MLB Underdog Betting Strategy in Five Angles

Baseball is the one major American sport where backing underdogs is a workable strategy, and it's also a sport where backing every underdog loses money. This list gives five angles for working in the gap between those two facts. It's written for bettors who already play MLB sides and props and want a method for the plus-money side of the board. The angles run in the order you'd apply them to a single bet: the price first, then the odds band, then the venue, then the matchup, then what happens after first pitch.

In betting, the underdog is the side the sportsbook expects to lose, and its price carries a plus sign: +150 pays $150 profit on a $100 stake. What matters for strategy isn't the payout. It's the win rate the payout implies. The formula, as Wizard of Odds sets it out, is 100 / (odds + 100) for plus prices and |odds| / (|odds| + 100) for minus prices. At +150 you break even if you win 40% of the time. At +200 you need 33.3%.

That's how a bettor can lose most of their dog bets and still come out ahead, and it's why baseball is the sport to do it in. Covers.com's cross-sport study found MLB underdogs won 42.7% of games straight up since 2021, up from 41.5% in 2016 to 2020. Over the same recent period NFL dogs won 33.3% and NBA dogs 32.0%. No other major league comes close.

The best argument against all this is that favorites still win most games: one DeucesCracked guide puts the league-wide rate at 58% to 62%. That's true, and it doesn't matter, because you pay for those wins in the price. SportsBettingDime gives the clearest example: Justin Verlander went 21-6 in a Cy Young season, yet $100 on his team's moneyline in every one of his starts lost $234. In practice, put the implied percentage next to every dog before you form an opinion of the team, then write your own estimate beside it. If you can't explain why your estimate is higher, you have no bet. The expected value of each bet comes from that gap and nothing else.

The table below applies the formula to every common MLB moneyline. Real markets add the bookmaker's margin, which is why the two sides of a game add up to more than 100%.

Most guides that rank for this search put the sweet spot at +120 to +150, or stretch it to +200. The only odds-band split in them that comes with a named source says the value runs out before +150. Action Network's 2025 home-dog numbers show home dogs priced shorter than +150 going 246-240 for a 7.2% ROI. Home dogs at +150 or longer went 30-62 and lost $1,009 at $100 a game. That's a 32.6% win rate in a band where the cheapest price needs 40% to break even.

The other claims don't come with that kind of evidence. One DeucesCracked guide says home dogs at +120 or longer win about 45% and have been profitable in 14 of the past 20 seasons, but it cites no dataset. A WagerLens newsletter says to bet +100 to +140 because value "flattens" past +150, also without a source. The strongest case for longer prices is that a bigger payout needs fewer wins, and long dogs do come in: Screwball lists the Giants' +264 win over the Dodgers on May 12, 2026 as the season's biggest upset by price. But one upset doesn't make a rate. My position: treat +150 as a ceiling, and only go past it when your own estimate clears the table's break-even figure by a clear margin.

Inside that band, a few cents of price make a big difference. DeucesCracked gives an example of one book offering +165 and another +180 on the same dog. That moves your break-even from 37.7% to 35.7%, two points of win probability for the cost of having a second account. The Action Network's closing line value explainer works out what one cent is worth (taking PIT at +121 when it closes +120): a bettor making about 150 bets a month at $100 units who beats the close by a cent each time picks up roughly $75 a month. Keep several sportsbook accounts, bet the best number, and record the closing price next to every bet in a tracker that logs closing line value. Matthew Freedman's warning in the same series holds for underdogs too: "It's almost never a good sign when the market moves against you."

The 45.9% home-dog win rate that shows up across page one comes from a newsletter that gives no source, dates or method. The sourced version is less impressive, and you can actually check it. Action Network's 2025 study backed every home underdog with a flat $100. July 2025 was the best month for home dogs since 2005: 68-55 (55.3%), a 20.6% ROI and $2,536 profit. The nearest comparison was July 2016, at 61-50 for a 26.6% ROI. Taken over the season to the point of the study, though, the edge was +4.1% ROI ($2,484), with four winning months out of five and June the loser. Twenty of the 30 teams were at or above break-even as home dogs.

One team distorts the average. The Rockies went 13-35 as home underdogs in that sample, and without them the rest of the league was about four games under .500 as home dogs. That conflicts with a DeucesCracked guide naming Colorado as profitable home dogs at Coors Field. In 2026, Screwball counts 38 Rockies wins as an underdog, the most in baseball, but it gives no ROI, and a count of wins tells you nothing about profit at long prices. So the method is to back home dogs in the sub-+150 band, check each team's own record as a home dog, and remember that a hot month like July 2025 is the exception and not the expected rate.

Road dogs have no sourced league-wide split anywhere in the material behind this article. VSiN's Steve Makinen tracks one narrow road spot: dogs from +101 to +187 in the final game of a series that ends a road trip went 426-488 for +85.85 units, a 9.4% ROI since 2022. That's a specific situation, not a general road-dog edge, and a separate DeucesCracked claim about divisional road dogs at +110 to +160 comes without data.

Public money is the usual explanation for any of this. Michael Shackleford of Wizard of Odds argues that "square bettors prefer to bet favorites, creating value on the underdogs", and ESPN reported the Dodgers drew 40% of all World Series bets and 25% of the handle at DraftKings at +375 before 2026. That's a reason to fade public money selectively, not automatically: VSiN's DraftKings data shows home moneylines carrying 90% or more of the handle went 282-127 for a +6.8% ROI since 2023. Sometimes the crowd is right.

Lines often lean on ERA, so pitchers whose ERA runs well above their underlying numbers are a good place to look for underpriced dogs. FIP, as Action Network defines it, uses only strikeouts, walks, home runs and hit batters, so it estimates what a pitcher's ERA would be with league-average defense and luck. xFIP goes a step further and swaps actual home runs for a league-average home-run-per-fly-ball rate, and the article calls it a better predictor than ERA. TheLines' example is Eduardo Rodriguez in 2021: a 4.74 ERA against a 3.32 FIP, a career-best 27.4% strikeout rate, and a .363 BABIP causing most of the gap. FanGraphs projected him at 3.74 and Baseball Reference's Marcel system at 4.37, both closer to his FIP than to his ERA.

The case against this deserves a fair hearing. TheLines calls FIP backward-looking, and some pitchers, Kyle Hendricks being its example, beat their FIP year after year by inducing weak contact. So treat an ERA-FIP gap as a reason to investigate, not a reason to bet. None of the sources here works through SIERA with figures, so I won't invent any.

Then look at the other side of the matchup. A DraftKings Network best-bets column shows the method: the Angels at +100 behind Reid Detmers (2.47 FIP, 2.53 xERA) against a Toronto lineup ranked 19th in wRC+ at 99, and the Astros at +102 with the second-best lineup by wRC+ (122) despite an 8-15 record. Those were picks, not results, but the reasoning is right. A team's bad record next to strong underlying stats is the sort of mismatch the moneyline misprices.

Don't treat bullpen quality as a dog angle by itself. VSiN's 2026 season-to-date numbers have better-bullpen underdogs at 316-380, down 22.16 units (-3.2% ROI), while better-bullpen small favorites from -106 to -126 went 254-181 for +9.2%. If your edge is the starter, DeucesCracked and SportsBettingDime both suggest the first-five-innings market, which takes the bullpen out of the bet.

Last, the park. Covers rates Coors Field as one of the most extreme home-run parks because Denver's thin air carries fly balls and flattens breaking pitches. Wrigley's reputation is more about wind than the park itself: its home-run factor was 97 over 2024 to 2026, close to neutral, while Bleacher Nation cited a single-season Statcast figure of 90. Action Network's weather piece puts it plainly: "Wind speed, direction, and other weather factors can greatly diminish the park factors that favor hitters or pitchers in a given game." Check the day's wind before relying on a multi-year average. None of the sources here gives figures for Tropicana Field, so it's left out.

In-game betting is where a dog's price moves fastest, and it has the least sourced MLB data behind it. DeucesCracked suggests backing underdogs live after early, small-sample scoring, before the third inning, when the market may overreact to one big inning. WagerLens adds checking whether the dog's key relievers are rested, since a fresh bullpen protects a late lead. Both make sense, but neither publishes a record, and none of the sources here has an MLB live-dog ROI dataset.

What the sources do show is how far live prices can stretch. ESPN reported that in Game 7 of the 2025 World Series, with Toronto leading 4-2 in the top of the 8th, the Dodgers were +750 on ESPN BET's live moneyline. That price implies an 11.8% chance. The bet only works if your read on the remaining bullpens and lineup spots says the real chance is higher, and the method is the same as before the game: convert the price, then compare it with your own estimate.

For anyone searching for an NBA underdog betting system, the most rigorous example here is an in-game one too. Action Network's Bet Labs system backs home teams that were full-game underdogs and lead by three or more at halftime on the second-half moneyline. Since 2006 it has made $12,603 on $100 bets, an 8.9% ROI, while winning only 35.1% of the time; Boston once paid +447 in that spot against Golden State. Compare that with Wizard of Odds' 6,567-game NBA study, where blindly backing every pregame dog on the moneyline returned -4.21%, slightly worse than favorites at -3.75%. In basketball, as in baseball, a specific spot can make money where betting every dog doesn't.

Start with the first two angles, because they cost nothing: converting each price to a percentage and holding more than one account to get the best number. The other three are filters, and each one cuts the number of bets you'll find.

Stake size matters more with dogs than with anything else on the board, because the edge is thin and you'll lose more bets than you win. The DeucesCracked guides recommend 1% to 2% of bankroll per standard bet, 3% at most for the strongest plays, and flat staking reviewed over several hundred bets. Flat staking is the right default. None of the sources tests letting winnings ride, and with an edge measured in single-digit percentages, increasing stakes after a win mostly makes the swings bigger. The worked example below shows how small the margin really is.

That 2026 figure is the most honest current-season example in the sources. A reasonable-sounding dog system lost across nearly 700 bets, and the only defense was stake size. No system guarantees a profit, and long cold runs are normal in a sport where even good dogs lose more than half their games. Plan for that before it happens, using how to survive a losing streak as the guide, and run your unit size through the calculator below to see what streak a full season is likely to throw at you.