Three ideas, and then the games
- There are two kinds of penalties. Some are black and white: a false start either happened or it didn’t. Others are judgment calls: holding, pass interference, roughing the passer. A referee can throw that flag or keep it in his pocket. If anyone is tilting a game, the judgment calls are where it would show.
- We score every flag by what it cost. A holding call on first down in the first quarter barely matters. The same call on third down in the fourth quarter can decide the game. For each flag we work out how much it changed a team’s chance of winning. If a flag drops Texas from a 60% chance of winning to a 50% chance, we write that as −10%.
- We also check the betting line, because that is where the money is. Before every game, oddsmakers set a number for how much the favorite should win by. If Texas is favored by 6 points and wins by 3 points, Texas won but everyone who bet on Texas lost. A crooked referee would not need to change who wins. He would only need to change the final margin. So for every flag we ask two questions: did it change who wins, and did it change whether Texas beat that number?
Always the scoreboard. “Favored by 5.5 points” means oddsmakers expected Texas to win by about that much. “61 total points” is both teams’ scores added together. Yards are always labeled as yards.
Always a chance, or a change in a chance. Never a score. If Texas had a 75% chance to win and a flag leaves it at 54%, we write that flag as −21%. A minus sign means the chance went down for Texas.
How rare something is. It means that if you looked at 28 games from this season, you would expect to find about one where the calls were this lopsided.
This page never names a referee and never claims a game was fixed. It shows where the calls were unusual and says how unusual.
Seven judgment calls went against Texas and three against Tennessee, and the SEC admitted one more mistake afterward
See the play-by-play for all flags
What probably happened. Texas played an aggressive defense and got flagged for it, and one botched ruling on the onside kick nearly gave the game away. People who bet on Tennessee did get paid, but a touchdown paid them, not a flag.
What a fixed game would look like, and whether we saw it. First, unusual money on Tennessee before kickoff. We did not see it: the betting price on Texas barely moved all week. Second, calls that shrink the margin without putting the win in doubt. We did not see those either.
What would change our mind. The same referee crew leaning the same way in its other games, or video showing the pass-interference and face-mask calls were simply wrong.
Replay has gone Texas’s way more often than not
A video review is when the game stops and a replay official checks the ruling on the field against the video. It ends one of two ways. The call is upheld, meaning it stands. Or it is overturned, meaning the video showed it was wrong. There have been 27 reviews in Texas games this season. Fifteen went Texas’s way, seven went against Texas, and for five the public record does not say enough to tell.
See all 27 reviews, play by play
Why replay matters. One reversed call can be worth more than a whole game of flags. If someone wanted to move a game with a single decision, this is the tool they would want.
Why it is the hardest place to cheat. Everyone at home sees the same video the booth sees. The rule says a call can only be overturned when the video is clear. And in the SEC the replay official in the stadium is on the line with a second group of officials at the league office in Birmingham. A bad flag is one man’s opinion. A bad reversal has witnesses.
What a crooked booth would look like. Reversals that keep helping the same side. Reversals late in close games that change the final margin. And the one we cannot see: plays that should have been stopped for review and were not.
What we see, and the most likely story. The reversals have helped Texas, not hurt it. Replay exists to fix mistakes, so the team it helps is the team the officials on the field got wrong the first time. Of the 15 calls overturned in Texas games, 11 were rulings that had first gone against Texas, 2 were rulings that had first gone Texas’s way, and 2 we cannot tell. That fits the rest of this page: the close calls on the field lean against Texas, and the video leans back. The other story, a booth that favors Texas, is the one Missouri and LSU fans would tell, and no team in the country has gained more from replay this season. The LSU reversal is also the first to match what we said a crooked booth would look like: late, in a close game, and decisive. Whether the runner was down is something anyone can check on the video. Ten games cannot rule the other story out.
What would change our mind. Video of a reversal that was not clear-cut, or the same lean holding up over another season.
Every Texas game this season
Each game gets a needle, the numbers, and our read on what happened. The numbers are measurements. The read is our best guess, and we say how strong the case is. Most games land near the middle. That is what makes a red one worth your attention.
What a normal game should look like
How we measure it
- We took every play of every major college football game since 2014, about 1.9 million plays, and used them to learn how likely a team is to win, and to beat the betting line, from any situation: score, time left, down, distance, field position.
- For every flag this season, we compare two versions of the game: the one where the flag was thrown, and the one where it was not. The difference is what that flag cost.
- We add up the judgment calls in each game. Then we subtract what a normal crew would be expected to call in the same situations, play by play. In a typical game that is about three judgment flags on each team. It is more in a game with more plays, and the calls are worth more in a close fourth quarter. What is left is the needle: calls made, minus calls expected. A team that should have drawn four flags and drew one shows up here, even though no flag was thrown. We then compare each game with every other game this season. The idea is an old one, and the story of where it comes from is on the blog.
- Then we look for the one pattern a paid-off referee would leave: judgment calls, late, all leaning one way, that change the margin but not the winner.
What this cannot tell you
- Who, if anyone, took money. Red means the calls in a game were unusually one-sided. It does not say why.
- Which flags were not thrown. A hold that goes uncalled leaves no record. What we can measure is the gap it leaves: how many calls a game like this normally produces on each side, against how many were made. That gap is built into the needle. It cannot say which play the missing call was on, and a bad spot or a clock mistake leaves no trace at all.
- Reviews that never happened. We can see every play the replay booth stopped. We cannot see the plays it should have stopped and did not.
- Whether a call was correct. A flag can decide a game and still be the right call. Only the video settles that.
- Which referee threw the flag. The public record does not say, and we do not name or accuse anyone.
- The last minute of a game, precisely. Our numbers get shakier in the final 60 seconds. In the Tennessee game the betting market disagreed with us about one late flag, and we went with the market.
None of this is betting advice.
This site is free and has no ads. If it’s been useful, consider chipping in to keep it running through the season.