The Flamingo Award
Most of the people worth electing have never held office. There is no roll call to read on them, so the only checkable evidence is what they have said in public and what can be sourced about their background. The Flamingo Award is what the movement gives for that evidence, and it is the only award with a state meaning we have not looked yet.
What it does
The Flamingo reads a person from their own recorded words and from what can be sourced about them: transcriptions of interviews and appearances where the person is speaking, background research, resume, education, and public commitments, plus the per-area qualification assessments built from them. Long-form video is the evidence base, and a two-hour interview counts for more than a clip, because that is where a person's reasoning becomes checkable. It has no reader for a roll call, an authorship, or a committee act. That code is not switched off; it is not there.
The evidence is sorted into eighteen qualification areas — the named things that make somebody worth electing. Each is read on its own, from individual claims up to the person, then the eighteen are combined at weights published in advance that sum to one hundred. Adding a nineteenth would mean re-pricing the other eighteen, in public.
One resolver then decides which of three states the person is in, and it checks the empty case first. No number at all is unknown. At or above the published threshold is earned, at the percentage computed. Too little known to make an adverse finding is unknown again. Only what is left is not earned.
- Earned renders the full-colour flamingo beside the percentage of the award given. A 36% Flamingo Award is an award, given in the percentage earned.
- Not earned renders a solid dark red fill with no bird: we had enough, we did the arithmetic, and no part of the award was earned.
- Unknown renders a solid gray-brown fill with no bird: we do not know enough to say yet. That is a statement about us, not a claim about the person.
The two blank states are different facts and are never collapsed into one another. Neither ever renders as 0%, because a zero would put a state back onto a scale. A blank carries an em dash, a typed cause in plain words — no videos yet, not assessed yet, they did not discuss this — and the action that would close it. In all three states the page prints how much of the person is known, as "N% known".
The goal it serves
The movement's strategy is elections: get a party's own base behind the strongest challenger before the primary rather than splitting three ways after it. A challenger has no voting record, so the ordinary way a voter sizes one up is party label, name recognition, and whoever bought the most advertising — exactly the levers money in politics already controls. The Flamingo exists so there is something else to look at: the person's own words, against rubrics anyone can read first, published with the quotes that moved the result. See Good New Leaders for who is on that roster and Qualifications for what they are read against.
The unknown state is the second reason. The cheapest way to damage a newcomer nobody has heard of is to publish silence about them and let the reader fill it in. If our own gap in evidence rendered the same as a finding, this would be a smear machine with arithmetic attached. So the gap is typed, named, coloured differently, and pointed at a queue of work.
Third, the unit of judgement stays a person, not a party. Nothing in the eighteen areas asks who somebody caucuses with. A reader who wants the partisan version goes to their own party's door; the assessment underneath is the same either way. See Replacing Captured Incumbents for how it gets used.
What keeps it honest
- The wall is enforced in code, not in style. A build test fails if a voting-record term ever reaches a Flamingo result, and the reverse is equally a defect.
- The red state is not the low end of a scale. The threshold is published in advance, and everything from it upward is conferred at the percentage computed. A middling reading gets an award at that percentage; it does not get the red fill.
- No adverse finding off almost nothing. A floor on how much is known guards the not-earned branch only. It never withholds an award that was computed: the honest answer to a thin record is the percentage with "N% known" beside it.
- Citizen stars sit beside the arithmetic, never inside it. Citizens cast one to five stars on how strong somebody is in a qualification area, and those stars are shown next to the movement's own computed result rather than folded into it, so neither one can quietly move the other.
- The weights are yours to disagree with. The movement publishes its own percentage for each of the eighteen areas, and a citizen can dial every one of them to their own number, with the movement's default shown beside it.
- The coverage gap is published as a finding about us. Every run reports how many people landed on each cause of a blank, and an area nobody has been assessed in is named as a hole in our own rubric authoring.
- An award decides nothing on its own. It does not endorse anyone. It is an input to a citizen vote, never a substitute for one.
Works with
- The Monkey Award — the damage side of the voting record, computed by a separate engine that cannot see any of this evidence.
- The Llama Award — the repair side; someone who has held office can hold a Llama and a Flamingo at once, shown side by side at the same size.
- The Wall — the rule deciding which evidence reaches which award, and why mixing them would make both unreadable.
Where to go next
- The Flamingo Award — the full area: what feeds it, and why not assessed is not a verdict.
- The three states — the exact difference between the red and the gray-brown.
- Qualifications — the eighteen areas and the rubrics, published before anyone is read against them.
- How someone gets on the roster — how a challenger comes to be assessed, and who can put one forward.
- Take Action — adding a video of somebody speaking is what turns an unknown into a known, and you can do it this week.