top of page
Footballer advancing with the ball in a floodlit stadium, illustrating first-half goals research

Two First-Half Goals Sounded Plausible. Our 2,007 Bets Lost 123 Units.

A first half with two or more goals is a familiar football betting proposition. It also sounds straightforward: identify a match likely to start quickly, back Over 1.5 first-half goals and wait. Our collector records tell a less comfortable story. Across 2,007 settled selections between 3 October and 8 November 2026, a standardised one-unit BACK calculation returned −123.49 units, or −6.15% on stakes, after 2% commission on winning profits.

This story is part of our Research Lab series. We place small live bets across betting markets to collect real-world results and see whether an appealing idea stands up to scrutiny. You can explore the wider Live Experiments programme, but the important question here is what these 2,007 selections actually did—not what we hoped they might do.

Footballer advancing with the ball in a floodlit stadium, illustrating first-half goals research

What the 2,007 bets actually delivered

The collector recorded 768 winning selections and 1,239 losses: a 38.27% strike rate. With a one-unit stake on every settled BACK bet, a loser costs one unit and a winner earns (matched decimal odds minus one) multiplied by 0.98. The 2% commission is applied to winning profit, not the returned stake. These are normalised research units, not the actual cash amounts staked in the export.

The arithmetic matters because the raw workbook contains small and varying cash stakes. Simply adding its recorded currency profit and loss would answer a different question. Standardising stakes lets us compare selection quality without the changing stake size distorting the result. It does not establish whether the same selections would have been executable at larger stakes or whether a future bettor could reproduce the prices.

Two matches, two very different prices

Consider Sabah U19 against Dortmund U19 on 20 October. The collector backed Over 1.5 first-half goals at matched odds of just 1.38, yet the selection lost. The next day, Ghazl El Mahallah against Al Ittihad in Egypt was a winner at 4.90. One result cost a full unit; the other earned 3.822 units after the assumed 2% commission. Neither match proves anything on its own, and the export does not tell us the minute-by-minute story of the goals. Together, though, they show why a winning selection and a good long-run price are not the same question.

There is also a striking cluster on 21 October. Sixteen Japanese fixtures settled as losses for this collector, including FC Tokyo against Kashiwa, Kawasaki against Kobe and Kyoto against Kashima. These were separate selections concentrated around the same match window, not necessarily 16 decisions made one after another. At one normalised unit each, that cluster represents a 16-unit hit. A trader running several markets at once would experience that as concentrated exposure, regardless of how neatly each individual bet fitted the original rule.

Where the losses concentrated

The busiest matched-odds band was 2.00 to below 3.00, containing 1,288 selections. It lost approximately 36.5 units, around −2.8% ROI. The 468 selections at odds 3.00 to below 5.00 lost about 72.9 units, roughly −15.6%. The 248 selections below 2.00 also lost approximately 11.1 units. Three selections at 5.00 to below 10.00 all lost, but that tiny sample cannot support a general conclusion about longer prices.

Geography exposed striking differences without proving a geographical edge. Japan contributed 165 selections and lost about 42.4 units (−25.7%), while Italy produced approximately +20.6 units from 81 selections (+25.4%). Spain's 76 selections gained about 9.5 units (+12.5%). These country-level samples are small, the competition mix may differ, and looking for winners after seeing the results creates a serious selection-bias problem. Missing country codes also affected 277 selections, which lost about 30.1 units.

A loss in both months

October accounted for 1,470 selections and a loss of approximately 91.9 units (−6.3%). The 537 November selections available through 8 November lost another 31.6 units (−5.9%). The result therefore was not created by one isolated losing month followed by a clear reversal. Both observed periods were negative, though the sample covers only a little over five weeks.

What the collector does not establish

These results do not isolate the influence of match state, league liquidity, late team news, market suspension or the timing of the recorded prices. Nor do they prove that every country and odds subgroup would behave similarly outside this sample. A positive-looking subgroup found after the fact should be treated as a hypothesis for prospective testing, not as a strategy already demonstrated to work.

The collector's last recorded selections were on 8 November 2026, while other collectors in the same master workbook continued recording bets through 17 November. That is consistent with this experiment having been retired, although the export does not record the decision itself. There is a practical lesson in knowing when to stop collecting: after 2,007 selections and losses in both observed months, continuing to risk money without a specific, testable reason for expecting a different outcome would not have been disciplined research. Stopping a weak experiment is not proof that no profitable variation exists; it is a decision not to keep paying to find out blindly.

The research conclusion

Backing every recorded First-Half Over 1.5 Goals selection in this collector would have lost 123.49 normalised units over the observed period. That negative result is worth keeping, rather than mining the historical results for a flattering country or odds filter. Any proposed change should be specified before new bets arrive and judged against fresh selections. The same discipline matters in other football markets: our away-team collector analysis found that a familiar-looking selection rule can also lose money over thousands of bets.

Comments


Contact us

Disclaimer

Trade Carefully is platform that focusses on using data-driven approaches to build the right mindset to have any chance of success long term. Sports Trading is an extremely difficult path to follow. It requires strict discipline, patience, and can result in losses.  

 

If you struggle with gambling addition, please get the support you need from organisations such as GambleAware

2026 - TradeCarefully

Contact: support@tradecarefully.com

bottom of page