About Custom Venom
Custom Venom is a fantasy sports analytics platform built on a simple belief: you deserve to know why we think what we think, and whether we've been right.
The Problem With Most Projections
Every fantasy platform gives you a single number. "Player X will score 18.2 points." No range. No explanation. No way to verify if their projections were any good last week, last month, or ever.
That number creates false confidence. Fantasy football is inherently uncertain — injuries happen mid-game, weather changes, coaches make weird decisions. A single projection pretends that variance doesn't exist.
Our Approach: Embrace Uncertainty, Earn Trust
Every projection displays a floor, median, and ceiling — what we call Strike Range™. This isn't hedging; it's honesty. Some weeks a player's outcomes are tightly clustered. Other weeks, they're volatile. You should see that.
When something around the game matters — weather, a tough matchup, an injury designation, a snap count that just moved — we show it on the player and widen or tighten his range to match. What it never does is quietly move the middle number on you. Bad weather doesn’t make us predict fewer points; it makes us less certain, and you get to see that. And if we’re not confident enough to say something useful, we stay quiet. Silence is a feature, not a bug. See exactly how a projection is built.
We grade every projection against what actually happened, and we grade the whole range — not just whether the middle number was close. The design target is that about 12% of results finish below the floor and about 15% above the ceiling. When the real rates drift away from that, the ranges get adjusted. Our public record starts with the 2026 season on the Scoreboard, published week by week as the season goes.
The Technical Foundation
We project the box score first — catches, yards, carries, touchdowns — and only then turn it into fantasy points using your league’s scoring. That is why one player has different numbers in different formats: it is one projection scored several ways, not several guesses. The weekly run moves through seven stages, each one gating the next, so a bad input stops the run instead of quietly reaching your screen.
The range around that number is built from the player’s own profile and widths the system has learned, capped so nothing gets absurd, and blended with last week’s so it doesn’t swing on one noisy game. Then context adjusts how wide it sits — never where the middle lands. Afterwards we grade the whole range, not just the middle number, and feed the misses back in.
Every projection includes a timestamp showing when the data was last refreshed. Stale data gets flagged. Broken signals get muted. The system is designed to be honest about its own limitations.
What We're Not
We're not trying to be right every time. That's impossible in a game with this much variance. We're not here to sell you certainty.
What we are is a system that gets better over time, shows its work, and lets you verify whether that trust is deserved.