diff --git a/live/daily.sh b/live/daily.sh index 30068968..3f59de98 100755 --- a/live/daily.sh +++ b/live/daily.sh @@ -37,7 +37,7 @@ python3 dashboard.py mkdir -p history && cp watch_skilled.json "history/watch_$(date '+%Y%m%d').json" 2>/dev/null echo "[daily] $(date '+%F %T') 6/6 publish (commit + push refreshed outputs)" PUBLISH="no changes" -git add watch_skilled.json watch_sharps.json conviction_wallets.json 2>/dev/null +git add watch_skilled.json watch_sharps.json conviction_wallets.json dashboard.html 2>/dev/null if git diff --cached --quiet 2>/dev/null; then echo "[daily] no output changes to publish" elif git commit -q -m "live: daily refresh — skilled + sharp wallets [skip ci]"; then diff --git a/live/dashboard.html b/live/dashboard.html index 80716a42..192b4d97 100644 --- a/live/dashboard.html +++ b/live/dashboard.html @@ -36,13 +36,13 @@ td.name{font-weight:600}a{color:var(--blue);text-decoration:none}a:hover{text-de .b{color:var(--green)}.s{color:var(--red)}

Skilled Wallets · Polymarket

-

107 wallets that beat their own entry prices and held up out-of-sample · generated 2026-06-16 07:37

+

61 wallets that beat their own entry prices and held up out-of-sample · generated 2026-06-23 10:33

-
Validated wallets
107
-
Median z
4.0
-
Median OOS z
3.0
-
Value / longshot
69
-
Favorite-rider
21
+
Validated wallets
61
+
Median z
4.2
+
Median OOS z
3.2
+
Value / longshot
37
+
Favorite-rider
12
@@ -68,7 +68,7 @@ td.name{font-weight:600}a{color:var(--blue);text-decoration:none}a:hover{text-de

How to read it. z = how far the wallet beat the prices it paid (skill; >3 is strong). OOS z = whether that skill held on held-out bets — the gate that separates real edge from luck. Alpha 0.2–0.4 = share of bets in the underdog value zone where research finds real edge concentrates. Value wallets win by beating longshot/mid prices (the copyable alpha); Favorite-rider wallets post high win rates buying near-certain favorites (real but thin, hard to copy). Click any row for live recent trades. Not financial advice — split-half OOS, forward-track before sizing.