> This is a template.
> The default signal is keyword-based market discovery combined with conviction-based sizing and esports_bias() — three stacked structural edges, no external API required.
> The skill handles all the plumbing (market discovery, trade execution, safeguards). Your agent provides the alpha.
Esports markets are mispriced in two directions simultaneously. Data-rich titles (CS2, LoL, Dota 2) have published Elo models, map win rates, and patch-level performance metrics that retail ignores entirely. At the same time, fan-favourite teams (T1/Faker) are systematically overcrowded by fanbases trading loyalty rather than skill assessment. Three structural edges compound cleanly without any API.
esports_bias() — three layers: game data quality × series format × Asian session timingmax(MIN_TRADE, conviction × bias × MAX_POSITION) — capped at MAX_POSITIONLayer 1 — Game / Market Type
| Game / market type | Multiplier | Key data source retail ignores |
|---|---|---|
| --- | --- | --- |
| T1 / Faker markets | 0.75x | Fandom overcrowds YES by 10–20% vs Elo model — documented 2023–2025 |
| CS2 / Counter-Strike | 1.20x | HLTV.org Elo ratings, map win rates, head-to-head history |
| League of Legends (non-T1) | 1.15x | Oracle's Elixir patch-level stats — meta shifts change team win rates ±15% |
| Dota 2 / The International | 1.15x | OpenDota comprehensive match stats — consistency rewarded in long series |
| Valorant / VCT | 1.10x | VLR.gg agent win rates, map pools — growing and increasingly accurate |
| Mobile esports (HoK, PUBG Mobile, MLBB) | 1.15x | Deep Asian stats with Western info lag |
| Game release date milestone | 1.10x | Publisher delay history documented — ~70% re-delay rate for prior delayers |
| Twitch / streaming peak viewership | 1.10x | TwitchTracker daily historical peaks — viewership growth curves trackable |
| Steam concurrent player milestone | 1.10x | SteamCharts real-time — launch peaks predictable from pre-order velocity |
The T1 / Faker Rule — The most precisely documented single-team overcrowding in all of esports. Faker's global fandom spans every region, every language, every platform. The result is systematic YES overpricing on T1 outcomes by 10–20% relative to what HLTV/Oracle's Elixir Elo models imply. T1 are genuinely elite — but the market price of T1 wins is almost always too high because the fan base is the dominant pricing force, not analysts. This is not a bet against T1 — it is a sizing discipline: trade T1 markets very conservatively.
Layer 2 — Series Format: Variance Reduction by Match Length
This is the cleanest mechanic in the entire trader — no data needed, just understanding how best-of series work:
| Format | Multiplier | Statistical reality |
|---|---|---|
| --- | --- | --- |
| Bo5 / Grand Final / Championship | 1.20x | Stronger team wins ~72–78% — retail says "anything can happen" which is statistically false |
| Bo3 / Playoff / Semifinal / Elimination | 1.10x | Stronger team wins ~65–70% — meaningful variance reduction |
| Bo1 / Group Stage / Swiss / Round Robin | 0.90x | ~40% upset rate — genuine uncertainty, reduce conviction |
The Grand Final insight: retail treats championship matches as the most uncertain because "the stakes are highest." The opposite is true statistically. Teams playing Bo5 Grand Finals have survived multiple elimination rounds — they are the two best teams in the tournament, playing the format that most reliably selects the winner. This is maximum-edge territory, not minimum.
Layer 3 — Asian Session Timing
LoL LCK/LPL, mobile esports, and Dota 2 SEA feature Korean, Chinese, and Southeast Asian teams competing at 01:00–09:00 UTC. Polymarket is US-dominated — match results in these regions take 30–90 minutes to fully reprice when US retail is asleep.
| Condition | Multiplier |
|---|---|
| --- | --- |
| Asian-dominant game + 01:00–09:00 UTC | 1.15x — lag window open |
| All other times | 1.00x |
| Market | Type | Format | Timing | Final bias |
|---|---|---|---|---|
| --- | --- | --- | --- | --- |
| CS2 Bo5 Grand Final | 1.20x | 1.20x | 1.0x | 1.35x cap |
| T1 Bo3 match | 0.75x | 1.10x | 1.0x | 0.83x |
| LoL LCK Bo5 at 04:00 UTC | 1.15x | 1.20x | 1.15x | 1.35x cap |
| Dota 2 Bo1 group stage | 1.15x | 0.90x | 1.0x | 1.04x |
| Any Bo1 group match | type_mult | 0.90x | 1.0x | Edge compressed |
esports, League of Legends, CS2, Counter-Strike, Dota 2, Valorant, Fortnite,
World Championship, tournament, Steam, Twitch, game release, PlayStation,
Xbox, Nintendo, gaming revenue, Riot Games, Blizzard, grand final, bracket,
LCK, LPL, LEC, BLAST, ESL, VCT, The International, HLTV, peak viewers,
concurrent players, T1, Faker, NaVi, Vitality, patch
The skill defaults to paper trading (venue="sim"). Real trades only with --live flag.
| Scenario | Mode | Financial risk |
|---|---|---|
| --- | --- | --- |
python trader.py | Paper (sim) | None |
| Cron / automaton | Paper (sim) | None |
python trader.py --live | Live (polymarket) | Real USDC |
autostart: false and cron: null — nothing runs automatically until you configure it in Simmer UI.
| Variable | Required | Notes |
|---|---|---|
| --- | --- | --- |
SIMMER_API_KEY | Yes | Trading authority. Treat as high-value credential. |
All declared as tunables in clawhub.json and adjustable from the Simmer UI.
| Variable | Default | Purpose |
|---|---|---|
| --- | --- | --- |
SIMMER_MAX_POSITION | 20 | Max USDC per trade — reflects esports market liquidity |
SIMMER_MIN_VOLUME | 3000 | Min market volume filter (USD) |
SIMMER_MAX_SPREAD | 0.10 | Max bid-ask spread (10%) |
SIMMER_MIN_DAYS | 2 | Min days until resolution — tournaments move fast |
SIMMER_MAX_POSITIONS | 10 | Max concurrent open positions |
SIMMER_YES_THRESHOLD | 0.38 | Buy YES if market price ≤ this value |
SIMMER_NO_THRESHOLD | 0.62 | Sell NO if market price ≥ this value |
SIMMER_MIN_TRADE | 5 | Floor for any trade (min USDC regardless of conviction) |
simmer-sdk by Simmer Markets (SpartanLabsXyz)
共 2 个版本