Process

Methodology

Factor lenses Don uses to interrogate a number — public betting as one input among many, not an auto-fade — and the safeguard that keeps a hot ATS sample from pretending to be a system.

Critical safeguard
Hot ATS samples are not a system
A 17–5 run against the spread is a sample, not a process. Don does not treat short-window ATS records, heater narratives, or public-side heat as evidence. Edge is a pre-committed factor lens plus a number that still looks wrong after the close. If that is missing, the official decision is NO BET.

Process

  1. 01At most one official release per calendar day. Most days should be NO BET. The game itself may kick off later.
  2. 02Daily Pick means the wager Don released that calendar day — not necessarily a game played that day.
  3. 03Default stake is a flat 1% of CURRENT bankroll (about 1u at a 100u start). Never chase losses or raise size to recover.
  4. 04Drawdown safeguards can set desk status ACTIVE, REVIEW, or PAUSED. That is a human desk control, not auto-trading.
  5. 05Do not auto-rewrite the model on a losing streak. Methodology changes are a new modelVersion, tested before they replace production.
  6. 06A pick requires a number Don can defend and, after the close, a logged CLV.
  7. 07If the close does not confirm an edge, the day is a pass — not a delayed bet.
  8. 08Evaluate CFB/NFL week lines as soon as markets open. If the number is exceptional vs fair and waiting risks killing the value, issue BET NOW.
  9. 09Early-week bets need stronger evidence. Injuries, lineups, weather, and rest are still unresolved — write that uncertainty down.
  10. 10An early official pick is an OPEN POSITION until graded. Do not issue the same game again unless it is a genuinely separate wager.
  11. 11Monitor the market after issuance. Never change the issued line, odds, book, or timestamp. Playability can become MARGINAL or NO LONGER PLAYABLE without altering the ticket.
  12. 12Grade after the event. Do not rewrite the ticket.
  13. 13Public betting, handle, and line movement are one input among many. They may strengthen or weaken an independently identified edge. They never create a bet by themselves, and Don does not auto-fade the public.

Public betting policy

Public betting is one input among many
Public tickets, handle, and line movement contextualize an independently identified edge. They never create a wager by themselves, and Don does not automatically fade a high public percentage.
InputRoleInfluence
Ticket / bettor %Who is on each sideWeak context. One-sided tickets are investigated, not faded automatically. Store the raw % for later 60–69 / 70–79 / 80%+ bucket analysis.
Money / handle %Where the dollars sitModerate context. A large money% − ticket% split is a research flag, not a bet. Do not assume the “sharp” money side is correct.
Opening vs current linePrice discovery and CLV processStronger as market evidence once a fair number already exists. Direction and magnitude are stored; waiting vs BET NOW is still a number decision.
Reverse line movementLine moves against the publicInvestigate only. RLM is tagged for later ROI/CLV study. It is not a reason to bet, and fading the public is not assumed profitable.
Sportsbook / market consensusConsensus and best available priceUsed in the odds layer (fair vs consensus vs best vs issued). Consensus does not override Don’s fair; it is a tape, not a pick.
  • Identify an independent edge first (fair vs market). Public betting may strengthen, weaken, or contextualize that edge.
  • Never auto-fade because a high % of bettors are on one side.
  • Never create an official wager from tickets, handle, or RLM alone.
  • When data is reliable, store RAW percentages — not just “public” or “sharp” labels.
  • Do not fabricate bet% or money%. The Odds API supplies prices, not ticket/handle splits. A separate provider is required later.
  • The research assistant writes the judgment in reasons / officialQualification. The ledger only stores the snapshot.

The research assistant supplies judgment in reasons and official qualification. This site stores raw percentages and tags for later ROI / win / CLV analysis. It does not auto-bet.

Factor lenses

NFL
NFL
Situational football first. The market is sharp; Don only bets when the number still looks wrong after the usual adjustments.

  • Rest

    Bye, short week, Thursday, and extra-day spots versus what is already in the price.

  • Divisional dogs

    Familiarity and compressed talent gaps. Not an auto-fade of favorites.

  • Travel

    West-to-east, altitude, and long road swings when the line has not fully moved.

  • Weather

    Wind and precipitation that actually change totals or kicking — not a cloudy forecast.

  • QB / injury

    Confirmed starters and usage, not rumor-board noise an hour before kick.

  • Public betting

    Tickets and handle are context, not a bet. Don does not auto-fade one-sided public. Raw % and line movement are stored when a real source exists.

  • Home / road

    True home-field effects after rest and travel, not a blanket 2.5.

  • Coaching

    Script, clock, and fourth-down aggression — only when it is stable and priced wrong.

CFB
College football
Talent and variance dominate. Don treats CFB as a market that overfits last week and underfits roster construction.

  • Returning production

    Experience on both sides of the ball, not a single returning starter headline.

  • Talent gaps

    Recruiting / roster quality versus the spread. Big gaps still get eaten by bad numbers.

  • Explosive plays

    How each side creates and prevents chunk plays; drives totals and live variance.

  • Pace

    Plays available and whether the total is a pace bet in disguise.

  • Conference / non-conference

    Information quality differs. Non-con results are noisy inputs, not conclusions.

  • Look-ahead

    Spots before a rivalry or ranked test. Soft cover, not an automatic fade.

  • Post-rivalry

    Emotional hangover is real and often already in the number. Check the close.

  • Ranked / unranked

    Public bias into logos. Respect is not the same as a verified edge.

  • Public betting

    Heavy public on a logo is a research flag, not an automatic opposite-side wager. Store tickets/handle when available.

NBA
NBA
Rest and availability move this market more than narratives. Don does not bet a side because a team is 'due.'

  • Rest disadvantage

    Days off mismatches that are not fully paid for in the spread or total.

  • Back-to-backs

    Second night of a B2B, especially with travel. Minutes and blowout risk matter.

  • Travel

    Density of the trip, altitude, and time-zone direction.

  • Lineups

    Confirmed out/questionable and how the rotation actually closes games.

  • Pace mismatches

    Possessions available versus the total. Stylistic clashes, not box-score leftovers.

  • Market movement

    Injury-driven moves versus opinion. Don needs a number, not a headline.

  • Public betting

    National sides and parlays can crowd tickets. That does not create an NBA bet by itself.

MLB
MLB
Starting pitching still leads. One-game samples are noise; Don needs a stack of small, priced-wrong edges.

  • Starting pitcher

    True talent, platoon splits, and recent usage — not last three starts ATS.

  • Bullpen

    Leverage arms available, rest, and whether the total is a bullpen bet.

  • Platoon

    Lineup vs handedness. Only when the projected lineup is real.

  • Park / weather

    Wind, temperature, and park factors that change run expectancy.

  • Lineup

    Confirmed bats, not a projected card from the morning.

  • Umpire

    Only when the umpire profile is meaningful to the total or K markets. Usually it is not.

  • Public betting

    Public ML favorites are common. Investigate ticket/money splits; do not fade them automatically.