Run a Small Tabletop Trial of a TV Format Before Filming a Pilot

By pitch.dogFoundation14 min read20 September 2026

Television

A tabletop trial is a small enactment of one rule, using paper, tokens and a few people at a table. It is not a cheap episode, and it is not a rehearsal for one. It is a way of buying one specific piece of information before you pay for the expensive version of the same question.

The promise is narrow, and it is worth stating plainly up front. You will not find out whether anyone would watch. You will not find out whether the format is fair, safe or kind. You will find out things like this: when three people are told to commit to a hidden number, do they know at what moment the number stops being changeable?

That is the whole shape of the exercise. Choose one uncertainty the table can settle. Build only the parts of the round that uncertainty needs. Keep the hidden answer hidden. Give the operator a script and the observer a blank sheet. Run a bounded number of rounds. Write down every time you helped. Then change one rule, and name the next question.

The output is a record: the wording you used, the questions people asked, the answers you improvised, and the point at which the round stopped. If the record comes out tidy, you probably wrote it afterwards.

Choose one question the table can actually answer

"Does the show work?" is not a question. It is a pile of them, and most of the pile is out of reach.

Break it up and look for the piece that plays out in a room. Some translations:

Will people enjoy watching this? Not a table question. Nobody sitting at your table is the audience, and their enjoyment measures the table.
Does the round produce tension? Closer, but still loose. Tension is something you infer from people, and inference is what an observation sheet handles worst.
Do players know that their written number is final before the reveal? This one you can watch. Someone will ask. Someone will reach for the slip again.
Can one person read the script card aloud and get three folded slips turned over together without adding a sentence of their own? Also watchable, and much closer to the problem you will have on the day.

The last two are worth building for. They are about the instruction in the narrow sense: what the words require, what order the actions happen in, and what a person is entitled to assume.

The MIT OpenCourseWare paper-prototyping lecture, from the Fall 2014 Creating Video Games course, is organised around roughly this move — take one question, represent the rules cheaply, separate the roles of facilitating, operating and observing, then revise after watching play. That is game-design teaching. Applying it to a television format is an editorial translation, not something the lecture claims to cover.

There is practitioner support for the underlying instinct. A Televisual development piece quotes the UK producers Dom Waugh and Juliet Denison describing the practice of trying game and panel-show mechanisms, and the interplay with talent, rather than assuming a written pitch already establishes how the thing behaves. Those are historical UK interviews. They are not a commissioning requirement anywhere, and they say nothing about audiences.

You also need to write down what your paper model keeps and what it throws away. One paragraph is enough:

Preserved: the order of actions, who knows what at each moment, the exact wording of the instruction, and where a round starts and stops.

Dropped: cameras, lighting, a host under pressure, an edit, a studio audience, and every question about whether the format deserves an hour of anyone's evening.

Keep that paragraph where you can see it. It is the sentence that stops a session from drifting into a conversation about stardom.

Build the rule, the information and the reset

Replace production with the smallest objects that still carry the rule. Tokens stand in for whatever the format moves around. Slips of paper stand in for whatever is hidden. A mat in front of each player stands in for their own knowledge.

Then apply the rule that beginners break most often: give each role only the information it would actually possess at that point.

If the observer has read all three slips before the reveal, the observer will write a different sheet — not dishonestly, just because knowing changes what you notice. If the person playing "player" has already seen the pool, they are not playing the same game. Paper makes this easy to get right, and paper makes it easy to get lazy, because the honest version is usually more paperwork.

Specify three things in writing before anyone sits down: how a round starts, how it ends, and how it resets. The reset is the part people skip, and skipping it is expensive. Improvised housekeeping — when tokens come back, when slips are discarded, who clears the mat — will quietly change the mechanism you are trying to examine. If tokens return to the tray at the start of the next round rather than the end of this one, a player's count at the moment of the reveal is different, and your notes describe a game you did not write.

Separate the operator from the observer

In a small trial you need three functions, even when you have two people.

The operator holds the script card, reads the instructions aloud, collects the slips and reveals them. The operator follows the script. That is the job, and adding to it is a departure worth logging.

The facilitator holds the boundary: briefs everyone before the round, fields questions, and stops things if they need stopping. This is the person players will turn to when something is unclear.

The observer writes. A blank sheet, one per round, with the rule version at the top, and a column for what happened that is physically separate from a column for what the observer thinks it meant.

If you have only two people, one of them will wear two hats. That is fine. Write down which two. A facilitator who answers a question and then reports on the round is a source of an intervention, and the record has to say so.

The observation sheet is worth designing in advance. Fields that have earned their place:

Round number and rule version.
The order of events, as they happened.
Anything the operator said that was not on the script card, written down in the words used.
Questions from players, in their words, and who asked.
Answers given, and whether anyone else heard them.
Slips folded, unfolded, rewritten or taken back.
Where the round stopped, and why.
A clearly headed column for the observer's interpretation, kept apart from the above.

That last column matters more than it looks. B picked her slip back up is an observation. B didn't trust the rule is an interpretation. Both may be useful; only one belongs in the event log, and mixing them is how a session becomes a story with a moral.

Before the first round, agree what an ordinary stop looks like. Anyone can pause or stop, at any point, for any reason, and a stop is a stop — it is not a data point about the format. That much is worth saying out loud once. Arrangements for actual participants are a separate production responsibility, and a paper kit does not discharge them.

Run the bounded round without manufacturing a clean result

Say in advance how many rounds you are running, and what would make you stop early. Two rounds is often enough to expose a wording problem; five is usually a way of enjoying yourself.

Then follow the script and record the departures. This is harder than it sounds, because the natural instinct when a round goes wrong is to fix it and move on.

Consider the specific failure. A player stops halfway and says she does not understand what to write. The facilitator explains. The round finishes. The notes read: round completed; instruction clear. Every part of that sentence is true except the last part, which is exactly the part you needed. The instruction did not carry the round. A person did. The finding is the two sentences the facilitator supplied, and it will vanish from the record unless someone writes them down while they are still in the air.

Two habits protect against this.

First, distinguish what happened from what it proves. A player who reopens a folded slip has demonstrated that the slip can be reopened. Whether that is a problem depends on what the rule says — which is a separate sentence.

Second, when the operation genuinely cannot continue, name the missing instruction instead of improvising a rescue. An abandoned round is a result. A round that only finished because the facilitator invented a rule on the spot is a disguised abandonment, and it is the more dangerous of the two, because it looks like a success.

One kit, end to end

The round below is invented for this article, as a worked pattern rather than a report. Nobody has played it. There are no timings, no results and no transcripts.

Materials. Fifteen white tokens, held in a supply tray. Three player mats, lettered A, B and C, each dealt five tokens at the start of a round. A pad of small blank slips. A pen per player. A pool tray in the middle of the table, empty at the start of a round. An operator script card. A blank observation sheet per round.

The round as written.

Each player secretly writes how many of their own tokens they want to put in, and folds the slip. Put your folded slip in the middle. When all three slips are in, the operator opens them together, and each player pushes that many tokens off their mat into the pool tray. The highest number takes the pool.

The operator script is five lines long: call the round; ask players to write and fold; wait until three slips are down; turn them over together, read the numbers aloud, and see that each player pushes their tokens into the pool; sweep the pool and the mats back into the supply tray, and deal five tokens to each mat for the next round.

The uncertainty this trial is for. Does a player know, before the reveal, that the number on the slip cannot be changed?

The draft wording has three soft spots, and it is worth being honest that they are soft spots rather than confirmed faults. "Want to put in" reads as a preference, not a commitment. Nothing in the sentence says that a folded slip is closed. And "the highest number takes the pool" leaves a tie undefined.

If you changed all three at once, a smoother session would tell you nothing about which change did the work. So change one.

A proposed revision (not a response to anything observed). Add a single sentence, and leave the rest of the wording alone:

Once your slip has left your hand, the number on it is final. You may not unfold it, rewrite it or take it back.

This is a proposal drawn from reading the draft rules. It has not been tested. The tie rule and the placement wording are separate questions for later rounds, and leaving them alone is the point.

The operator branch. Suppose a player asks, "Can I still change it?" Version 1's script card has no answer for that.

The operator should not invent a rule. Give the shortest answer that lets the round continue, and write it down. On the sheet: the question in the player's words, the answer given, and whether the other two players heard it. That last detail is not fussiness. If the answer was given out loud, two players now hold information the written rule never gave them, and any later result from that round is conditional on it.

The question then becomes a decision for the next trial rather than a patch for this one. Is finality at the fold, at the moment the slip leaves the hand, or at the moment the operator collects it? Each of those is a different rule, and each would need its own round to separate.

If the operator has to write the question down, that note is not the observation record — it is a margin note. The observer transcribes it onto the sheet so the record has one author.

Change one rule, name the next question

At the end of the session, look at the record and choose one of three moves:

Revise the wording. A question or a stalled action points at a sentence that did not carry.
Change the information timing. If the difficulty was about who knew what and when, the wording may be fine and the sequence wrong.
Run nothing further and think. A genuine possibility, and often the honest one when the sheet shows the same ambiguity in two different forms.

Whichever you choose, version the instruction. Call the tested wording v1 and the revision v2, keep both, and put the version on every sheet. In three weeks you will not remember which sentence was in play, and the record will be worthless without it.

Then write the next question before you pack up, in the form the table can answer. Not "is the tie rule working" but "when a tie is announced, does the operator's next sentence tell players what happens to the pool without the facilitator adding anything?"

And be careful about what a pleasant session has established. Three people who understood the round and enjoyed it have shown that three people understood the round and enjoyed it. That is a real finding about the instruction. It is not evidence about audience demand, it is not a fairness finding, and it does not survive contact with the general public. Repeated enjoyment across several sessions moves the needle no further, because the thing being measured is no longer the rule.

One piece of outside work is worth doing before any of this becomes a claim. Have someone who works in formats read the kit and check that the hidden-information mechanism actually survives the move to paper. A rule that depends on a studio device, a screen or a reveal you cannot reproduce with three slips will not be tested by this method — it will be tested by something that looks like it, which is worse.

What you have at the end

The deliverable is thin, and that is correct. A folder containing:

The kit: materials list, rule version, operator script.
The blank observation sheet, with the question the trial is meant to answer written at the top.
A line for the result, still empty.

That empty line is the honest part. A cheap trial is useful when it exposes a defined rule question and records the help it required. It stops being useful the moment a lively evening at a table gets written up as though it had proved something about a show.

Frequently asked questions

What can a tabletop trial actually establish?

It can settle one narrow uncertainty the table can watch, such as whether players know when a written number becomes final or whether an operator can read a script without adding sentences. It cannot show whether anyone would watch, whether the format is fair, safe or kind, or how an audience would respond.

Why separate the operator, facilitator and observer?

The operator follows the script card and reads instructions, the facilitator holds the boundary and fields questions, and the observer writes one record on a blank sheet. If only two people are available, one wears two hats; write down which two, because a facilitator who answers a question and then reports on the round has created an intervention the record must note.

What belongs in the observation sheet, and what should be kept apart?

Record round number and rule version, order of events, anything the operator said off script in the words used, player questions and answers, slips folded or taken back, where the round stopped and why. Keep observer interpretation in a clearly separate column; 'B picked her slip back up' is an observation, while 'B didn't trust the rule' is an interpretation.

Why change only one rule at a time?

If the draft has several soft spots, changing all of them at once makes a smoother session tell you nothing about which change did the work. Leave the tie rule, placement wording or other questions for later rounds so the trial can attribute an effect to one revision.

What should happen if an operator is asked a question the script does not answer?

The operator should not invent a rule. Give the shortest answer that lets the round continue and write it down, including the question in the player's words, the answer given and whether other players heard it. If others heard, later results from that round are conditional; the question then becomes a decision for the next trial.