https://x.com/NWSPhoenix/status/1836205206557868339 enough to resolve (only 93F, same entity).

Phoenix meteorologist saying the streak is over https://x.com/AmberSullins/status/1836180739727200722 will wait for the day to show up at https://www.weather.gov/wrh/climate?wfo=psr presumably tomorrow to resolve, unless it has a big surprise.

https://x.com/NWSPhoenix/status/1836066968673382891

Forecast highs for today are 6-12 degrees below normal. Yes, you heard that right, BELOW NORMAL. Phoenix's streak of 100+ °F days is expected to finally come to an end today, a streak that has been going for 113 days (~3.5 months).

High was 102 today, streak up to 113, predicted to be broken (high under 100) tomorrow https://x.com/NWSPhoenix/status/1835850060061618371

https://ktar.com/story/5604584/100-degree-days-phoenix-streak-nears-end/

Just noticed X from what I think is the data source https://x.com/NWSPhoenix

Note the high yesterday (2024-09-15) according to https://www.weather.gov/wrh/climate?wfo=psr was 106F, like 6 degrees higher than hour by hour forecasts were showing. I'm not sure if this is typical, hopefully serious traders have investigated and are trading accordingly. Anyway the streak has reached 112 days. Will it continue today?

I've been trading in this question as I said I would in the description, however I won't trade after the data for a new day has been released at the URL above if it indicates the streak has been broken as that feels like cheating. Of course others are welcome to before I notice and resolve!

Note the 100th consecutive day was September 3 (I incorrectly stated September 2 in the description; updated). The first day was May 27. Today (September 11) the 108th.

Anyone can look at the forecast but here's a writeup of when it might break anway https://www.azcentral.com/story/news/local/phoenix-weather/2024/09/10/when-will-phoenix-drop-below-100-degrees/75143812007/

I created this in part to see how the "experimental" numeric market type works, so I'm going to trade in it.

It's not clear to me that it's possible to buy "lower"; either option seems to result in "higher" for chosen range. Though I might be misunderstanding something!

Ah, "higher" and "lower" just preselect ranges above and below current expected value. There doesn't seem to be any way to buy down a particular range, other than by buying up other ranges. Though I might be misunderstanding something else!

Hmm, the "Mike Linksvayerbought Ṁ6 108-109 YES2m" seems to show an implementation detail. I last bought "Ṁ99 of 108 - 120 from 126.00 to 125.56" and 99/13 is...well 7 with a remainder. So though I suspect an implementation detail is showing, I also wonder whether there's an off by one. Also probably I'm misunderstanding something.

Test.

Another implementation detail leak "92 more trades" (I've made 5, spanning a range of 89 by my eyeball math...hmm 89+5 = 92+2?)

Also noticing the "holders" tab lists the number of traders holding each increment (ranges of 1 in this case, eg 102-103, 103-104, etc). That feels more like that's just how the market works rather than a showing an implementation detail that doesn't strike me as useful to expose in the way it is (such as the above 92 more trades thing).

Not specific to the numeric result market type, but in case a Manifold person happens to read this thread, another random suggestion: turn down the opacity of Source: manifold.markets so it doesn't obscure potentially important parts of the chart, eg

Another odd thing, perhaps about this market type: sell is all or nothing for a given value -- sell all your holdings or not. Unless I'm missing something.

You're always trading a range, like 102-103 rather than discrete values. That makes perfect sense where the thing being predicted is continuous. It's a little unintuitive when what is being predicted is discrete, like n days of a streak here.

Thinking about resolving 50% to both sides of the final streak when it ends, so if the steak is 130, 50% to both 129-130 and 130-131. However when you buy it's always

Lower bound <= value < upper bound

So maybe in above scenario 130-131 should get 100%, 129-130 nothing.

@mods any guidance on how a numeric question with a resolution value touched by two ranges should be resolved? The higher range, due to how buying is specified (see above)?

Thanks for making this experimental market type available!

Another thing in favor of going with the higher range: that's what's shown in the graph; eg 113-114 is shown as 113:

when you trade, you see this:

so the integer '113' corresponds to the option '113', and you'd pick only that. that's the way the market type is intended to work

@jacksonpolack thanks for the confirmation. Idle feedback: it might be more intuitive (well, would be for me) to always display something like "113 <= to < 114" where now in the chart you see "113" and in holdings, trades, resolution, perhaps elsewhere you see "113-114".

ps while not a numeric market I see the creator of https://manifold.markets/yaakovgrunsfeld/how-many-hezbollah-operatives-were used "lower <= x <= upper" to make the answers clear (though not completely consistently, precisely 2000 matches none).

How many Hezbollah Operatives were killed by the pager operation

https://www.jpost.com/breaking-news/article-820536 Israel apparently managed to kill or injure a significant number of hezbollah operatives there are conflicting reports as to how many operatives were killed I will resolve according to Wikipedia's estimate of number killed, they need not be confirmed (so if wikipedia says approximately 50 killed it will resolved to 10->100) If the wikipedia page for this incident doesn't include an estimate (or if there is no wikipedia page) I will resolve to the best estimates of generally reliable sources such as the NYT, WSJ, Jerusalem Post or other sources I consider reliable. If there are conflicting numbers numbers, I'll prioritize according to the order I listed here (that is to say if the NYT and WSJ give different numbers I'll resolve acc to the NYT) I will wait until at least new years 2025 to resolve to give some time for the fog of war to clear regardless of if numbers are published beforehand, if there is still no clear info I'll wait until new years jan 2026 If it turns out that the operation was carried out by someone other than Israel that won't affect my resolution similarly, if some other israeli target were killed they will be included in the count (eg. PIJ Hamas IRGC)

@MikeLinksvayer note to self, noticed numeric markets were disabled a few days ago, assuming https://github.com/manifoldmarkets/manifold/commit/5a40fa7f88a89eacd17ea4201a1ec1d1a09e3120 corresponds to what's on the site. Looks like they've been enable/disabled/iterated on a few times, which is great.

disable numeric markets · manifoldmarkets/manifold@5a40fa7

Manifold Markets: A market for every question. Contribute to manifoldmarkets/manifold development by creating an account on GitHub.

Couldn't find a question on the current streak.

Topically related (2024) Phoenix seems very likely to "win" but it's just highest for any one day this ~summer:

Questions about heath deaths in [Phoenix] Arizona in future years:

Also in the other direction (cold) for 2024 Phoenix there's