Prompting technique

Time-Constrained Reasoning (Coarse-to-Fine & Answer-and-Verify)

An Empirical Study of LLM Reasoning Ability Under Strict Output Length Constraint, arXiv:2504.14350 (2025) · read the paper

What it does

A method of structuring prompts for working with output length constraints. Includes pre-informing the model about limits and enforcing final output requirements. Uses two approaches: Coarse-to-Fine and Answer-and-Verify, which allows obtaining quality answers even under strict constraints.

Task typereasoning
Shapechain of thought
Efforttwo or three steps
Addressesthe model losing what was agreed earlier; answers that are close but wrong; output that ignores the requested format
Published2025-04-19
AuthorsYi Sun, Han Wang, Jiaqiang Li et al.

When it helps

Use this technique when working with constraints on model output length (tokens, time, budget) while requiring a quality, structured response. Especially effective in production systems with strict generation limits, in chatbots with fast response times, when solving mathematical and logical tasks where the final answer matters. Suitable for scenarios where the full chain of reasoning might be too long, but you need to guarantee obtaining the final output.

A prompt you can paste

Generic by design: it applies the technique without knowing your task. Adapt the marked parts.

Prompt
You are an AI assistant tasked with generating a response under a strict output length constraint, simulating a time-limited scenario. Your goal is to provide the most valuable information efficiently.

Apply the **"Coarse-to-Fine" (c2f)** and **"Answer-and-Verify" (aav)** strategies, combined with a forced concluding statement, to generate [NUMBER] distinct ideas for [USER'S TASK DESCRIPTION].

**Constraints:**
- The final output must be concise and actionable.
- Prioritize delivering the core ideas and essential justification.
- Do not exceed a total of [MAXIMUM WORD COUNT] words for the entire response.

**Execution Plan:**

**Part 1: Coarse-to-Fine Reasoning (High-Level Outline)**
First, provide a high-level overview of the [NUMBER] ideas. For each idea, include:
1.  **Idea Name:** A brief, descriptive title (2-4 words).
2.  **Core Concept:** A single sentence summarizing the essence of the idea.

**Part 2: Answer-and-Verify (Detailed Justification)**
After presenting the high-level outline, detail each of the [NUMBER] ideas. For each idea, provide:
1.  **Detailed Description:** Elaborate on the core concept (2-3 sentences).
2.  **Key Benefit/Value Proposition:** Explain why this idea is compelling or useful (1-2 sentences).
3.  **Applicability to [USER'S TASK CONTEXT]:** Briefly state how it fits the specific requirements (1 sentence).

**Part 3: Final Conclusion (Forced Summary)**
In the very last part of your response, after all detailed descriptions, write:
"**Final Conclusion:** [Your single most recommended idea or a synthesized takeaway statement]."

**Structure your response with clear headings for each part and numbered lists for ideas.**

If this one does not fit, the two closest alternatives in the corpus are Structured Data Prompting and Blueprint Prompting, which target the same failure from a different angle.

Worked example

The same technique applied to a concrete job: work through a multi-step planning problem. Use it as the pattern for your own case rather than as a finished artefact.

Worked example
You are an expert AI assistant tasked with complex planning and problem-solving under strict constraints. Your goal is to provide a structured, high-quality output that prioritizes essential information, even when faced with limitations.

Apply the **"Coarse-to-Fine" (c2f)** and **"Answer-and-Verify" (aav)** combined strategy, drawing inspiration from the "Times-Up" research on time-constrained reasoning.

## User Task: Multi-Step Planning Problem

**Objective:** Develop a comprehensive plan for launching a new community garden initiative.

**Key Stages of the Plan:**
1.  **Community Outreach & Engagement:** How to inform and involve local residents.
2.  **Site Selection & Preparation:** Criteria for choosing a location and necessary groundwork.
3.  **Resource Acquisition:** Securing funding, tools, seeds, and volunteers.
4.  **Garden Design & Planting:** Layout and initial planting strategy.
5.  **Ongoing Maintenance & Management:** Long-term care and community involvement.

**Constraints:**
*   The final output must be a complete, actionable plan.
*   Prioritize clarity and conciseness in each section.
*   Assume a limited "budget" for detailed explanations, meaning the most critical information should be presented upfront or summarized effectively.

## Execution Plan:

**Phase 1: High-Level Overview (Coarse-Grained Reasoning - "Coarse-to-Fine")**

Provide a concise, bulleted summary of the entire plan. For each of the five key stages listed above, outline the primary objective and 1-2 critical actions. This section should be a complete roadmap, even if details are sparse.

*   **Example Structure for this Phase:**
    *   **Stage 1: Community Outreach & Engagement**
        *   Objective: Build awareness and recruit initial participants.
        *   Critical Actions: Host informational meetings, launch social media campaign.
    *   **Stage 2: Site Selection & Preparation**
        *   Objective: Secure a suitable and ready plot.
        *   Critical Actions: Identify potential sites based on criteria, perform soil testing.
    *   *(Continue for all 5 stages)*

**Phase 2: Detailed Action & Verification (Fine-Grained Reasoning & Answer-and-Verify)**

After providing the high-level overview, elaborate on the **first THREE stages** (Community Outreach, Site Selection, Resource Acquisition) with more detail. For each of these stages:

*   **Detailed Actions:** List specific, actionable steps.
*   **Key Considerations/Challenges:** Identify potential obstacles or important factors to address.
*   **Success Metrics:** Define how to measure the success of this stage.

**Phase 3: Final Conclusion (Forced Conclusion - "Answer-and-Verify" adaptation)**

In the very last part of your response, after all detailed explanations, provide a concluding statement that summarizes the most critical element for the *overall success* of the community garden initiative. Frame this as:

**"Critical Success Factor:"** [Your single most important takeaway for the entire project.]

**Important Note:** Ensure that **Phase 1 (High-Level Overview)** is fully completed and presented first. This guarantees that the essential roadmap is always delivered, regardless of how much detail can be provided in Phase 2.

Get this written for your actual task

Paste what you are trying to do and the corpus will be matched against it directly. Free, no account, about ten seconds.

Free · no signup · ~10s
0.00match confidence
single retrieval pass
Prompt for your task

      

That number is low on purpose, and it is real. It is the raw similarity of one retrieval pass: no specialist read the paper, no judge compared anything, the first plausible match won.

6,235techniques in the corpus
one of which is this page

Picking the right one for a specific task is the work, and it is the work GetDecision does.

This pageone technique, generic prompt
What you just ranone technique matched to your wording, nothing verified
Full runten specialists read the papers in full, a judge ranks the top three for your task and shows its reasoning, generation on the model you pick, saved to your history

See the top three for your taskTen specialists, a judge, and the reasoning shown. Free account, first run included.

Run the full analysis

Related techniques

Structured Data PromptingA method of formatting structured data in a prompt based on the principle of information locality. Related fac…Blueprint PromptingA technique for providing a language model with a pre-prepared structured reasoning plan (Blueprint) for solvi…Layered Context PromptingLayer-by-layer enrichment of the prompt with context with clear division into logical blocks: TASK + FORMAT + …Requirement-Oriented Prompt Engineering (ROPE)ROPE (Requirement-Oriented Prompt Engineering) is an approach to creating prompts as detailed technical specif…

All techniques · Failure modes and fixes