First Half Over, Game Under: The Parlay That Fights Itself

Across 3,028 NFL games from the 2015 season through the 2025 playoffs, a ticket pairing "first half goes over half the game total" with "full game finishes under the total" cashed 12.6% of the time. Multiply the two legs as if they were separate events and you would expect 24.5%, a fair price near +309. The honest fair price is closer to +691. The Best Bet on Sports has bet football live for more than twenty years with a verified $367,520+ in profit across all sportsbooks, and this is the two-leg structure we see misread more than any other, because the mistake is arithmetic rather than opinion.

The appeal is obvious. You watch a game open fast, you decide the pace cannot hold, and you want to get paid for both halves of that story. Sportsbooks will happily build you the ticket. What almost nobody checks is whether the two legs can both win often enough to justify what the ticket pays.

Everything below comes from public nflverse play-by-play data covering the 2015 through 2025 NFL seasons, regular season and playoffs, 3,028 games in total. The first-half score is the score at the end of the second quarter. The number each game is measured against is the closing full-game total recorded for that game in the dataset. Nothing here is a projection.

One note on method before the numbers. Sportsbooks post an actual first-half total, and it typically sits a bit above half the game number. We are using half the game total as the benchmark because it is the only line available consistently across 11 seasons. In this sample, first halves produced 50.4% of all points scored and averaged 50.8% of the closing game total, so half the number is a close and slightly conservative stand-in.

Why Does a Fast First Half Make the Under Harder?

Start with the thing that makes this ticket different from a normal parlay across two games.

When you parlay the Chiefs and the Ravens, the two legs are separate events. One result does not enter the other's box score. When you parlay a first-half over with a full-game under, the legs share a term. The final score is the first-half score plus the second-half score. Every point that helps leg one hurts leg two, one for one.

That is not correlation in the usual sense. It is containment. Leg one is a bet that a specific chunk of leg two's number has already been spent.

The numbers show exactly how expensive that is:

Each leg on its own is close to a coin flip, which is what you would expect from a market that prices totals well. Put them together and the ticket hits barely half as often as independent pricing suggests. A game that starts fast finished under only 26.2% of the time, against 50.7% for a random game.

Games went over 48.4% of the time, under 50.7%, and landed exactly on the number 0.9% of the time. The market is doing its job on the full-game number. It is the joint outcome where the gap opens.

How Often Does a Fast First Half Still Finish Under?

The blunt version above splits games at one threshold. Here is the full picture, with games sorted by how much of the closing total was already on the board at halftime.

Read down the under column and the ticket dies in front of you. Once a game has 65% of its number on the board at halftime, it finished under 15.5% of the time. Past 80%, it finished under in 8 games out of 224.

Now read down the second-half column. That is the part that should stop you.

Does the Second Half Slow Down After a High-Scoring First Half?

No. Average second-half points across those six groups: 22.0, 23.1, 22.6, 23.2, 22.4, 22.9. The first halves in that table range from 11.4 points to 40.7 points. The second halves that follow them are, for practical purposes, identical.

Sorting by raw first-half points tells the same story:

First-half scoring moves across a 30-point range. Second-half scoring responds by less than four points, and it moves in the wrong direction for the under. The correlation between the two halves across all 3,028 games is 0.075, which is close enough to zero to treat as noise, except that it leans slightly positive.

There is one more way to state it, and it is the cleanest. Second halves stayed under half the game total 53.1% of the time overall. After a first half that went over half the number, second halves stayed under half the number 53.2% of the time. The fast start told you nothing about what was coming.

This matters because the entire folk theory behind the ticket is regression. Teams trade scores early, coaches tighten up, the clock runs, the total settles back down. It is a good story. Eleven seasons say the second half does not know what the first half did.

What Would This Parlay Have to Pay to Be Worth It?

At 12.6%, break-even is +691. You will not be offered +691.

Sportsbooks know these legs overlap. Same-game parlay engines reprice correlated combinations rather than multiplying them, and this pairing is one of the most obviously correlated in football, so it gets shaded hard. In practice you will see something in the +200 to +300 range on a standard total, which is at or below the naive independent price and less than half of what the historical rate is worth.

Work an example at +250 and $100 a ticket. Over 1,000 tickets you would expect 126 winners at $350 back, which is $44,100 returned against $100,000 staked. That is not a bad price. That is a structural hole in the floor.

The reverse ticket is no better. Pairing a first-half under with a full-game over cashed at a similar disadvantage: after a slow first half, games went over only 25.4% of the time against a 48.4% baseline. Same mechanism, opposite direction. You cannot fix a containment problem by flipping which side you take.

Compare that with a two-leg over parlay across two different games, where the legs really are close to independent and the correlation you would need to justify the ticket is small but at least the arithmetic starts from a fair place. This ticket starts from a false one.

What to Do With a Fast First Half Instead

If a game opens fast and you genuinely think the pace is wrong, you have cleaner ways to express it than a ticket built to lose.

Bet the second half by itself. The second-half total is its own market with its own number, and it is priced off what just happened rather than off what was expected three days ago. You give up the parlay payout and you get a bet that is not carrying the first half on its back. The trade-offs on those numbers are worth understanding before you use them, and we walk through them in our guide to second-half spreads and parlays.

Bet the live full-game under on a reason, not a number. The live total will have moved up to reflect the fast start. If your read is that the fast start was an illusion caused by two short fields and a special-teams score, the live number may now be genuinely too high. That is a real edge. "It feels like too many points" is not.

Take the first-half over as a standalone bet before kickoff. If your actual opinion is that these teams start fast, buy that opinion by itself at -105 or so. Adding the under leg does not make the opinion stronger. It makes you pay for a second opinion you never had.

Stop treating parlay payouts as the goal. The payout on a two-leg ticket is the compensation for two things going right. When one of those things makes the other less likely, the compensation has to be much larger, and no book is going to hand it to you. This is the same lesson that shows up in why most parlays lose, just with unusually clean arithmetic behind it.

Season to season, none of this wobbles. Under rates by year ranged from 45.6% in 2024 to 54.9% in 2022, and average second-half scoring ranged from 21.3 to 24.3 points. There is no era in the sample where a fast first half predicted a quiet second one.

Get Tonight's Live Picks

Want tonight's live in-game picks delivered to your phone during the game, with the reason attached instead of a bare number?

The Best Bet on Sports is the only live betting handicapping service limited on all six major U.S. sportsbooks (FanDuel, DraftKings, Caesars, BetMGM, Fanatics, ESPN BET) for winning too much during in-game action. Verified profit: $367,520+. Picks delivered via Email and Discord during games.

→ Get tonight's live picks: $199 first month — 1-Unit package, full live betting access → Try a live pick first — reserve your spot for tonight's pick

Compare every package on the buy sports picks page, see what a sports handicapping service is supposed to do for the money, or check the graded record on the results page. Weekend football coverage runs through NFL picks, college football picks and NFL over under picks.

Frequently Asked Questions

How often does an NFL game go over the first-half total and under the full-game total?

Across 3,028 NFL games from 2015 through the 2025 playoffs, both happened together 12.6% of the time, using half the closing game total as the first-half benchmark. That is a fair price of about +691 with no sportsbook margin. Pricing the two legs as independent events would suggest 24.5%, or about +309, which is why this ticket looks better than it is.

Why is a first-half over plus full-game under parlay priced worse than multiplying the legs?

Because the legs overlap rather than being separate events. The final score is the first-half score plus the second-half score, so every point that wins the over leg pushes the final closer to the total you need it to stay under. Sportsbook same-game parlay engines account for this and reprice the combination instead of multiplying the two prices.

Do NFL second halves score fewer points than first halves?

Barely. Across the sample, first halves averaged 23.0 points and second halves averaged 22.7. First halves produced 50.4% of all points scored. The idea that scoring drops sharply after halftime is not supported by 11 seasons of data, and it is the assumption most first-half-over, game-under tickets are built on.

Does a high-scoring first half mean the second half will slow down?

No. Grouping games by first-half points, second-half scoring averaged 21.2 points after a first half of 10 or fewer and 25.0 points after a first half of 39 or more. The correlation between the two halves is 0.075. Second halves stayed under half the game total 53.1% of the time overall and 53.2% of the time after a fast first half.

What happens to the full-game total once a game starts fast?

It becomes very hard to reach the under. Games that had 55 to 65% of their closing total on the board at halftime finished under 33.4% of the time. At 65 to 80% it fell to 15.5%, and at 80% or more it fell to 3.6%, or 8 games out of 224. The under does not need the second half to be quiet, it needs it to be historically quiet.

Is the reverse ticket, first-half under plus full-game over, any better?

No, it fails the same way. After a first half that stayed under half the game total, games finished over only 25.4% of the time, against a 48.4% baseline for all games. The two legs still share a term, so the structure works against you no matter which direction you point it.

What is a better way to bet a fast start than this parlay?

Bet one thing at a time. Take the second-half total as its own market, priced off what actually happened, or bet the live full-game under only when you have a specific reason beyond the score, such as an injury to a key player or a defensive change. If your real opinion is that these teams start fast, buy the first-half over by itself rather than paying for a second leg you do not believe in.