Two problems addressed:
1. Favorites' P(1st) was too sharp (e.g. NHL top teams ~20% vs ~12% implied).
- The NHL simulator hardcoded its parity factor (1000) and ignored the
season config's parityFactor, so the knob meant to flatten the
distribution did nothing. It also re-blended raw futures odds into every
game on top of the odds->Elo conversion, double-counting the same signal.
- NHL now reads parityFactor/iterations/seasonGames/overtimeRate from config
and no longer re-blends odds per game (odds enter once, via the central
odds->Elo resolver). Honoring parity 2500 flattens a top team from ~29% to
~13% title odds.
2. "Season Config" and "Input Policy" were two forms over the same stored
object that didn't reflect each other, and the engine-knob half was inert
for many simulators.
- Every simulator now reads its engine knobs (iterations everywhere;
parityFactor for all Elo-based sims) from the merged config, passed in by
the runner via the Simulator interface. Defaults equal the former
hardcoded constants, so behavior is unchanged unless a season overrides.
- The admin simulator page is now a single "Simulator Configuration" card
with structured Engine and Input-derivation sections (profile-driven, so
each sport shows only the knobs it honors) plus an Advanced raw-JSON
escape hatch — all writing the same config.
Also: centralized the duplicated configNumber helpers into config-access.ts;
the central odds->Elo resolver now maps onto the configured Elo floor/ceiling
so those bounds set the odds-derived spread (a real flattening dial).
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01PAFMogMkFJf52YpHyCDvuf
* Add MLS standings sync and fix simulator conference resolution
- New MlsStandingsAdapter using ESPN's free soccer API (no key required),
returning Eastern/Western conference data, W/D/L/GF/GA/GD/PTS, conference
rank, and home/away records; registered under mls_bracket
- Display route now shows soccer table columns and a 9-team playoff cutoff
line for mls_bracket (top 9 per conference qualify for MLS Cup Playoffs)
- MLS simulator gains a third conference fallback: reads externalId="Eastern"
or "Western" on the participant, mirroring the LLWS pool-assignment pattern
so admins can bootstrap conference data before the first standings sync
- 9 unit tests covering stat mapping, conference normalization, rank ordering,
and error paths
https://claude.ai/code/session_01WhzXHpv6taXdHzhgvnv83u
* Address code review feedback on MLS standings + simulator
1. Update mls-simulator.ts module comment to document the new step-3
externalId fallback in the conference resolution order
2. Tighten normalizeConference to exact-match "Eastern"/"Western
Conference" instead of broad substring, preventing false matches
on names like "Northeast"
3. Pre-parse statsMap for each entry before the sort so it isn't
rebuilt O(n log n) times during comparison
4. Document the externalId name-matching tradeoff on
parseConferenceFromExternalId
5. Try ESPN's gamesPlayed stat before falling back to wins+losses+ties
sum; export parseConferenceFromExternalId for direct testing
6. Add tests: normalizeConference passthrough, winPct=0 at preseason,
and parseConferenceFromExternalId (case-insensitivity, null/undefined,
numeric ESPN IDs, unrecognized strings)
https://claude.ai/code/session_01WhzXHpv6taXdHzhgvnv83u
* Fix lint: replace != null with !== null && !== undefined
oxlint enforces eqeqeq; the five != null checks in mls.ts were flagged.
https://claude.ai/code/session_01WhzXHpv6taXdHzhgvnv83u
---------
Co-authored-by: Claude <noreply@anthropic.com>
* Add MLS sport simulator (mls_bracket)
Adds Major League Soccer as a draftable sport with a full-season Monte Carlo
simulator. Models both preseason regular-season projection (34 games across
Eastern and Western conferences) and the MLS Cup Playoffs bracket, including
the Wild Card (single game + PKs), Round 1 best-of-3 series, Conference
Semis/Finals (single game), and MLS Cup.
P1–P8 mapping: MLS Cup winner, finalist, Conference Finals losers, Conference
Semifinals losers. Conference assignment reads from regularSeasonStandings.conference
or falls back to the region simulator input ("Eastern"/"Western").
Admin inputs: projectedTablePoints (primary, max 102 for 34×3), with derivation
chain to sourceElo via existing input-policy; sourceOdds as alternative.
No hardcoded team data — all inputs are admin-managed per season.
- database/schema.ts: add mls_bracket to simulatorTypeEnum
- drizzle/0104_chief_boom_boom.sql: migration for the new enum value
- mls-simulator.ts: MLSSimulator + exported pure helpers for testability
- registry.ts / manifest.ts / simulator-config.ts: register mls_bracket
- mls-simulator.test.ts: 38 unit tests covering all helpers and sync checks
https://claude.ai/code/session_015wkBJ3SYGcMGjsddKKGkwa
* Fix MLS simulator: config loading, sourceOdds fallback, normalization
Three issues found in code review comparing against EPL and NFL simulators:
1. Load simulator config from DB via getSportsSeasonSimulatorConfig so
admins can override iterations, seasonGames, parityFactor, drawRates
per season without code changes. Previously all constants were hardcoded.
2. Add sourceOdds → convertFuturesToElo fallback when no sourceElo is
present. EPL and NFL both do this; MLS was throwing immediately instead
of attempting the odds conversion that the manifest declares as optional.
3. Call normalizeSimulationResultColumns before returning results, matching
EPL's local normalization pattern for consistency (runner also normalizes
globally, but EPL calls it locally too).
https://claude.ai/code/session_015wkBJ3SYGcMGjsddKKGkwa
* Polish MLS simulator: configNumber zero, logger warning, Map lookup
Three small fixes from secondary code review:
1. configNumber: allow value >= 0 (not just > 0) so admins can
legitimately set baseDrawRate or drawDecay to 0 without the value
being silently discarded and replaced with the default.
2. Add logger.warn when a participant is excluded from simulation due
to a missing Elo rating, matching the NLL bracket-aware pattern.
Gives admins a visible signal instead of a silent exclusion.
3. getBySeeds: build a Map once instead of calling Array.find() per
seed. Eliminates 4M linear scans across 50k iterations for a 9-
element array — trivially fast either way, but Map is the right tool.
https://claude.ai/code/session_015wkBJ3SYGcMGjsddKKGkwa
---------
Co-authored-by: Claude <noreply@anthropic.com>
