Polymarket Bundle Dota2 Props Trader

Bundle Dota 2 Props Trader

This is a template. The default 签名al 检测s action-score inconsistencies across correlated Dota 2 match props -- remix it with OpenDota API data, hero draft analysis, or live match feeds. The 技能 handles all the plumbing (market discovery, prop parsing, bundle grouping, inconsistency 检测ion, trade execution, safe防护s). Your 代理 provides the alpha.

Strategy Overview

A single Dota 2 match on Polymarket spawns 28+ prop markets:

"Total Kills Over/Under 47.5" = 58% "Total Kills Over/Under 50.5" = 52% "机器人h Teams Beat Roshan?" = 30% "Will Any Barracks Be Destroyed?" = 45% "Will There Be a Rampage?" = 15% "First Blood Before 3:00?" = 62% "Will There Be an Ultra Kill?" = 22% "Will the Game End in Daytime?" = 40%

These props are fundamentally correlated. High-kill games feature more teamfights, which means more roshan contests, more barracks pushes, more multi-kill sprees. When the kills O/U market implies a high-action game but another prop implies low action, the bundle is internally inconsistent -- and one of them is wrong.

Edge

Polymarket prices each prop in isolation via its own order book. There is no mechanism to enforce cross-prop consistency the way a professional sportsbook would. This 创建s 系统atic mispricings:

Correlated props priced independently -- RetAIl traders bet on "Roshan?" without 检查ing what the kills market implies about game action level Action-score anchor -- The kills O/U market is the most liquid and best-priced prop; it serves as a reliable anchor for what kind of game to expect Thin liquidity on exotic props -- Rampage, ultra kill, and daytime markets have much less volume, making them more prone to mispricing Multi-game series multiply opportunities -- BO3 matches have separate prop bundles for Game 1, Game 2, and Game 3, tripling the surface area 签名al 记录ic Discover active Dota 2 prop markets via 获取_markets(limit=200) as primary discovery (Dota markets often missed by find_markets) + keyword 搜索 supplement 解析 each question to 提取: match_key, game_number, prop_type, prop_value Prop types recognized: kills_ou (with threshold), roshan, barracks, rampage, first_blood, ultra_kill, daytime Group by (match_key, game_number) into bundles Compute action score from kills O/U markets (weighted average, higher thresholds weighted more) For each boolean prop, compute expected probability given the action score: High-action props (roshan, barracks, rampage, ultra_kill, first_blood): should be HIGH when action score is HIGH Low-action props (daytime): should be LOW when action score is HIGH (high-action games extend into night phases) Trade props where actual probability diverges from expected by more than MIN_INCONSISTENCY (default 10%) Conviction 扩展s with inconsistency magnitude (30% gap = full conviction) Remix 签名al Ideas OpenDota API: Pull hero picks for the match draft -- teamfight-heavy drafts (Enigma, Magnus, Tidehunter) predict higher kill totals and more action; split-push drafts (Nature's Prophet, Tinker) predict lower kills but faster barracks destruction Team aggression 性能分析s: Some teams (e.g. Spirit, Tundra) historically play high-kill games averaging 55+ kills/game; others (e.g. GAImin Gladiators) play disciplined low-kill styles averaging 40-45 kills. Weight the action score by team 身份 补丁 meta analysis: Major Dota 补丁es shift average game length and kill totals. After 7.36 (teamfight meta), average kills jumped 15%; after 7.35 (laning meta), they dropped 10%. 追踪 补丁 timing to adjust the action-score baseline Live kill feed: During a BO3, if Game 1 had 65 kills, Game 2 and Game 3 action expectations should shift upward -- the teams are in an aggressive mood. Wire in live Game 1 结果s to adjust Game 2/3 prop expectations Cross-match consistency: If two different matches from the same tournament have wildly different roshan probabilities despite similar team 技能 levels, one is likely mispriced Safety & Execution Mode

The 技能 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

auto启动: false and cron: null mean nothing 运行s automatically until 配置d in Simmer UI.

Required 凭证s Variable Required Notes SIMMER_API_KEY Yes Trading authority. Treat as a high-value 凭证. Tunables (Risk Parameters)

All declared as tunables in ClawHub.json and adjustable from the Simmer UI.

Variable Default Purpose SIMMER_MAX_POSITION 40 Max USDC per trade at full conviction SIMMER_MIN_TRADE 5 Floor for any trade SIMMER_MIN_VOLUME 3000 Min market volume 过滤器 (USD) SIMMER_MAX_SPREAD 0.10 Max bid-ask spread SIMMER_MIN_DAYS 0 Min days until resolution (0 = allow same-day) SIMMER_MAX_POSITIONS 10 Max concurrent open positions SIMMER_YES_THRESHOLD 0.38 Buy YES only if market probability <= this SIMMER_NO_THRESHOLD 0.62 Sell NO