What raters reward
Logical predictions grounded in the picture (Content/Coherence, Task Fulfillment), accurate future forms (will, going to, about to), and fluent delivery (Listenability). Each prediction needs a reason tied to what's visible.
Your time plan
How to structure it
Fill the [slots] with your own ideas — adapt the frames, don't recite them.
Ground the listener before you look ahead.
Grammar Present continuous
- Right now, [what's happening].
- In this scene, [person] is about to [action].
Tie the prediction to a visible clue.
Grammar Future (will / going to / about to) + because
- In a moment, [person] will probably [event] because [clue].
- Judging by [clue], I expect [prediction].
Move the story forward — what happens after that.
Grammar Future + sequence (after that, then)
- After that, [next event] is likely to happen.
- This will most likely lead to [consequence].
End with the likely outcome.
Grammar Summary / future perfect
- By the end, [outcome].
- So overall, I think [final prediction].
A worked model answer
A high-scoring sample that follows this shape — use it as a model, not a script.
Looking ahead, the quietude of this hall might shift as more patrons arrive for an afternoon workshop. The children on the central carpet could soon finish their story and transition to a hands-on craft activity, which would likely increase the noise level. We can expect the young woman at the desk to complete her research project, at which point she might return her borrowed books to the circulation counter. The elderly man could decide to trade his newspaper for a biography found on the nearby shelves. Furthermore, as the sun moves across the sky, the library staff would probably adjust the lighting to maintain a comfortable reading environment. Eventually, as closing time approaches, the librarians may begin tidying the workstations, and the diverse group of visitors would slowly depart, leaving the stacks in stillness.
Useful vocabulary
Vocabulary is one of the four scored dimensions — weave a few in (don't force all of them).
Common mistakes to avoid
- Describing the scene (that's Task 3) instead of predicting.
- Predictions with no reason from the image.
- Avoiding future forms.
- Only one prediction stretched thin.
Quick tips
- Use future language: will, going to, likely to, about to.
- Anchor every prediction to something visible.
- Two justified predictions beat many vague guesses.
