About5 min readJanuary 2026

We built the prop research tool we wanted to use

Aggregate the data. Surface the signal.

Player prop research has an information problem. Market lines are published in seconds using models you'll never see. To evaluate them you need a browser tab open to the box score, another to a defensive ranking site, a third to the market data, and a fourth to whatever pace-adjusted projection you trust this week.

props.fyi is the answer to that problem. We pull every input that matters — game logs, defensive matchups, aggregated market data, game context — and run them through a model that surfaces one number: the gap between the published line and the model's projected hit rate. Then we put it on one screen.

What we built

The core of props.fyi is a single, sortable, filterable table covering 23+ player prop types — Points, Assists, Rebounds, PRA, Steals, Blocks, Threes, Fouls, Double-Doubles, quarter-specific stats, and full alt-line ladders. Every row gives you the eight numbers that summarise the prop:

L10 hit chart — the last 10 games visualized as a hit/miss grid, so you can see streaks at a glance
Season hit rate — full-season percentage at the current line
DvP rank — defensive vs. position ranking from 1 (toughest matchup) to 30 (easiest)
Aggregated market data — pulled in real time from major public market feeds
Value Score (0–100) — our proprietary composite ranking, weighing hit rate, projection gap, matchup, and game context
Signal % — the gap between the published market's implied probability and our model's projected hit rate
Streak detection — automatic hot/cold flags
Alt-line analysis — test ±0.5 line adjustments without leaving the page

Filter by hit rate range, position, home/away, prop type, market range, or just hide everything except today's market movers. Sort by Signal to see the most mispriced lines on the slate. The URL updates as you filter — share a view, bookmark a search, come back tomorrow and pick up where you left off.

How we calculate the signal

The Signal % isn't a guess. It's the difference between two probabilities:

01The market's implied probability — derived from the posted line, with the spread removed.
02Our model's projected hit rate — built from historical performance, defensive matchup adjustments, game environment (pace, total, spread, home/away), and recent form.

When our model says LeBron clears 7.5 assists 65% of the time, but the published market is pricing it at 58% implied, that 7-point gap is the signal. A positive signal means the projection sits above the market. A negative signal means the projection sits below it.

We're explicit about one thing: this is probabilistic research, not predictions. Variance is real. Past performance does not guarantee future outcomes. The math only works if you respect sample size and process.

Who props.fyi is for

You'll get value here if you're:

A serious sports analytics enthusiast who already understands hit rates, alt lines, and implied probabilities.
A DFS player using prop data to anchor lineups.
A content creator, journalist, or researcher tracking signals across hundreds of props per slate without manually pulling data.
Anyone tired of paying $300 a month for a Discord that posts the same screenshots you could generate yourself.

You won't get value here if you're looking for "lock of the day" picks, parlay tips, or guarantees. We don't sell those because they don't exist.

Sports & coverage

Live now: NBA, EuroLeague, football, and CS2 — full coverage across every game, every prop type, every night.

On the roadmap: NFL, NHL, MLB, NCAA Basketball, NCAA Football, WNBA, and MLS. We'd rather ship one league well than four leagues poorly, so coverage expands as we can guarantee the same data quality our live products have.

The interface ships in English and Turkish today, with more locales as the platform grows.

Why we built this

We're sports analytics enthusiasts first. props.fyi exists because the existing options were either fragmented free tools that took 20 minutes per prop to research, or expensive picks services that hid the math behind subscription paywalls and influencer credibility.

We think the math should be public. Build the model, expose the inputs, let the user judge whether the signal is real. No tout culture, no "trust me bro" picks, no parlay-of-the-day marketing. One product, one price, one job: surface the statistical signals you'd otherwise miss.

We're independent. We don't take third-party operator money, we don't sell user data, and we don't run an affiliate funnel disguised as a research platform.

Pricing