TD Optimizer
Pick your payout. Build your touchdown parlay.
Searches within 10% of your target return, or shows the closest combinations found.
Customize my picks Model default
Position rules apply to every option, including locked picks.
The TD model remains the baseline. Weights change which picks we favor, not their scoring probabilities. Missing data stays neutral.
One player per game · Finish with GamblyBot
Your TD parlays
Built around your targetYour TD picks will appear here
Choose your target odds, then build. Lock a favorite below or let the model choose.
Lock or exclude players 0 projected players
Projected players are shown here. Players need current market prices and sufficient evidence for their current scoring role. A lock cannot override either requirement.
No matching players.
Why these picks
Explore a player’s evidence now. Build a parlay to review all of its picks here.
Player preview only · This player has not been selected for a parlay. Current prices and injury availability are checked when you build.
Model default
Picks are ranked using the provisional TD model, within your odds target and selection rules. Extra preference weighting is off.
Signals below run from 0 to 100. They are ranking inputs, not scoring probabilities. Missing signals stay neutral. A player can fit a parlay because of the target odds, locks, and one-player-per-game rule, even if another player has stronger individual signals.
Player evidence will appear when projections load.
How reliable is the model?
Historical validation pendingRecorded TD hit rates describe a player’s past. Model accuracy requires forecasts saved before kickoff, compared with actual outcomes. The audit below uses the earliest stored pregame forecast per player and game; it does not test your preference weights or promise a parlay win rate.
Probability accuracy
Compare projected scoring chances with actual TD rates over held-out games.
No measured result yetDo the weights help?
Compare the baseline model with weighted picks under the same odds and leg constraints.
No measured result yetStill to add
Calibrate goal-line and end-zone evidence, model teammate competition, and test injury workload scenarios.
Evidence first; probability changes need validationHow these combinations are selected
We use the median available anytime-TD price across supported sportsbooks as a reference estimate, then rank combinations near your target using our provisional player TD model and your optional preferences. We keep your selected leg count, from 2 to 13. This is a bounded search and does not guarantee the best possible combination.
The model uses projected opportunity and team scoring context. Matchup preference uses overall defense, not position-specific TD defense. Current-role hit rates use verified appearances with comparable lineups, including zero-opportunity games. Players with insufficient comparable evidence are held from automatic selection. Preference weights add a bounded adjustment to the model ranking; they do not represent percentages of the complete model. Missing signals stay neutral. Inside-the-5/10 and end-zone usage are shown as supporting evidence when available; they are not yet calibrated into the TD probabilities. Targets and scoring opportunities respond to the current available lineup and quarterback. Player scored-TD estimates share a team budget. Injury exits and snap restrictions remain uncertain. No win percentage or pricing edge is claimed.
We exclude players marked out or doubtful and games that have started. Reference prices can vary by sportsbook and do not guarantee a combined offer at any one book. GamblyBot matches the chosen picks to available bets. Final odds, availability and settlement rules are confirmed there. Sharing before creating a GamblyBot link saves an estimate snapshot.