
The theme is what makes a jam a jam rather than a weekend of solo projects. OYU revealed three themes (Urban, Connection and Escape), and a 30% weight made interpreting them worth real effort without letting it outweigh a working game.
| Sub-criterion | Share of the category | Share of the final score |
|---|---|---|
| Theme interpretation | 50% | 15% |
| Originality | 50% | 15% |
Reward interpretation over illustration. A game that puts the theme's word in its title scores lower than one where the theme changes how the game plays. Tell judges this before they start, because the literal reading is the one that jumps out on a first pass. If you still need a theme, our 50 game jam theme ideas run from one-word prompts to mechanic constraints, and the open-ended ones leave the most room to interpret.
Originality is judged against the other entries, not against every game ever made. Judges should ask whether they have seen this idea elsewhere in the room.
Presentation and pitch: 20%
Presentation matters because judges have limited time, and a team that shows its game well helps them judge it accurately. It is capped at 20% so a confident speaker cannot carry a weak game.
OYU gave each team five minutes to present and five for questions, kept on the clock with a timer so the last team got the same attention as the first. That format suits this weight well.
| Sub-criterion | Share of the category | Share of the final score |
|---|---|---|
| Demo and clarity | 60% | 12% |
| Answers to judges' questions | 40% | 8% |
Score the demo on whether the judges understood the game by the end of it, not on slides. The questions are where a team shows it made decisions on purpose. "We cut the second level to polish the first" is a good answer.
Anchor the scale so judges agree
A weighted rubric is only as fair as the scores that go into it. Without anchors, one judge's 4 is another judge's 2. Pick a scale (1 to 5 is easiest to hold in your head) and write down what the ends and the middle mean for each criterion.
For game quality, for example:
- 1: does not run, or cannot be played without the team's help
- 3: playable start to finish, with a clear core loop that is fun in places
- 5: the judge kept playing after they needed to, and would show it to someone else
Anchors take twenty minutes to write and save an hour of argument at the end. Our guide to seven jury scoring criteria covers anchors for criteria beyond these three.
Six rules that keep jam judging fair
- Publish the rubric before the jam starts. Teams build toward what you tell them you are measuring. OYU defined its weights up front, so every participant knew how they would be scored before they wrote a line of code.
- Every judge plays every game. OYU's three judges scored all 13 teams. When there are too many entries for that, give each game to at least two judges and normalize their scores, so a strict judge's panel is not ranked against a generous one's. How judging software keeps scores fair explains how normalization works.
- Declare conflicts and step aside. A judge who mentored a team, works with a member or is related to one does not score that team. Decide what happens to that team's missing score before the jam, not during it.
- Say how AI tools are treated. OYU's rules included an AI-use disclosure policy. Whatever you allow, put it in the rules and tell judges whether disclosed AI assets affect the art or originality scores, so two judges do not apply two policies.
- Set the tiebreaker in advance. The simplest is the higher game quality score. Announce it with the rubric, so a tie is settled by a rule rather than a debate.
- Score live, reveal later. At OYU the jury scored each team as it presented, so the leaderboard was ready the moment the last team finished. Hold the reveal until the scores have been checked, and have a plan for results day that includes certificates. Our free certificate generator makes one for every participant from a list of names.
Adjusting the weights for your jam
The 50/30/20 split suits a competitive jam with a jury and real prizes. Change it deliberately when your jam is different:
- A beginner or student jam: drop stability and completeness to a small share and move the weight to fun. First-time teams learn more from finishing something playable than from polishing it.
- A theme-led or experimental jam: raise creativity and theme to 40% and take it from presentation, so bold interpretations are rewarded even when they are rough.
- An online jam with no live pitches: replace presentation with a short written description and a gameplay video, and weight it at 10 to 15%.
If you would rather start from four criteria, our free judging criteria builder has a game jam template (fun, theme interpretation, originality and presentation), checks that your weights add up to 100%, and exports a score sheet for your judges. For hackathons, where impact and technical execution matter more, see hackathon judging criteria.
FAQ
What criteria should a game jam be judged on? Game quality, creativity and theme, and presentation. Weight game quality highest, at around 50%, because the game is what the jam produced. Creativity and theme take about 30%, and presentation about 20%.
How much should presentation count in a game jam? No more than 20%. Enough that a clear demo helps judges understand the game, but not so much that a confident speaker can win with a weak game.
Should judges play every game? Yes, whenever possible. Watching a demo tells a judge how the game looks, but playing it is the only way to judge controls and fun. If there are too many entries, have at least two judges play each game and normalize their scores.
How many judges does a game jam need? Three is a practical minimum: enough that one judge's taste does not decide the result, and small enough that every judge can play every game. OYU GameDev Bootcamp 2026 used three judges for 13 teams.
How do you break a tie in a game jam? Set a rule before the jam and announce it with the rubric. The most common is that the higher game quality score wins, since it is the criterion that carries the most weight.
Should the theme be its own criterion? Yes. If theme only counts inside originality, teams learn they can ignore it. Giving theme interpretation its own score, around 15% of the total, keeps the jam about the theme you chose.
