Prompt Battle.

Test your prompt engineering skills. Score points for accuracy and prove your mastery.

This battle covers

Choosing the right prompt technique for each situation
Understanding temperature and its effects
System prompts and how they shape behavior
When to use zero-shot vs few-shot vs chain-of-thought

Strategy Guide

The prompt engineer's decision tree.

Before you battle, let's build your mental toolkit. Prompt engineering is about choosing the right strategy for the right situation — like picking the right tool from a toolbox. A hammer is great for nails, terrible for screws. Same with prompt techniques.

Zero-Shot — when the task speaks for itself

Use zero-shot when the task is common and well-defined. Classifying sentiment, translating text, summarizing a paragraph — the AI already knows how to do these. No examples needed. Just give a clear, specific instruction. Think of it as asking a skilled colleague to do something they already know how to do: "Summarize this in 3 bullet points." Simple. Direct. Effective.

Few-Shot — when you need a specific pattern

Use few-shot when you need the AI to follow a format it would not guess on its own. Showing 2-3 examples of input-output pairs teaches the pattern better than describing it. It is like training a new employee: instead of writing a 10-page manual, you show them three completed examples and say "do it like this." The AI picks up the pattern from your demonstrations.

Chain-of-Thought — when accuracy matters most

Use chain-of-thought for math, logic, debugging, and any multi-step reasoning. The magic phrase "think step by step" forces the AI to show its work — and showing work dramatically reduces errors. Research shows up to 2x accuracy improvement. It is like asking a student to show their math work: the process catches mistakes that a rushed final answer would miss.

Role-Play — when you need an expert voice

Use role-play when you want specialized expertise or a specific tone. "You are a senior security engineer" produces different output than "check this code." The persona shapes word choice, focus areas, and depth. It is like the difference between asking your friend about a legal issue versus asking a lawyer — same question, vastly different quality of answer.

Here is the decision tree. Follow it before every prompt you write:

  PROMPT TECHNIQUE DECISION TREE

  Is the task simple and well-defined?
  ├── YES → Zero-Shot (just ask clearly)
  └── NO → Does it need a specific output format?
       ├── YES → Few-Shot (show 2-3 examples)
       └── NO → Does it need reasoning or accuracy?
            ├── YES → Chain-of-Thought (think step by step)
            └── NO → Does it need expertise or tone?
                 ├── YES → Role-Play (set a persona)
                 └── NO → Zero-Shot (default)

  PRO TIP: You can combine techniques!
  Role-Play + Chain-of-Thought = expert reasoning
  Few-Shot + Role-Play = patterned expert output

Temperature — the creativity dial

Temperature is a number from 0 to 1 that controls randomness. At 0, the AI always picks the most likely next word — deterministic, consistent, factual. At 1, it sometimes picks less likely words — creative, surprising, but error-prone. Think of it as a dial between "accountant mode" and "poet mode." Code and facts want low temperature. Brainstorming and creative writing want high temperature.

System prompts — the invisible rulebook

A system prompt is a hidden message processed before any user input. The user never sees it, but it shapes every response. Think of it as giving an actor their character brief before the show starts. System prompts define persona, rules, constraints, and tone. Every major AI product uses them — ChatGPT, Claude, Copilot all have system prompts running behind the scenes.

Now you have the playbook. The battle below will test whether you can pick the right technique for each situation. Remember: there is usually one best answer, but combining techniques is a sign of mastery.

The Battle

Answer fast, answer right.

Prompt Example — Zero-Shot vs Few-Shot vs Chain-of-Thought

# ZERO-SHOT (no examples — just the task)
Classify this review as POSITIVE, NEGATIVE, or NEUTRAL:
"The battery lasts forever but the screen is dim."

# FEW-SHOT (teach by example)
Review: "Absolutely love it!" → POSITIVE
Review: "Broke after one day." → NEGATIVE
Review: "It's fine, nothing special." → NEUTRAL
Review: "The battery lasts forever but the screen is dim." → ???

# CHAIN-OF-THOUGHT (force step-by-step reasoning)
Classify this review as POSITIVE, NEGATIVE, or NEUTRAL.
Think step by step before answering:
"The battery lasts forever but the screen is dim."

Step 1: "battery lasts forever" = strong positive.
Step 2: "screen is dim" = moderate negative.
Step 3: Mixed sentiment, but positive outweighs.
Answer: POSITIVE

Technique Mastery

Match the technique to the task.

Advanced Moves

Combine techniques for maximum power.

The real masters do not use one technique at a time — they layer them. Here are the most effective combinations:

Role-Play + Chain-of-Thought

"You are a senior financial analyst. Analyze this quarterly report step by step before giving your recommendation." The role sets the expertise level and vocabulary. The chain-of-thought forces rigorous reasoning. Together they produce expert-quality analysis with visible logic.

Few-Shot + Constraints

"Here are 3 examples of how we format support tickets. Now format this one. Never include the customer's email address in the summary." The examples teach the pattern. The constraints add guardrails. This is the standard pattern for production AI applications where consistency and safety matter.

System Prompt + Few-Shot + Temperature

Set the persona in the system prompt, provide examples in the user message, and tune temperature for the task. This triple-layer approach is what powers most commercial AI products. The system prompt sets the foundation, examples calibrate the format, and temperature controls the creativity level.

The best prompts are designed, not written. Think of each technique as a tool in your belt. Simple tasks need one tool. Complex tasks need a combination. The quiz below tests whether you can pick the right tool — or the right combination — for each situation.

Final Challenge

Collect the prompt engineering knowledge.

Remember

The prompt engineer's cheat sheet.

  PROMPT ENGINEERING CHEAT SHEET

  Technique        Best For                   Key Phrase
  ─────────        ────────                   ──────────
  Zero-Shot        Simple, clear tasks        "Classify this as..."
  Few-Shot         Custom formats/styles      "Here are 3 examples..."
  Chain-of-Thought Math, logic, debugging     "Think step by step..."
  Role-Play        Expert voice/persona       "You are a senior..."

  TEMPERATURE GUIDE
  0.0 - 0.2  Code, math, facts, data extraction
  0.3 - 0.6  Business writing, explanations, general use
  0.7 - 1.0  Brainstorming, creative writing, idea generation

  POWER COMBOS
  Role + CoT        Expert-level reasoning with visible logic
  Few-Shot + Guard  Consistent format with safety constraints
  System + Examples Production-ready AI applications

Prompt engineering is a skill, not a talent. The more you practice, the better your instincts become. Save this cheat sheet and reference it every time you write a prompt until the patterns become second nature.

Next: Words as Numbers →

🔒

This lesson is for Pro members

Unlock all 520+ lessons across 52 courses with Academy Pro.

Go Pro — $49/mo ← Back to course

Already a member? Sign in to access your lessons.