CELPIP Speaking Task 4 Template — Making Predictions

Speaking Task 4 is the second of the two picture-based tasks on the CELPIP Speaking test, and it asks you to do one thing well: predict what will happen next in the scene you just described in Task 3. You get 30 seconds to prepare and 60 seconds to speak into the microphone — there is no live examiner. This guide gives you a repeatable structure, two reusable language banks (future forms and prediction starters), a 60-second model answer, and a clear breakdown of what separates a CLB 7 response from a CLB 11 one.

The key thing to understand is that Task 3 and Task 4 share the same picture. In Task 3 you described what is happening now; in Task 4 you look at that same image and say what is likely to happen in the moments, hours, or days that follow. The examiner is not testing whether your guesses are "correct" — there is no right answer. They are testing whether you can produce confident, grounded predictions using a natural range of future-tense grammar.

What the examiner is looking for

CELPIP Speaking is scored on four criteria. Here is what each means for Task 4 and how to hit it.

• Content & Coherence — Your predictions should be logically connected to the picture and to each other. Hit it by always pairing a prediction with a reason drawn from the scene ("because the sky is darkening, they'll probably pack up soon"). Order your ideas so one flows into the next.
• Vocabulary — You need precise, varied language, not just "they will go home." Hit it by using specific verbs, descriptive adjectives, and a range of certainty words (likely, probably, almost certainly, there's a chance).
• Listenability — This is the spoken equivalent of readability: clear pronunciation, smooth pacing, natural stress, and few distracting fillers. Hit it by speaking at a steady pace, grouping words into phrases, and using transitions instead of long silences.
• Task Fulfilment — You must actually make predictions for the full 60 seconds, not slip back into describing the present. Hit it by using future forms throughout and covering two or three distinct outcomes with reasons.

The 30-second prep: the "Predict the Rule of 3"

Thirty seconds is short, so do not script full sentences. Instead, jot three quick prediction triggers. The most reliable structure is to make three predictions, each with a reason, framed by an opening and a closing line.

Step 1 — Open with a confident lead prediction

Start immediately with your strongest, most obvious prediction so the listener knows you are on task. Anchor it to something visible. For a picture of a family at a beach with grey clouds, you might open: "Looking at this scene, I think the family is going to leave the beach quite soon." This signals task fulfilment in the first five seconds and buys you momentum.

Step 2 — Give the reason from the scene

Never leave a prediction floating. Immediately justify it with evidence from the image, which is what lifts Content & Coherence. For example: "I say this because the sky is getting darker and a few people in the background are already folding up their umbrellas." Linking cause to effect shows the examiner you are reasoning, not guessing randomly.

Step 3 — Add a second prediction with a different time frame

Vary the time scale to show grammatical range. Your first prediction might be about the next few minutes; your second could reach further out: "After they pack up, they'll most likely head to a nearby café to wait out the rain." Using "after" and a sequencing word demonstrates you can chain future events together.

Step 4 — Offer a less certain or alternative outcome

Strong responses acknowledge uncertainty, which sounds natural and shows off modal verbs. Say something like: "Of course, the storm might pass quickly, in which case they could decide to stay and enjoy the rest of the afternoon." This single sentence adds might, could, and an alternative outcome.

Step 5 — Close with a summary prediction

Wrap up so you do not trail off into silence when the timer ends. A short closing prediction works: "Either way, I'm fairly sure they'll remember this as an eventful day at the beach." A clean ending boosts Listenability and signals control of the full minute.

The fill-in template

Use this skeleton in your head during prep. Fill the brackets with details from your specific picture.

OPEN:   Looking at this picture, I think [person/people] is/are going to [main prediction] quite soon.
REASON: I say this because [visible evidence from the scene], so it's likely that [consequence].
NEXT:   After that, they'll probably [second prediction], since [reason / context].
MAYBE:  Of course, [alternative possibility] might happen instead, in which case they could [different outcome].
CLOSE:  Either way, I'm fairly confident that [summary prediction about how it ends].

Prediction sentence-starter bank

Mix these so you do not repeat "I think... I think..." Group them by function and pull one from each row.

Future forms & certainty language

The single biggest score-lifter in Task 4 is using a range of future grammar rather than "will" ten times. This table shows each form, when to use it, and a ready example.

Aim to use at least four different forms from this table across your 60 seconds. That single habit is often the difference between a CLB 7 and a CLB 9.

Model answer

This sample responds to a common Task 4 image type: a family setting up a picnic in a park while dark clouds gather overhead. It runs about 125 words — a comfortable 60 seconds.

Why this answer scores high

• Opening line — It states a clear main prediction ("going to have to cut their picnic short") in the first sentence, nailing Task Fulfilment immediately and using be going to based on visible evidence.
• Second sentence — "I say this because…" ties the prediction directly to two scene details (darkening sky, wind on the blanket). That grounded reasoning is exactly what Content & Coherence rewards.
• Third and fourth sentences — They sequence events with "So it's very likely…" and "After that, they'll probably…," showing the speaker can chain future actions and vary certainty (very likely, probably).
• Fifth sentence — "Of course, the storm might blow over… they could decide to…" introduces an alternative outcome with two new modals (might, could), adding grammatical range.
• Closing — "I'm fairly confident… is about to become…" delivers a tidy summary prediction and slips in the is about to form. The varied vocabulary (scramble, restless, blow over, wait it out) lifts the Vocabulary score, and the smooth transitions support Listenability.

Across the response the speaker uses six different future forms — going to, will, it's very likely, probably, might, could, is about to — without ever sounding mechanical.

What CLB 7, 9 and 11 look like

• CLB 7 — Makes predictions and stays on task, but relies heavily on "will" and "going to." Reasons are present but brief ("because it's raining"). Some hesitation and simple vocabulary; the listener follows easily but the range is limited.
• CLB 9 — Uses three or more future forms naturally, pairs every prediction with a specific reason from the scene, and includes at least one alternative outcome with might/could. Pacing is smooth with few fillers, and vocabulary is noticeably more precise.
• CLB 11 — Sounds effortless and confident, weaving four or more future forms, hedged certainty ("almost certainly," "there's an outside chance"), and vivid, idiomatic vocabulary. Predictions are layered across different time frames and connected with sophisticated transitions, all delivered with natural stress and rhythm.

Common mistakes to avoid

• Describing instead of predicting — Slipping back into present tense ("The family is sitting on a blanket") wastes seconds and hurts Task Fulfilment. Keep your verbs in the future.
• Using only "will" — Repeating one form flattens your Vocabulary and grammar scores. Force at least one going to, one might/could, and one likely to.
• Predictions with no reasons — A list of guesses with nothing from the picture reads as random. Always answer the silent "why?"
• Running out of things to say — If you finish early and freeze, the silence hurts Listenability. Have your closing summary ready, and use the alternative-outcome line to extend.
• Over-hedging — Saying "maybe, possibly, perhaps, I'm not sure" in every sentence sounds unconfident. Lead with strong predictions; reserve uncertainty for one or two ideas.
• Rushing or going silent at the start — Spend your 30 seconds planning three triggers, then open immediately. A confident first sentence sets the tone for the whole minute.

Buying time & staying fluent

Natural fillers keep you fluent without dead air while your brain forms the next prediction. Sprinkle these in — but lightly.

Keep these to one or two per response. Used well, they make you sound thoughtful and in control; overused, they read as hesitation. With the structure, banks, and model above, you have everything you need to turn the same picture from Task 3 into a confident, high-scoring minute of predictions.