The Spreadsheet That Changed How I Bet on Baseball
I spent my first two seasons betting MLB the way most UK punters do — eyeballing batting averages and win-loss records, backing teams that “felt” strong. My results matched the approach: mediocre. Then a mate who works in data analytics shared a spreadsheet tracking FIP, wOBA, and BABIP for every starting pitcher in the American League. Within a month I found pricing gaps I had never noticed. Within a season I understood why sharp bettors treat traditional stats like training wheels and advanced metrics like the actual steering wheel.
Sabermetrics — the empirical analysis of baseball through statistics — is not new. Bill James popularised it in the 1980s, and the Oakland Athletics made it famous a decade later. What is new is its direct application to betting markets. The UK gambling industry generated GBP 11.5 billion in gross gambling yield in 2023-24, and a growing slice of that comes from punters using data-driven approaches to American sports. MLB’s 2,430 regular season games produce a staggering volume of data, and the bettors who know how to filter signal from noise hold a genuine edge.
Why Batting Average and ERA Are Not Enough
Here is something that still surprises people: a pitcher can have a 4.50 ERA and still be a better bet than a pitcher with a 3.20 ERA. How? Because ERA includes outcomes the pitcher cannot control — namely, the defence behind him and the sequencing of hits. I backed a starter in 2024 whose ERA sat at 4.30 while his FIP read 3.10. The gap told me his team’s defence was costing him runs, not his arm. The market priced him off the ERA. I priced him off the FIP. Over the next six starts, his ERA fell to 3.40 as the luck evened out.
FIP — Fielding Independent Pitching — isolates the outcomes a pitcher directly controls: strikeouts, walks, hit batsmen, and home runs. It strips away the noise of defensive alignment, outfielder range, and batted-ball luck. When a pitcher’s ERA sits well above his FIP, the market tends to undervalue him. When it sits well below, the market overvalues him. That divergence is where value lives.
Batting average suffers from a similar problem. A hitter slashing .330 with a .380 BABIP (batting average on balls in play) is riding luck. The league-average BABIP hovers around .300, and individual hitters regress toward their career norms over time. The batter with a .280 average and a .260 BABIP is the one I want to back in player prop markets — he is due for positive regression, not the other way round.
wOBA and Expected Stats — the Metrics That Price Runs
Last April I placed a series bet on a team the market had as underdogs in all three games. Their offence ranked 22nd in batting average but 8th in wOBA. I trusted the wOBA. They won the series 2-1, and the moneyline prices had been generous because casual bettors and even some algorithms still anchor on traditional slash lines.
wOBA — weighted on-base average — assigns run values to each offensive event. A single carries roughly 0.882 run value while a walk sits around 0.691. Unlike batting average, wOBA treats a home run as worth more than a bunt single, which sounds obvious until you realise how many markets still price teams off raw average. The metric correlates more tightly with run scoring than any traditional stat, and run scoring is what determines baseball outcomes.
Expected stats take the concept further. xwOBA uses launch angle and exit velocity data from Statcast to calculate what a hitter’s wOBA “should” be based on the quality of contact rather than the actual outcomes. A hitter whose actual wOBA trails his xwOBA by 20 or more points is hitting the ball hard but getting unlucky — line drives caught by well-positioned fielders, rockets that found gloves instead of gaps. That gap typically closes over a sample of 200-plus plate appearances, and the closing creates value for anyone positioned on the right side of the correction.
FIP, xFIP, and SIERA — Choosing the Right Pitching Metric
Three pitching metrics sit on the betting shelf, and reaching for the wrong one at the wrong time costs money. I learned this the hard way when I bet heavily on a flyball pitcher at Coors Field based on his FIP alone, forgetting that FIP does not adjust for ballpark-inflated home run rates.
FIP works best as a general-purpose measure. It tells you what a pitcher’s ERA “should” be based on strikeouts, walks, and home runs. xFIP goes a step further by normalising the home run rate to the league average — useful for identifying pitchers who have been either lucky or unlucky with how many fly balls leave the yard. A pitcher with a high ERA, a low FIP, and an even lower xFIP is practically shouting “back me before the market catches up.”
SIERA — Skill-Interactive ERA — adds complexity by accounting for the types of batted balls a pitcher generates. Ground-ball pitchers suppress home runs naturally, so xFIP can underrate them by normalising away an actual skill. SIERA handles this better. For my money, SIERA is the most predictive single-number pitching metric available, but it requires access to batted-ball data that not every free source provides. I use FIP for quick screening, xFIP for fly-ball pitchers at extreme parks, and SIERA when I have the data for a deeper dive.
Putting Sabermetrics Into Practice on UK Platforms
Numbers on a spreadsheet mean nothing until they connect to a betting decision. I run a straightforward process every morning before the day’s slate opens. First, I pull each game’s starting pitcher and check the gap between their ERA and FIP. Any gap larger than 0.50 gets flagged. Second, I look at the team-level wOBA for the past 30 days against the handedness of the opposing starter — left-handed or right-handed. Third, I cross-reference with the closing line to see whether the market has already priced in the information. If it has not, I bet.
In-play wagering now accounts for 62.35% of online betting revenue across the UK market, and sabermetrics translate seamlessly to live betting. If a starter’s first-inning exit velocity against is averaging 95 mph-plus, his expected outcomes are deteriorating in real time regardless of whether runs have scored yet. I use Statcast’s live data to front-run bullpen changes — if a pitcher’s stuff is degrading, the in-play line has not always caught up to the reality the underlying data reveals.
The practical edge for UK bettors is that most platforms still price MLB games primarily off win-loss record, ERA, and recent form. The bettors using sabermetrics are competing against a market shaped by those surface-level inputs. You do not need a PhD in statistics. You need FIP, wOBA, and BABIP. You need to understand when a player’s results are likely to improve or deteriorate based on the quality of contact and the sustainability of performance. And you need the discipline to bet the data when it contradicts the narrative, which remains the hardest part of this entire exercise — something I explore further in the context of closing line value and long-term profitability.
