2 ChatGPT Prompts to Analyze HT Under/Over 1.5 From Match Screenshots

A screenshot of football fixtures or betting odds can look full of useful information, but it often leaves out the most important question: how often do these teams score and concede before half-time? ChatGPT can inspect uploaded screenshots and help organize match data, but it can also misread small text or make mistakes. The best prompt must tell it when not to make a pick.

In the HT Under/Over 1.5 goals market, Under 1.5 means zero or one goal in the first half; Over 1.5 means at least two first-half goals. Second-half goals do not count toward this market.

These two prompts are designed for different tasks: the first screens a screenshot with several upcoming matches; the second takes a closer look at one selected match with a more detailed statistics screenshot. Neither can guarantee a winning bet. Their main value is making the analysis more structured and stopping ChatGPT from inventing missing evidence.

Prompt 1: Screen several matches from one screenshot

Use this when your screenshot contains a list of fixtures, odds or a match-selection page. For better results, upload a clear image and include first-half statistics if you have them.

🟦 COPY PROMPT 1 — MULTI-MATCH SCREENING

Act as a cautious football data analyst. Analyze the football matches visible in the screenshot I uploaded and assess the first-half total goals market: HT Under 1.5 versus HT Over 1.5. Your job is to screen matches using evidence, not to force betting tips or promise winners.

STEP 1 — READ THE SCREENSHOT
Transcribe every match you can read, including teams, competition, match date/time, current or scheduled status, and any visible odds or statistics. Do not guess blurred names, missing numbers, league names, or odds. Mark unreadable fields as UNKNOWN and ask for a clearer crop if they matter.

STEP 2 — VERIFY RELEVANT EVIDENCE
If web search is available, confirm the exact fixture and date, then look for trustworthy, recent data for each team. Prioritize first-half-specific evidence from roughly the last 8–12 matches when available:
- First-half scores and total goals at half-time.
- Percentage of first halves ending Under 1.5 and Over 1.5.
- First-half goals scored and conceded by each team.
- Relevant home/away splits and the size of each sample.
- Current team news or lineup information only if it can be verified.

Cite the sources and state when you checked them. Give recent head-to-head results less weight than broader first-half records. Do not use full-time over/under percentages as a substitute for first-half statistics.

If browsing is unavailable, use only information genuinely visible in the screenshot. If it shows only team names or odds and not enough relevant first-half evidence, do not invent statistics—return PASS.

STEP 3 — APPLY STRICT DECISION RULES
For every match, assess both sides separately. A YES means the available evidence is strong enough to consider that side, not that it is certain to win. Never mark both Under 1.5 and Over 1.5 as YES for the same match. If the evidence is weak, conflicting, stale, unreadable, or insufficient, mark both NO and choose PASS / INSUFFICIENT DATA. Do not force a pick simply because I asked for an answer.

Do not manufacture exact win probabilities or confidence percentages. If you calculate a historical rate, show the numerator, denominator, and the matches counted. Explain uncertainty and avoid treating a small sample as proof.

OUTPUT — USE THIS FORMAT FOR EACH MATCH
1. Match and kickoff time:
2. Screenshot data readable? YES / NO
3. HT UNDER 1.5: YES / NO
4. HT OVER 1.5: YES / NO
5. Final decision: UNDER 1.5 / OVER 1.5 / PASS
6. Strongest supporting evidence:
7. Important missing or conflicting information:
8. Sources and check time, if web search was used:

A YES recommendation is optional. A well-supported PASS is better than a guess. Do not tell me to stake a particular amount or describe any result as guaranteed.

Why this prompt helps: It requires first-half data instead of relying on general full-time goal trends. It also prevents ChatGPT from inventing statistics when a screenshot only shows fixtures and odds.

Prompt 2: Deep-check one match using a statistics screenshot

Use this when you have a more detailed screenshot showing form, half-time results, team statistics or odds. If it is a live match, the screenshot must clearly show the current minute and score; live statistics can be delayed, so refresh the original source before acting.

🟩 COPY PROMPT 2 — SINGLE-MATCH DEEP CHECK

Act as a skeptical football analyst reviewing one match for HT Under 1.5 or HT Over 1.5 goals. Analyze the screenshot I upload, but do not treat an AI opinion as certainty. I want a careful YES / NO decision for each side, with PASS whenever the evidence does not support a responsible conclusion.

A. FIRST, READ AND CONFIRM
Identify the teams, competition, date, match status, match clock if live, score if live, visible odds, and every readable statistic. List any unclear or cropped values as UNKNOWN. Do not silently fill gaps. If you cannot confidently identify the match or read the key numbers, stop and ask for a clearer screenshot.

B. CHECK THE RIGHT STATISTICS
If web search is available, confirm the fixture and date and research current, reliable sources. Focus on:
- Each team's first-half goal totals in recent competitive matches, ideally 8–12 matches if available.
- Under 1.5 and Over 1.5 first-half frequencies, with the number of matches behind each rate.
- Home team's home first-half record and away team's away first-half record when sample sizes are meaningful.
- First-half goals scored and conceded, not just full-time goals.
- Verified lineup news, key absences, and match context only when reliable sources support them.
- If the screenshot is LIVE: the exact current minute and score, goals already scored, and visible relevant live indicators. Treat live stats and odds as potentially delayed and state their timestamp where possible.

Do not invent data or rely on possession, corners, team reputation, or a short winning streak alone. Head-to-head records should be secondary and should not outweigh a larger, relevant sample. If the sources disagree, explain the disagreement. If browsing is not available, clearly say that the analysis is limited to the screenshot.

C. MAKE A CONSERVATIVE DECISION
For a pre-match screenshot, assess whether the available first-half evidence leans toward Under 1.5, Over 1.5, or neither. For a live screenshot, first check whether the first half is still in progress; account for the current score and time remaining, but do not pretend to know how much stoppage time will be added. If half-time has already ended, return NO BET / TOO LATE for this first-half prediction.

If two goals have already been scored in the first half, state that the Over 1.5 condition has already been reached rather than presenting it as a fresh prediction. Do not encourage a bet on a market that may already be settled or unavailable.

Never mark both sides YES. Use YES only when the evidence is relevant, readable, current enough, and reasonably consistent. If evidence is insufficient, data is missing, the match is hard to verify, or the conclusion depends on guessing, mark both sides NO and return PASS. Do not fabricate a probability or call a selection a sure bet.

REQUIRED OUTPUT
- Match / competition / date:
- Pre-match or live; screenshot time if visible:
- Key facts read from screenshot (mark uncertain facts):
- First-half statistics verified and sample sizes:
- HT UNDER 1.5: YES / NO — brief evidence:
- HT OVER 1.5: YES / NO — brief evidence:
- Final decision: UNDER 1.5 / OVER 1.5 / PASS:
- Main uncertainty and information still needed:
- Sources and time checked:

If neither side has enough support, the final answer must be PASS. Do not give staking instructions, claim that a result is guaranteed, or present this as financial advice.

Why this prompt helps: It distinguishes pre-match analysis from live analysis, requires the current score and clock for live screenshots, and prevents a confident-looking answer when the image is unreadable or the data is stale.

Tips to improve the quality of the analysis

1. Upload a clear screenshot. Include the match names, league, date and any first-half stats. Crop away unrelated content, but do not crop out the labels or units.
2. Check ChatGPT's read-back. Verify team names, odds, match times and statistics against the original source before using the analysis. OpenAI warns that unclear image inputs may be interpreted less accurately.
3. Use first-half numbers. Full-match goal averages do not directly tell you whether a game will have two goals before half-time.
4. Respect the pass. If only fixtures and odds are visible, the prompt should not manufacture the missing stats. A clear PASS is more useful than a forced pick.
5. Remember the risk. These prompts cannot guarantee accuracy or profit. Do not gamble money needed for essentials, chase losses, or use betting as an income plan.

The bottom line: A strong prompt cannot turn a screenshot into certainty. Its job is to extract the information accurately, verify relevant first-half evidence, show where that evidence is weak, and return YES / NO / PASS without inventing data.

Sources

• OpenAI: ChatGPT Image Inputs FAQ
• Sky Bet: Football in-play rules — HT Under/Over goals
• Responsible Gambling Council: Plan Before You Play