I Had a Winning Record and Still Lost Money. Here's How That Happens.

At the end of 2024, I calculated my full-year results. My record was 298-282, a 51.4% win rate. Not great, not terrible. I expected to be close to break-even.

I was down $1,400.

How does someone win more bets than they lose and still lose money? It's easier than you'd think, and I'm sharing this because I know I'm not alone.

The problem: inconsistent bet sizing driven by emotion.

When I was confident (often unjustifiably), I bet bigger. When I was less confident, I bet smaller. Sounds logical, right? The issue is that my big bets performed worse than my small bets. I bet big on marquee games where the lines were sharp. I bet small on niche games where I actually had edges.

I also chased losses. After a bad Sunday, I'd come back Monday with bigger NFL bets trying to get even. These revenge bets performed terribly.

And I bet more on my favorite teams. Emotional attachment led to poor decision-making. My Eagles bets were my worst-performing category.

The math:

My average winning bet was $87. My average losing bet was $104.

Even at 51.4% wins, that gap killed me: (298 × $87) - (282 × $104) = $25,926 - $29,328 = -$3,402

Wait, that's even worse than the $1,400 I reported. The difference came from a few parlay wins that partially offset the straight bet losses. But the core problem is clear: I won more often at lower stakes.

What I changed:

Flat betting. Every play is now 2% of my bankroll, period. No exceptions for confidence, for grudges, for "I really like this one."

Banned myself from Eagles games. Just completely removed them from my betting universe.

No Monday bets based on Sunday results. If I want to bet Monday games, I have to have identified them as targets before Sunday started.

These changes turned my 2025 into a profitable year so far. Same analysis approach, same win rate, completely different financial outcome.

Betting skill is only half the battle. Betting discipline is the other half, and I was failing at it completely.