Per @GnatykA:今日南美气温市场出现模型大偏差——布宜诺斯艾利斯 SAEZ 模型预报最高 24–25°C,实测已达 29°C 且仍在上升;圣保罗 SBGR 已达 34°C。预报失灵时,区间价格会剧烈双向波动,盯紧 METAR 实测才是关键。
4
11月12日,两个“稳赢”同一天翻车。Per @polytempbot:安卡拉 14°C+ 早上7点还 89¢,雾和毛毛雨把机场压在 10°C;巴黎 11°C 上午10点 92¢,天一放晴就冲到 12°C。89¢、92¢都不是零风险——同一天、不同城市,相关性才是真正的风险。先看结算站,再看预报。
13
One Polymarket temperature wallet has bought No on outlier highs in 50 cities since March. Per @rcom1337: 5,017 bets, 4,995 paid, 21 missed, +$100K on $2.56M volume, 179 green days, 6 red. Key risk is correlation: one pattern can hit many cities at once.
1
30
Polymarket 气温市场手续费将从 11 月 20 日起下调:taker 费率 0.05→0.01,50¢ 合约每份费用 1.25¢→0.25¢。Per @polytempbot,100 份 50¢ 合约费用为 $1.25→$0.25。费用下降会降低模型入场门槛,但结算口径仍是关键:时间、站点、来源须公开可查。
15
纽约气温市场 11 月 13 日 13:11 发生了什么,交易员仍在等一个公开说法。Per @ShokoyX,截至 11 月 16 日,这个问题还没有得到公开回应。天气市场的结算价就是产品本身:时间、站点、来源都应公开可查。
13
Tokyo is running hot this morning: RJTT already at 20°C, 5°C above its 5-year hour normal. Per @tempterminal, their 7-model blend called 18°C — peak still ahead. 🌡️
1
27
今日要点:Kalshi 次日高温市场在 7 个城市中的 6 个跑赢 NBM,开盘 1 小时后 RMSE 低约 10%,Per @polytutor(Crosier,未经同行评审)。11月14日伦敦结算 10°C,Per @OddsEngine,区间 8/9/10°C。先看结算站,再看预报。
14
London did it again. Nov 13 high settled at 8°C — same 6/7/8°C block, same 78c entry, same 1.28x payout as Nov 10. Per @OddsEngine.
Two 8°C settlements in three days. When a station is this sticky, the block is the trade.
Check the station, then the model.
13
圣保罗 SBGR 又来了。
Per @GnatykA:11月10日预报最高 26–27°C、部分模型还报雨,NO on 30°C 在 98–99¢ 被打穿。
Per @tempterminal:11月12日最高气温结算 28°C,高过 DWD ICON 的 26°C;当天 7 个模型跨度 4.8°C(22–26.8°C)。
98% 不是 0 风险。先看结算站,再看预报。
18
Kalshi temperature markets settle on one named station — and stations just 12 miles apart can differ by 13F. Knowing the station and its quirks is half the edge. Per @polytutor
8
Polymarket changed its weather rules 6 times this year — latest Oct 23: settlement now uses the airport's raw METAR reports. Per @polytempbot. Rule changes are settlement risk, not trivia: track the source, not just the forecast.
11
Miami hit 85°F on Nov 4 — but the weather market paid 82-83°F. Both numbers are correct. The market doesn't settle on 'Miami'. It settles on one named station, and the station can be 2-3°F away from the headline number. That's why we track 76 settlement stations.
6
Per @dizpers: each Polymarket temperature market settles on one named airport station — and the station matters. In Seoul the two airports matched the official result 93% vs 21%. A free site matched 77%; the raw report matched all 329 checked. We track 76 settlement stations.
4
Munich recap: Polymarket settled the daily high at 4°C — 3°C nearly hit. Settlement keys off EDDM routine METAR T-groups; SPECI reports don't count. The thermometer that matters is the one in the rules, not the one on your weather app.
6
Daily-low books bite differently than the highs. You can mark the coldest overnight METAR, then the climate report comes out a degree warmer. Official min isn't always the lowest tick on the tape; a replaced obs can drop out, and some markets still wait for the CLI.
4
US Saturday daily-high books: the dinner METAR max is not settlement. A late SPECI can still lift the day's airport high, and the official number usually only locks when the next-morning CLI prints whole degrees for that ICAO.
5
Chicago is the wider thermometer gap. Kalshi vs Polymarket NYC is ~7.5 km (KNYC vs KLGA); Chicago is Midway vs O'Hare (~24 km). Same city label, different buckets on a hot afternoon — read the ICAO before you size the cross-venue arb.
8
Same city on two venues ≠ same thermometer. Kalshi NYC and Polymarket NYC can settle different stations/series (e.g. KNYC CLI vs KLGA METAR). Read the ICAO in the docs before you size the cross-venue arb.
20
Cheap weather tickets aren’t free money — London/Paris/NYC each name a station and a rounding rule. METAR body vs T-remark tenths vs next-morning CLI. Same degree on the chart, different thermometer in the docs.
7
