Will resolve based on NOAA global monthly report. Usually gets released in the middle of the following month. A tie will resolve “no”.
Edit: Selling some NO per manifolio ( still think it’s a much lower chance though )
@JaimeSantaCruz Are you sure about those 2k limit orders? I'm happy to take it all but not if it was a typo....
@JaredAsh Last September anomaly was record breaking (1.43C for NOAA).
My own analysis use the mixed data sets in this flow: GEFS long range forecast (6 hour intervals) to fill in the ERA5 (pulse) data with a forecast until the end of the month (a linear model, minimizing the MSE for GEFS->ERA5), and then using that data with another linear model to match NOAA's records.
A middle, point estimate from a run a couple days ago implies 1.21 C (implying a chance of 3%), while the NOAA average from last year is 1.43 C. This is a 0.22 C difference, which is fairly large. It's hard to estimate the uncertainty with the above workflow but I estimate the average error should be roughly ~0.1 C (maybe a few hundredths more or less). This puts a point estimate for the error to break the record at about twice (or more) this amount (<5%).
@JaredAsh my analysis is much more basic than JRPs.
We're coming off El Nino, when in last September we were fully into it, breaking all past records by large margins (every single day last september was a record high for that day, mostly by like half a degree). It was really unusual, to put it mildly. This would mean to me the chance of coming in at a record temperature this September should be much less than 50/50, I'd bet it down to 20% at the beginning of the month and let those who do the fancier analysis argue over how low to go below that. I also let the graphical representation of the ERA5 dataset inform my bets: https://climatereanalyzer.org/clim/t2_daily/?dm_id=world and so far September is coming in cooler than a year ago, and for every day that continues, it becomes less likely the month will break a record (be careful to factor in sea surface temperatures too, but note the scale - you can get them from the menus at that link). So at this point if the probability was still at 20% I'd be willing to bet it down to say 10, unless I was betting against JRP.
Edit to add: my prior starts out at 50/50 for this month, before adjusting as above, because last month's NOAA report came in at 0.01 degrees celsius above August 2023. So, basically equal, August almost wasn't a record. And the "this is the xth record month in a row" records have been by a smaller margin each month for the past few months, so actually my real starting prior would be slightly lower than 50/50.
Edit: deleted
Had to update after discovering I was using the wrong NOAA dataset 😢
Acceleration bailing me out a bit...
@PaperBoy No worries! I just go so fast that I sometimes don't look at the description, so helpful to have it in the title. Thanks!