The Musburger Baseball System: Three-Game Streaks Backtested

Why a 47-year-old broadcaster’s MLB system trended in 2026

I had Brent Musburger’s system filed under “broadcaster nostalgia” until last May, when the Chicago Sun-Times published a backtest that forced me to dig my old notebook out of a drawer. The premise is almost embarrassingly simple: ride MLB teams on a three-game winning streak, fade MLB teams on a three-game losing streak, and repeat the trade until the streak breaks. It is a system that fits on a beermat and somehow returned roughly 77.72 units of profit on a $100-per-unit basis across the 2025 regular season.

That number is what dragged it back into UK MLB chat groups this spring. I have spent nine years modelling run lines and NRFI markets, and I rarely change my mind because of a single data point. But a profitable, easy-to-follow rule on a 162-game season demands a fair hearing, especially when most of the public material I see in the UK is still recycling vague “fade the public” advice. So I sat with the season log, ran the rules myself, and came away thinking the Musburger framework deserves a spot in any UK bettor’s research stack, with caveats.

The rules: three-game streaks, two ways

The first time I wrote these rules out for a colleague, he laughed. There are no FIP thresholds, no umpire splits, no park factors, no Vegas-line tests. Just two triggers and a stop.

The “ride” leg works like this. Any MLB team that wins three games in a row becomes a backable side. You stake their next game on the moneyline. If they win, you ride them again the following day. If they lose, the streak ends and the system disengages until they string together three more wins. Stake size is flat: one unit per game, no progression, no doubling.

The “fade” leg mirrors it. Any MLB team that loses three games in a row becomes a fadeable side. You bet against them on the moneyline in their next game. If they lose again, you fade them once more. If they win, the streak breaks and you wait. The stake is identical, one flat unit, against whichever opponent the schedule presents.

That is the whole system. There is no closing-line check, no power rating, no weather screen. Musburger built it to be readable on television, in a hurry, by anyone holding a remote control. The moment you start adding adjustments you are no longer testing his rules, you are building your own.

What the 2025 MLB sample produced

Here is where the numbers get interesting. The 2025 regular season backtest of the unfiltered system produced approximately 77.72 units of profit for a bettor staking $100 per game, which translates to roughly $7,772 by season’s end. For UK readers used to thinking in singles and percentage ROI rather than American “units”, that is effectively a strategy that hit somewhere in the low single digits ROI range across hundreds of plays. Not life-changing, but well above the breakeven a bettor needs at standard prices.

What I want to flag, because the headline number is seductive, is the structure of the sample. A 162-game schedule generates streaks unevenly. Bad teams produce more three-game losing runs than three-game winning runs, which means the fade leg sees a heavier volume of bets. Good teams cluster wins, which means the ride leg gets fewer attempts but those attempts are concentrated on better sides. The net result is a system that pays mostly through the fade leg, and mostly through plus-money payouts on opponents of struggling teams.

I tested it for variance against my own NRFI flat-bet history from the same year. Musburger’s drawdowns were shorter but his cold streaks were sharp, two or three weeks where he would go 8-15 before snapping back. If you are bankrolling this, that is a reality you have to budget for from day one. A flat one percent of bankroll per play is the maximum I would run on something this rigid.

The six-team filter that nearly doubled the edge

The most interesting part of the backtest is not the headline 77.72 figure. It is the filter the Sun-Times applied afterwards. When the same rules were limited to six teams only – Cleveland, Miami, Pittsburgh, San Diego, San Francisco, and Seattle – the ride-a-winner leg jumped from 21.03 to 39.95 profit units, and the combined season figure climbed to roughly 96.94 units. So a tighter universe, applied to the same mechanical rules, produced more profit than the broad version.

That tells you everything you need to know about why this kind of system can work in MLB. It is not because three-game streaks contain magic information. It is because certain kinds of teams are systematically mispriced when their results swing. Mid-market clubs with low public visibility, teams who attract no parlay action, teams whose lines are set by the book and barely moved by money – these are the sides where the moneyline stays a little stale, a little wide, a little favourable to the contrarian.

Musburger himself was clear about what he thought of the result. As he told the Sun-Times reporter, “Maybe you want to keep that to yourself. Think about that. As soon as you, well, the FanDuel and DraftKings bookies … you know, we’d better be careful.” That line is funny, but it carries a real warning. Once a system is on television, and once it starts moving handle, the books adjust. The 2025 numbers are not a guarantee for 2026.

Why streak-fading still has theoretical bite

I get pushback on this whenever I mention Musburger to other handicappers. The argument is always the same: streaks are not predictive, recency is a fallacy, you are betting on noise. I take the point seriously, and I want to address it head-on, because the answer is more nuanced than either side admits.

Streaks themselves are noise. A three-game losing run by a 90-win team tells you nothing useful about the next game. But the bookmaker line for that next game is not noise. It is built partly on opening price, partly on betting handle, and partly on how aggressively a trader wants to defend it. When a team has just lost three, public ticket count piles up against them – because casual bettors fade the recent loser – and the line drifts. That drift is what the fade leg actually exploits, and it is also why the related contrarian principle holds: across a long sample, betting on a sub-.400 team immediately after it wins one game following a loss has produced more than $19,000 in profit for a $100 bettor and has not had a losing season. That is a related fingerprint of the same effect, public overcorrection on weak teams, and it is structurally similar to what the Musburger fade leg captures.

So the system is not predicting future results from past results. It is harvesting a small public-money distortion that recurs in MLB more reliably than in other sports. Whether that distortion will survive 2026 sharper-priced books and the rise of mid-tier UK contrarian MLB betting material is the question I cannot answer yet. I would rather sit with the live sample for a month before I started staking it seriously.

Caveats: survivorship bias and 2026 risk

Two things worry me about treating the 2025 result as a forecast. The first is survivorship bias in the six-team filter. Picking the six clubs that produced the strongest signal in retrospect is exactly the kind of post-hoc fitting that makes backtests look better than live betting will. If you ran the same selection a priori on January data, you would not have arrived at that exact six. So the 96.94 unit figure is best read as a ceiling, not a target.

The second is line tightening. UK books especially, where margin pressure from the 2025 statutory levy is now a permanent overhead, are no longer leaving as much air on weak-team moneylines as they did even three seasons ago. Pinnacle has tightened its MLB pricing year on year, and most UK books anchor closer to Pinnacle than they admit. If the Musburger fade is exploiting a five-cent line softness, and that softness shrinks to two cents in 2026, your edge halves before you place a single ticket.

If you do run this in 2026, I would advise three things. Cap stake at 0.5% of bankroll per play, not the customary one percent, because the strategy clusters losses. Track every entry against closing line value, not just result, so you know within 50 plays whether the edge is still alive. And accept that this is research-grade, not portfolio-grade – the kind of system you run alongside your main book, not the one you build the bankroll on.

The cold honest verdict

The Musburger system is not a holy grail and was never sold as one. It is a clean, public, repeatable rule that happened to print money in 2025 because public action and book pricing left a small edge in plain sight. For a UK bettor who wants something to backtest and watch in parallel with their existing approach, it has real value. As a sole strategy, it has too much downside variance and too much exposure to line correction. Treat it the way Musburger seemed to treat it himself – as a curious thing you keep an eye on, not a system you bet the house against.