Skip to content

House model: NFL totals

Paper bot

Bets on NFL totals when the model and market meet the published selection policy.

Forward test9 trades recorded · 8 settled

Simulated paper trading only — orders are not sent to Polymarket.

Model vs market picks →

House model: NFL totals returned −1.1% from its first trade on 24 Oct. Buying the S&P 500 returned −1.1%, measured from its 23 Oct base close. −0.04 pts vs the S&P 500.

All recorded dates · 24 Oct – 16 Nov

1M and 3M unavailable: only 22 days recorded so far.

What $100 in House model: NFL totals became, against the S&P 500The chart starts with the first trade on 24 Oct. The S&P 500 comparison uses its 23 Oct base close, the last trading close before that trade. 8 trade marks.What $100 became$96$98$100$10225 Oct8 NovLost on 27 Oct 2026 · $20.29 · Cowboys vs. Eagles: O/U 44.5Won on 1 Nov 2026 · $20.30 · Ravens vs. Bills: O/U 52.5Won on 2 Nov 2026 · $20.29 · Patriots vs. Dolphins: O/U 44.5Won on 8 Nov 2026 · $20.27 · Lions vs. Dolphins: O/U 50.5Lost on 9 Nov 2026 · $20.31 · Packers vs. Patriots: O/U 46.5Lost on 9 Nov 2026 · $20.29 · Buccaneers vs. Bears: O/U 46.5Won on 15 Nov 2026 · $20.27 · Vikings vs. Packers: O/U 42.5Lost on 15 Nov 2026 · $20.28 · Texans vs. Browns: O/U 40.5S&P 500−1.1%Strategy−1.1%
Inspect a date with the arrow keys, or drag across the chart.
Drawdown for House model: NFL totalsHow far House model: NFL totals fell from its own running peak between 24 Oct 2026 and 16 Nov 2026; worst −2.8% on 16 Nov 2026.Below its own peak0%−2.8%Worst−2.8%
The strategy line is green when it is ahead of the S&P 500 and red when it is behind; amber is the same $100 in the S&P 500 over the same dates. Selected ranges rebase both lines within the selected dates. The drawdown band uses recorded chart observations and keeps earlier running peaks. The facts below cover the lifetime record. Marks: 0 entries, 4 wins, 4 losses. Hollow circles are raw BUY fills; filled green circles are settled wins; red diamonds are settled losses. Each sits on the recorded equity at or before its timestamp, not a fill-time valuation. Marks show recorded entries and settlements in this chart only; lifetime totals are in the facts and log.
Return
−1.1%Lifetime · from $1,000.00
vs S&P 500
0.0 ptsLifetime · S&P base close 23 OctS&P −1.1% from that close
Equity
$988.75Paper account value
Trades
98 settled
Max drawdown
5.6%Lifetime · worst fall from a peak
Days live
23Since the paper record began

The bet is selective. The edge is uncertain.

The rule was chosen before the live record, and it decides which published picks become paper bets. The backtest is evidence, not a promise.

Bets the under in windy outdoor stadiums when the price is between 42¢ and 58¢ (the rule's past hit rate gives at least 2% expected profit), at most 4 a week.

Test ROI +8.1% across 104 test bets. The 95% interval is −9.4% to +25.8%. That range includes losing money.

All-edge baseline ROI −4.7%.

Backtest prices: closing line (nflverse), vig included. A backtest uses historical prices; the paper record above uses live quotes.

The points model currently has zero weight: this policy bets a pre-registered base rate, not the model's chance.

The model publishes a chance for every game and is graded on every game, whether or not this policy places a bet. A published pick is a prediction; a paper bet is an execution.

Every published pick and its grade →
Show policy methodology

A live test of a classic handicapping rule: wind hurts passing and kicking, so this bot bets the under in outdoor stadiums that are windy at that time of year (average wind of at least 10 mph in past seasons) and skips everything else. Because Polymarket prices carry little vig, it will bet nearly every windy-stadium under priced 0.42–0.58, capped at 4 a week. At bookmakers' closing odds the rule made +6.1% on 2010–22 (432 bets) and +7.6% on 2023–25 (101 bets), which is not statistically significant.

Periods
train 2010-2022 · validate none (pre-registered rule, no tuning) · test 2023-2025
Prices
closing line (nflverse), vig included
Test bets
104 · hit 56.9%
Test ROI
+8.1% (95% CI −9.4% to +25.8%)
Every ≥4¢ edge
−4.7%
Probability
blends the model (weight 0.00) with the market price (weight 1.00)
Stake
$20.00 a bet
Version
2026-10-02

In plain words

TL;DR

Bets NFL unders in windy outdoor stadiums at 42–58¢, at most 4 a week (a base-rate rule; the points model currently has zero weight).

How it bets

The model estimates how many points both teams score together and the chance the total goes over each line. Polymarket lists several lines per game; the bot only trades the main one (priced closest to 50¢), once per game; then see "Why 4¢" below.

When this bot bets

  1. Before each game the model publishes its chance for every side, next to the market's price.
  2. Bets the under in windy outdoor stadiums when the price is between 42¢ and 58¢ (the rule's past hit rate gives at least 2% expected profit), at most 4 a week.
  3. $20 per bet, best expected value first; held until the game ends; nothing within 10 minutes of kickoff.
  4. Backtest (2023-2025): ROI +8.1% on about 35 bets a season, vs -4.7% betting every edge.

Exact rules

Bets the under in windy outdoor stadiums when the price is between 42¢ and 58¢ (the rule's past hit rate gives at least 2% expected profit), at most 4 a week. $20 per bet, best expected value first; held until the game ends; nothing within 10 minutes of kickoff.

Bets

Record 4–4Realized −$10.35Unrealized −$0.90Open 1Best win +$19.70Worst loss −$20.31

Price is per share; a share pays $1 if it wins. The fee is Polymarket's taker fee for the whole bet. Polymarket links include our referral code.

Signed amounts show gains and losses. Open amounts are unrealized; settled results retain their recorded status.

9 settled and open bets on this record.

Show raw fills · 17 recorded

All 17 recorded fills available on this page are shown (up to 100).

  • Dolphins vs. Bills: O/U 46.5

    $20.00

    notional

    Under 46.5 at 0.500 · 40 shares · $0.50 fee

  • Texans vs. Browns: O/U 40.5

    $0.00

    notional

    SETTLE at 0.000 · 37.03703703703704 shares · $0.00 fee

  • Vikings vs. Packers: O/U 42.5

    $36.36

    notional

    SETTLE at 1.000 · 36.36363636363636 shares · $0.00 fee

  • Texans vs. Browns: O/U 40.5

    $20.00

    notional

    Under 40.5 at 0.540 · 37.03703703703704 shares · $0.28 fee

  • Vikings vs. Packers: O/U 42.5

    $20.00

    notional

    Under 42.5 at 0.550 · 36.36363636363636 shares · $0.27 fee

  • Buccaneers vs. Bears: O/U 46.5

    $0.00

    notional

    SETTLE at 0.000 · 38.46153846153846 shares · $0.00 fee

  • Packers vs. Patriots: O/U 46.5

    $0.00

    notional

    SETTLE at 0.000 · 40.816326530612244 shares · $0.00 fee

  • Lions vs. Dolphins: O/U 50.5

    $36.36

    notional

    SETTLE at 1.000 · 36.36363636363636 shares · $0.00 fee

  • Lions vs. Dolphins: O/U 50.5

    $20.00

    notional

    Under 50.5 at 0.550 · 36.36363636363636 shares · $0.27 fee

  • Packers vs. Patriots: O/U 46.5

    $20.00

    notional

    Under 46.5 at 0.490 · 40.816326530612244 shares · $0.31 fee

  • Buccaneers vs. Bears: O/U 46.5

    $20.00

    notional

    Under 46.5 at 0.520 · 38.46153846153846 shares · $0.29 fee

  • Patriots vs. Dolphins: O/U 44.5

    $39.22

    notional

    SETTLE at 1.000 · 39.21568627450981 shares · $0.00 fee

  • Ravens vs. Bills: O/U 52.5

    $40.00

    notional

    SETTLE at 1.000 · 40 shares · $0.00 fee

  • Cowboys vs. Eagles: O/U 44.5

    $0.00

    notional

    SETTLE at 0.000 · 38.53564547206165 shares · $0.00 fee

  • Ravens vs. Bills: O/U 52.5

    $20.00

    notional

    Under 52.5 at 0.500 · 40 shares · $0.30 fee

  • Patriots vs. Dolphins: O/U 44.5

    $20.00

    notional

    Under 44.5 at 0.510 · 39.21568627450981 shares · $0.29 fee

  • Cowboys vs. Eagles: O/U 44.5

    $20.00

    notional

    Under 44.5 at 0.519 · 38.53564547206165 shares · $0.29 fee

Last hourly run

Considered 1 · Trades 0 · Skipped 1

Skips are normal when the market fails the bot's published filter, the price is outside its band, the book is empty, or the full stake cannot fill.