Using DebateFlow
Speaker Points
Speaker points are the number new judges ask about most, and the one most likely to go wrong. This page explains the scale, what the app suggests, and what it deliberately refuses to look at.
The scale
Public Forum speaker points run from 25.0 to 30.0, in half points. That gives you eleven choices: 25.0, 25.5, 26.0, and so on up to 30.0.
Points are an individual measure. They are separate from the win: you can give the losing team's second speaker the highest points in the round, and judges regularly do. The win says which team's arguments carried the round. The points say what each of the four speakers did.
25.0 is the floor. The app will not accept a number below it or above 30.0, and it will not accept a number that is not on a half point.
What each band means
This is the rubric the app shows you, word for word. Each band describes what a speaker did with arguments, not how polished they were.
| Points | What it describes |
|---|---|
| 30.0 | Complete arguments, extended with new reasons, answered everything that mattered, and weighed clearly. Rare. |
| 29.0-29.5 | Complete arguments and clear weighing; one small gap. |
| 28.0-28.5 | Solid arguments; some extensions were name-checks, or one important response was missing. |
| 27.0-27.5 | Claims present but thin warrants or impacts; weighing implied rather than stated. |
| 26.0-26.5 | Several arguments incomplete or dropped; little weighing. |
| 25.0-25.5 | Reserved for conduct problems or a speech that did not engage the round. |
Three words in that table come from debate, and it helps to have them: a warrant is the reason an argument is true, an impact is why it matters, and weighing is saying why your argument matters more than the other side's. The glossary has the rest.
The bottom band is narrower than it looks. 25.0-25.5 is not for a nervous speaker or one you found hard to follow. It is for conduct problems, or a speech that did not engage the round at all.
Where the suggestion comes from
On the last step of the decision wizard, each speaker gets a suggested number and a short list of the factors behind it. The list is the point: you can see what the number rests on and disagree with any part of it.
Every speaker starts at 27.0 - the "claims present, but thin" band - and moves from there on what the round recorded about their own speeches:
- Complete arguments. How many of the arguments they made had a claim, a reason and an impact, rather than only a claim.
- Extensions. Whether their side developed an argument in a later speech with new reasoning, or only named it again. Naming it again counts against them; that is the standard debaters are coached to.
- How much of the other side they answered. Out of the opposing arguments in front of them, how many their speeches engaged.
- How those answers landed. Whether the analysis rated their responses strong or weak.
- Weighing. Whether they compared their argument against the other side's, or asserted it and stopped.
There is one delivery input, and it is yours: if you answered No to "could you follow this speech?" after a speech, that takes half a point off. It only ever moves the number down, because you cannot give credit for an argument you could not follow. It never works in reverse - nobody gains points for sounding clear.
Each speaker's speeches are attributed from your own speaker picks during the round. If you did not say who was speaking, the round has nothing to attribute.
What it will not look at
The suggestion never reads how a student sounded: not their accent, not their confidence, not their volume, not their poise, not whether they seemed credible.
This is not squeamishness. Speaker points are a documented place where bias enters a ballot, and it lands on the same students every time - students with accents, students who are nervous, students who do not perform confidence the way a judge expects. An earlier version of DebateFlow built its suggestion from two sliders, clarity and persuasiveness, which is a measure of how a student came across to one listener. Two speakers whose arguments were nothing alike got the same points whenever they came across the same way.
So the suggestion reads the round instead. You are still free to weigh delivery in the number you assign - that is your judgement to make - but the app will not do it for you, and the ballot carries a reminder to check that your points rest on the arguments and how well they were explained.
The check worth running
Before you finalise, read your four numbers back and ask what each one rests on. If the answer is "they sounded more sure of themselves", that is the moment to look again at what the two speakers actually argued.
What the suggestion will not do
It never suggests 30.0. The highest it will go is 29.0. A 30 is a judgement that a speech was close to flawless, and that is yours to give by hand.
It never goes below 26.0. The bottom band is reserved for conduct problems or a speech that did not engage the round, and neither of those is something the app can see. It will not put a student at the bottom of the scale on argument quality alone.
It says "no suggestion" rather than guessing. If the round recorded no points or answers for a speaker - because a speech was not recorded, or because no speaker was picked for it - you get "No suggestion" instead of a number. A default like 27.0 for a speaker nobody recorded anything about reads as a judgement, and judges anchor to numbers they are shown.
"No suggestion yet" means something different. The suggestion is built from the app's read of the round, which runs once the round ends and takes a minute or two. Before it finishes, every speaker says "No suggestion yet - this round has not been analysed". That is the app not being ready, not a comment on the speakers. Score from your notes, or come back to the round from your dashboard once it has finished.
You assign the points
Nothing is filled in for you. The suggestion sits beside the box; the number in the box is the one you choose. You can leave a speaker unscored and come back, and you can ignore the suggestion entirely.
This is deliberate. An earlier version quietly pre-filled the boxes with its own number, which made the points the app's rather than yours. Points on your ballot follow a student through the season, and they should be yours.
Ties within a team
If you give both speakers on the same team identical points, you get a warning: many tournaments do not accept tied speaker points within a team, and the suggestion is to separate them by at least half a point.
It is a warning, not a block. You can submit tied points if that is what you saw. Check your tournament's own rules if you are unsure - this is a tab room convention, not a rule the app enforces.
What's next?
- Making Your Decision - where you assign the points, on the last step
- Voting Issues - how the win is decided, separately from the points
- Debate terms, explained - warrant, impact, extend, weigh, and the rest
- Quick Reference - the checklist to print, and where to look things up mid-round