CELPIP Speaking Task 4, Making Predictions: how to predict what happens next

Making Predictions is the Task 3 picture turned into an argument. You are not asked what the picture shows, but what will happen next and why you think so. The common failure is to spend 40 of the 60 seconds describing, then offer a prediction with no reason.

What does CELPIP Speaking Task 4 ask you to do?

You see a picture and say what you think will most probably happen next. You get 30 seconds to prepare and 60 seconds to record, the same clocks as Task 3.

CELPIP's Speaking Pro study pack says Task 4 uses the same illustration as Task 3, so you will already have looked at the image once. Speaking has eight tasks in about 15 minutes. All are recorded replies to on-screen prompts, there is no examiner, and on the official test the screen moves on when the clock ends.

Timings checked 16 October 2026 against the Speaking Pro study pack. The eight-task, 15-minute format is in the CELPIP Guidebook for Test Takers, July 2026.

What do raters look for in a prediction answer?

CELPIP's performance standards use the same four dimensions for every Speaking task. This is our reading of how they apply to predictions, not CELPIP's wording.

• Content/Coherence. The number and quality of ideas, organisation, and examples and supporting details. Here the supporting detail is the clue in the picture that justifies your prediction.
• Vocabulary. Range and precision. Modal and probability language, such as "is likely to", "might", "is bound to", gives you accurate ways to be uncertain.
• Listenability. Rhythm, pronunciation and intonation; pauses, interjections and self-correction; grammar and sentence structure. Future and conditional forms are where errors appear.
• Task Fulfillment. Relevance, completeness, tone and length. Relevance means predicting from this picture. Completeness means giving a prediction and the reasons, not only a description.

See how CELPIP-General is scored for the full criteria.

How do you use the 30 seconds of preparation?

You need a prediction and two clues. Everything else can be improvised.

1. Seconds 0 to 8: find the moment. What is happening right now, and what is about to change? Look for a person, an object and something unstable or unfinished.
2. Seconds 8 to 18: choose the likeliest outcome. Pick the prediction the picture supports best, not the most dramatic.
3. Seconds 18 to 26: pull two clues. Each clue is a visible fact: a position, an object, an expression. These are your reasons.
4. Seconds 26 to 30: think of an alternative. One short "it's also possible that" gives you a way to fill remaining time without repeating yourself.

If the picture shows two people and a situation, the prediction usually concerns what one of them does next, so decide who your answer is about.

How do you structure a 60-second prediction?

This is ScoreFluent's own technique, with approximate timings.

• Situation, about 10 seconds. One or two sentences setting the scene. Only what the listener needs.
• Main prediction with two clues, about 25 seconds. "I think the most likely thing is that…, because…, and also…".
• What follows, about 15 seconds. The consequence of the prediction. This extends the answer without adding a new topic.
• Alternative and close, about 10 seconds. "Of course, it's also possible that… but I think… is more likely."

Use "will probably" and "is likely to" for your main prediction. Reserve "might" and "could" for the alternative, so your stance is clear.

Worked example

This example is built from the text of a bank item (item 3 below), not the image: a man is in the middle of painting a house while a dog wanders over, and the picture shows a man on a ladder painting a house wall while a dog sniffs a paint can. In your real answer, use the clues you actually see.

In this picture, a man is up on a ladder painting the wall of a house, and a dog has wandered over and is sniffing a paint can. I think the most likely thing is that the dog will knock the can over. The dog is clearly interested in the paint, and the man is busy with the wall, so he probably won't notice in time. If the paint spills, it will probably run across the ground, and the dog might step in it and leave paw prints across the path. The man will most likely climb down quickly, a bit annoyed, and spend the next while cleaning up instead of painting. Of course, it's also possible that he sees the dog in time and calls it away. But I'd say the spill is more likely, and next time he'll probably keep the paint out of reach.

This sample was written by ScoreFluent for this guide. It is not an official CELPIP model answer and has not been scored. The first two sentences describe, then the prediction arrives. Each prediction has a reason ("because" is implied by "clearly interested" and "busy with the wall"), the consequence extends the story, and the alternative appears late and briefly.

What mistakes make prediction answers weaker?

• Describing for too long. Forty seconds of description leaves too little for the thing you were asked.
• A prediction with no reason. "The dog will spill the paint" is a claim. "The dog is sniffing the can and the man isn't watching" is a reason.
• Predictions the picture does not support. An answer about a storm that nothing in the picture suggests has no evidence behind it.
• Over-hedging. If every sentence says "maybe", your stance disappears. Commit to one main prediction.
• Present tense for the future. "The dog spills the paint" in the middle of a prediction is a grammar slip. Use "will", "is going to" or "is likely to".
• Repeating the same reason. Saying "the dog is curious" three ways is one reason. Find a second clue.

What do Task 4 practice items look like?

These are practice items from the ScoreFluent bank. They are not real exam questions. In ScoreFluent the picture itself is shown with each item. Question images are not public files, so this page gives the bank's situation text and its own one-line text summary of the picture. They are listed from easier to harder on the bank's difficulty rating.

Item 1. The child is standing at the top of a snowy hill with a sled.

Picture summary: A child in a red coat with a sled at the top of a snowy hill.

Task: Predict what may happen next in this situation and explain why you think so.

Item 2. Two hikers have reached a stream on their trail.

Picture summary: Two hikers at a log bridge over a stream.

Task: Predict what may happen next in this situation and explain why you think so.

Item 3. The man is in the middle of painting the house while a dog wanders over.

Picture summary: A man on a ladder painting a house wall while a dog sniffs a paint can.

Task: Predict what may happen next in this situation and explain why you think so.

Item 4. The campers are setting up for the evening.

Picture summary: Two campers beside an orange tent and a pile of firewood.

Task: Predict what may happen next in this situation and explain why you think so.

Item 5. The baker is serving a customer while a child waits nearby.

Picture summary: A baker handing a bag to a customer while a child reaches for a cookie.

Task: Predict what may happen next in this situation and explain why you think so.

The bank does not include model answers for these items. Practise from the situation line and summary: name the likeliest outcome, find two clues the line supports, add a consequence, and then check your prediction against the real picture in the app.

How do you practise Task 4?

1. Predict from still images. Pick a news or advertising photo, set a 30-second timer, and make a prediction with two clues. Then speak for 60 seconds.
2. Practise the language. Write five ways to say "probably" and five ways to say "possibly", then use two of each in the next answer.
3. Cap the description. Set a rule that the prediction must start before the 12-second mark.
4. Pair Task 3 and Task 4. Describe a picture for 60 seconds, then predict from the same picture for 60. This mirrors the structure CELPIP's Speaking Pro study pack describes and shows the difference between the two tasks.

On ScoreFluent you can record Task 4 against the real timings and get a practice estimate on the four published dimensions. That estimate is provisional: the CELPIP speaking grader has not been calibrated against human-rated CELPIP responses, so use it to see what to fix, not as a level you will receive. New accounts get 40 free credits, credits never expire, and there is no subscription. Start with 40 free credits →

Read Task 3, Describing a Scene for the description half of the pair, and Task 7, Expressing Opinions for another task where a reason has to follow every claim. The whole section is covered in CELPIP Speaking, and the exam overview is on the CELPIP-General hub.

Where do these timings and criteria come from?

The task name and the 15-minute, eight-task Speaking format come from the CELPIP Guidebook for Test Takers, July 2026. The 30-second preparation, 60-second recording, shared picture and move-on rule come from the Speaking Pro study pack. The four Speaking dimensions are from CELPIP's performance standards. Checked 16 October 2026. Exam details change, so confirm against the official pages before you rely on a number.

ScoreFluent is an independent practice platform. It is not affiliated with, endorsed by or connected to Paragon Testing Enterprises or CELPIP.