A disciplined college-football workflow built for people who are tired of guessing. Profile the week's slate to your own rules, let the GRIDIRON engine project every score from real 2025 team data, and size each play with conviction-based stakes — all live off the actual ESPN board, all in the browser, no spreadsheets and no touts.
You bring the bankroll and the discipline. The Process brings the number, the stake, and the verdict — the same three-stage loop, every Saturday, on every game that fits your shape. What you see on your screen is what a full-time bettor would build for themselves if they had the time: a live board, a real model against a real market, and a stake that answers to math instead of mood.
There is no chat room, no upsell funnel, no "premium play of the year." One tool, one process, one number per game. Run it every week and let the sample size do the talking.
Set your parameters once — favorite location, points laid, and total window. The board pulls the live FBS slate from ESPN and surfaces only the games that match your shape, so you never spend a Saturday scrolling past games you were never going to play. The profile is a pre-commitment, not a suggestion: it decides what you're allowed to bet before the emotion of kickoff can talk you into something else.
GRIDIRON runs each qualifier through shrinkage-stabilized PF/PA, the market-weighted Claw-Back, and 8,000 Monte-Carlo sims down to a projected final score, a cover probability, and an over/under read — a number that answers to the tape, not the narrative. When the model and the market disagree, the disagreement is the whole point; that is where the edge lives.
Every edge becomes a definitive side or total, then gets sized by conviction under a hard weekly exposure cap. The stake rises directly with the win probability — a 86% cover probability is an 8.6% of-bankroll bet on a $1,000 roll — so a heavy card can't overcommit the roll and a light card doesn't leave money on the table.
The Process is not a picks service. There is no Discord full of last-second parlays, no cappers screaming lock of the year, no vague "unit" system that expands or shrinks whenever it needs to. It's a repeatable workflow that turns the week's live board into a projected final score, a definitive side, a definitive total, and a stake you can defend on Sunday morning — win or lose. You run it. You own it. Every Saturday looks the same.
The reason the same workflow can be trusted week after week is that it never changes shape to fit the outcome. Losing weeks don't shrink the model and winning weeks don't inflate it. The stake is the same fraction of the roll it was in Week 1, the tiers are the same tiers, and the number is built the same way. That is what makes it a process instead of a run.
Every game, spread, total, and moneyline is pulled straight from ESPN's public scoreboard at the moment you load the tool. Nothing is cached, nothing is stale, and the model is projecting against the number a real book is actually posting — not a screenshot from last night. When the line moves at 11am, you refresh and the model moves with it.
Team ratings come from actual season PF and PA, shrunk toward the national mean so a fluky 55-point Week 2 doesn't distort September. In Week 1 the model leans on the prior season baseline and quietly widens its confidence bands until real games bank. The number gets sharper as the season does — it doesn't pretend to be sharp in Week 0.
Conviction-based staking, a weekly exposure cap, and tiered verdicts (HAMMER / BET / PLAY / LEAN) mean the same math sizes a five-game Tuesday MACtion card and a twelve-game Saturday slate. The rules don't bend because the week feels big, and they don't shrink because last week hurt.
Every game shows the projected score, the market spread, the market total, and the gap between them at a glance. You don't have to guess where the edge is — the tool draws it for you, and the LeakBand shows how confident the model is in the outcome distribution before you ever look at a stake.
When the market piles on a team the fundamentals don't support, the Claw-Back pulls that favorite's rating toward the underdog's before the projection runs — scaled by how hard the market is committed. It is the mechanic that stops the model from following the same narrative that is currently paying the books' bills.
You set your bankroll and your weekly cap once. The tool sizes every stake off those two numbers, and refuses to let a heavy week silently push you past the cap. Winning is optional; not blowing up is not.
No touts, no lean-on-vibes, no "trust me on this one." A repeatable process you can run every Saturday, defend to yourself Sunday morning, and improve on across a full season instead of chasing last week's variance. Sign up free, run the board, and see the difference between an opinion and a number.
Three stages, one discipline — exactly what happens between the live slate and the stake on your ticket, with nothing hidden in a black box. Read it once and you will know why the tool recommends what it recommends, every single Saturday.
You define the shape of a bet you want to take before you ever look at a game. Home or away favorite, the range of points laid (default 3 to 4.5), and a total window (default 47 to 57). Every Saturday the tool loads the full FBS board straight from ESPN's public scoreboard and filters it down to just the games that match. In-season it also reads each favorite's recent form — cover streaks, scoring trend, pace drift — and surfaces those flags next to the matchup. In Week 1 those flags sit out until games bank, and the model tells you so rather than pretending it knows.
The point of profiling is not to find more bets. It is to pre-commit to the kind of bet you're actually good at, so Saturday becomes a filter and not a temptation. If a game doesn't fit the shape, it never makes the board — no matter who's playing, what time it kicks, or how loud the narrative is.
The default profile — mid-sized road/home favorites in a mid-range total — is the meat of the college-football schedule and the zone where public perception drifts furthest from real efficiency. You can tighten it if you're targeting a niche, or widen it if you want to see the whole slate; the math on the back end doesn't care, but your discipline will.
Each qualifying game is projected from real seasonal scoring, not power ratings scraped from a message board. Team PF and PA are shrunk toward the national mean so small samples don't scream in September, then turned into expected points via an offense-versus-defense interaction with a home-field adjustment. That expected-points estimate is fed into a Markov drive model and run across 8,000 Monte-Carlo simulations of the game.
What lands on your screen is a projected final score, a cover probability against the live spread, and an over/under read against the live total — plus a LeakBand showing how tight or wide the distribution of outcomes is. A one-possession edge on a game with a huge variance profile is not the same as a one-possession edge on a low-variance game, and the model refuses to pretend it is.
Shrinkage is the piece most casual bettors skip and most sharp bettors live by. In Week 3 a team that has scored 45, 52, and 10 does not have a "true" offense of 35.7 points per game — it has a small, weird sample. Pulling those numbers toward the national mean keeps early-season projections from over-reacting to two games, and the pull automatically weakens as the sample grows. By November the model is trusting the tape almost entirely; in September it is deliberately humble.
Ported from the ERA Claw-Back used in baseball totals modeling. It fires only when the spread favorite has a worse season net differential (PF−PA) than the underdog — that is, the market is backing a team in spite of its own fundamentals, usually because of brand, recency, or a highlight-driven narrative. In that spot the engine claws the favorite's rating toward the dog's before projecting, and the strength of that pull scales with market conviction.
Conviction is read from the real posted moneyline when a book has one, and derived from the spread when it doesn't. A -2.5 / -135 favorite barely gets touched — the market is only mildly committed. A -7 / -350 favorite whose PF−PA is worse than its opponent's gets clawed hard, because the market has planted a flag on a team the tape doesn't support. The Claw-Back is the single biggest reason The Process's number will sometimes look nothing like the consensus, and it is the mechanic you're paying attention to when a "boring" underdog suddenly grades as a HAMMER.
Crucially, the Claw-Back does not fire on every underdog. If the favorite deserves to be favored — better PF, better PA, better net — the mechanic sits out and the projection runs clean. It only intervenes when the market and the fundamentals are actively pointing in opposite directions, which is exactly the spot the model was built to exploit.
The cover probability feeds the stake directly: win probability × 10 = percentage of bankroll. A 86% cover probability sizes to 8.6% of the roll — so on a $1,000 bankroll the tool recommends $86. A hard weekly exposure cap scales the whole card down if a heavy week would overcommit the roll — so a twelve-game Saturday with five real edges never quietly turns into a bankroll-threatening night.
Every play then lands in a tier — HAMMER, BET, PLAY, or LEAN — driven by the model's cover probability against the live number. The tier tells you how hard the model is leaning; the stake tells you what that lean is worth in dollars. Together they replace the two questions that ruin most bettors' Saturdays: "How much do I like this?" and "How much do I put on it?"
The sizing is aggressive by design because the edge is the probability itself. If the model prices a side at 86% to cover, a 1.6% Kelly bet wastes the edge. The cap is the guardrail, not a shrunken fraction of Kelly — it keeps the total weekly exposure inside the limit you set, and the tool scales every stake down proportionally when the card is too heavy.
The Process does not scrape injury Twitter, guess at weather, or price in coaching drama. It works off the number the book has actually posted and the scoring the teams have actually put up, and it assumes you — the bettor — will decide when a late scratch or a monsoon makes a play a no-go. The model gives you an honest edge against the live market; the human still has the last word before the ticket is written.
It also does not chase. There is no "double up to catch up" logic, no bump on plays because last week was rough, no shrink after a big win. Every week starts from the same rules and the same bankroll math, and every game gets the same treatment regardless of what the previous Saturday looked like.
Open the tool on Friday night or Saturday morning. Confirm your profile still matches the kind of bets you want to take this week. Refresh the board — the numbers are pulled live — and let GRIDIRON project every qualifying game. Scan the tiers: LEAN and PLAY are low-conviction directional reads, BET is a real edge worth a stake, HAMMER is the model's strongest read of the week. Cross-check any game you're going to bet against your own eye — injuries, weather, motivation — and either take the play at the stake the tool recommends, or pass. That is the loop. It is deliberately boring; that is the feature.
Everything runs in your browser off the live board. Nothing here is a guarantee, and every probability is a model estimate — the market is efficient and edges are thin by design. 21+. If gambling stops being fun, take a break, and call 1-800-GAMBLER if you or someone you know needs help.
Browse the board free. Unlock the definitive picks, stakes, and Claw-Back readouts with Pro. One tool, one process, transparent pricing — no upsells, no add-ons, no premium-play-of-the-year email at 11pm.
The Free tier is not a demo. You get the live ESPN board, the full profiling filter, and every projected final score the model produces — the same score Pro users see. You can compare the model's number against the market number on every game, all season, without paying a dollar. What Free doesn't do is tell you what to bet or how much: no definitive side, no total call, no stake, no tier.
That's on purpose. Free is designed for people who want to sanity-check the model against their own reads before trusting it with a bankroll. Run it for a couple of weeks, watch the projections against the closing numbers, and if the model is doing what it says it does, upgrading to Pro is the easy call.
Definitive side and total on every qualifying game, a conviction-based stake sized off your bankroll and weekly cap, the HAMMER / BET / PLAY / LEAN tier for each play, and the full Claw-Back readout showing when and why the model is pulling away from the market. In short: everything you'd actually need to bet the card without a second tool open.
Weekly ($7.99) is designed for the bettor who plays a handful of Saturdays a year — bowl season, rivalry week, a hot conference stretch — and doesn't want to commit up front. SZN ($299) is the full-year, all-sports pass: CFB from Week 0 through the natty, plus NFL, CBB, MLB and every new model as it ships. Straight math: $7.99/wk × 52 = $415.48, so SZN pays for itself in roughly 38 weeks and covers everything else free after that.
Pro access is granted on the account you sign in with. There is no card on file with the tool itself and nothing renews automatically — you decide each week (or each season) whether to keep going. Cancel by simply not paying the next invoice; the account reverts to Free and you keep the projections.
Send payment via Cash App to $TuttlemathWinsGames with your account email in the note. Access is flipped on manually once payment lands — usually within a few hours on weekdays and faster on Saturdays. Prices above are placeholders; the operator reserves the right to adjust them as the tool matures.
Refunds are handled case-by-case for the current billing period only. Once a week's slate has been unlocked and viewed, that week is considered delivered. Season refunds are pro-rated on request within the first two weeks of activation.
Payment note: "Unlock" sends you to Cash App at $TuttlemathWinsGames. Include your sign-in email in the payment note — access is switched on manually after payment lands, and it stays locked until then. If you've been given lifetime access, that flows through the account and follows you across devices automatically.
Run the Free tier for a Saturday or two, compare the projected finals to the closing numbers, and see whether the model is doing what this page says it does. If it is, upgrade. If it isn't for you, keep the Free board — it's yours regardless. That's the whole pitch.
The full FBS board straight from ESPN — live scores, quarters, and clock updated on demand. This is the same feed The Process reads to build your number.
Radical transparency. No cherry-picked screenshots, no back-tested "would-have-been" ROI, no fabricated units. When live results exist, they'll live here — every pick, every stake, every close, win or lose.
The Process is a new public tool. Rather than publish backfit numbers from a model that already knew the outcomes, we're doing this the honest way: every pick The Process generates from kickoff forward gets logged here with the closing line, the stake size the model recommended, and the graded result. No edits, no deletions.
Until real games have closed under this framework, this page will show exactly what you're looking at — an empty ledger and a promise to fill it in public.
Anyone can post a spreadsheet. Back-tested results built on final data — even honestly built ones — reliably over-fit and cannot be independently verified. We'd rather show you nothing than show you numbers you have no way to trust. When the first slate is graded, this page updates.
Build your game profile once. The board matches the week's slate against it and surfaces the games that fit, GRIDIRON projects each score, and conviction-based sizing sizes the stake off your bankroll — with a hard weekly exposure cap so a heavy card can't overcommit the roll.
| Matchup | Fav | Line | Total | Fav last | PF / PA |
|---|
Live board loads from ESPN's public college-football scoreboard in your browser (no key): it reads each game's spread and total from the posted odds. Team PF/PA isn't in that feed and no team has 2026 scoring yet, so Week 1 priors are neutral — set your preseason ratings inline or via Import, and the projection will move off the market. The Claw-Back fires when the spread favorite has a worse 2025 net (PF−PA) than the dog, and its strength scales with market conviction — the favorite's implied win probability from the real posted moneyline when the book has one, otherwise derived from the spread. GRIDIRON: James-Stein shrinkage on PF/PA → expected points (offense × opponent defense + HFA) → Markov drive model → Monte Carlo, ~8k sims, margin variance calibrated to ~17 pts. Cover probabilities feed conviction-based sizing: win probability × 10 = percentage of bankroll. If the live load is blocked by CORS on your host, the verified Week 1 snapshot stays put and you can paste a board via Import. Probabilities are model estimates, not guarantees; nothing here is a lock.
Prototype auth — credentials stay in this browser. Don't reuse a real password.