McKinleyMorganfield

• Updating and significantly revising my forecast on latest data. According to CDC and other sources, cases as of 18 September are (depending on which source you believe) between 1491 and 1497 cases. At the current infection rate, cases are virtually certain to fall between 1500 and 1649--i.e., the third bin. 
• I assign less than 0.5% odds--rounded down to zero--that cases won't reach 1500. To do so, weekly infection rates would have to suddenly plunge to level not seen since 12 January. I'm aware of no plausible reason to suspect they might do this.
• By the same token, to reach or exceed 1650, weekly rates would have to suddenly leap to heights not seen since May. I'm aware of no reason to suspect this, and historical data indicate that rates have only twice since 2022 spiked so pronouncedly, both times in March, suggesting a seasonal effect not applicable now. As with the alternative possibility of sudden near-total cessation of new cases, I assign less than 0.5% probability to such growth in cases, on such an unusual spike, in the time that remains before 30 September. 

• Fixing a cloying oversight in my last forecast, which inexplicably--and in any case, mistakenly--recorded a 1% probability to >1800 cases (i.e., bin 5).
• For reasons laid out in my last forecast, I had assigned a (rounded-up) 100% probability to bin 3 (i.e., to cumulative cases between 1500 and 1649).
• The issue is now essentially already resolved, with CDC currently reporting 1514 cases. Judging especially from historical CDC data, there is no plausible reason to entertain the possibility, in the space of a week, of a potential surge in cases large enough to land us in bin 4 (let alone bin 5). But let the record reflect: I'm not cheating by tweaking my distribution, which now accurately reflects the 100% analysis of my (intended) previous forecast. 
• Thank you ctsats for catching my error!

• Updating/downgrading forecast to less than 0.5% probability, rounded down to zero. As with my just-filed adjusted forecast on India-Pakistan, there is no evident reason to expect such a scenario at this time, especially not at this time, in the still-warm afterglow of Xi-Modi summitry.
• I arrived at < 0.5% odds given (a) low base rate, (b) approaching deadline, and (c)  influence of recent high-level diplomacy. The odds such a trend-defying scenario could unfold within the next 10 days are vanishingly small.

Doh! I just filed an updated forecast...and forgot to adjust my numbers on the board. I'm doing so now, for reasons just explained.

• Updating/downgrading forecast to less than 0.5% probability, rounded down to zero. With no plausible evident predicate to anticipate an outbreak of hostilities--Pakistan's recent signing of a defense pact with Saudi Arabia isn't one, imho--there is vanishingly little time before deadline for (a) something to happen that (b) triggered hostilities that (c) produced 100+ casualties. 
• I continue to assess that the most plausible trigger would be a major terrorist attack, whose incidence makes for a low base rate. I further assess that the chain of events from attack to hostilities to 100+ casualties would take at least several days to unfold. Which by implication means that a triggering terrorist incident big enough to produce a criteria-satisfying outcome would need to happen in the next few days. The odds of such a development on such a timeline are surely below 1%--and imho below even 0.5%.

Adjusting forecast, now rounded down to zero percent. With less than 7 business days remaining before deadline and US military and civilian leadership seemingly focused on issues unrelated to ROK, I assess the odds of an announcement by 30 September to be less than 0.5%

Belatedly moving my forecast to 100% on same reasoning ScottEastman convincingly offers in his updated forecast today. The odds  of two Cat 3 landfalls within the next two weeks, while not impossible, appear so low as to, in effect, round down to zero.

• Adjusting downward my last forecast, based on VTC yesterday with three RFI team members, one of whom (Scott Eastman) usefully pointed out that, even if the next Atlantic tropical depression (AL91) turns into a major hurricane, it would not likely make landfall (if at all) before circa 20 September, i.e. 10 days before deadline.
• AL91 is the only tropical depression currently observed between Africa and the Americas as of 6 September, and NOAA accords it only a 20% probability of cyclonic formation within 7 days.
• The odds, therefore, that a second major hurricane would materialize and hit the US within 10 days of a first, major landfalling hurricane--if indeed AL91 were to even become one--are extremely small, judging from historical NOAA data. I accord generous 4% odds only because experts are forecasting an above-average hurricane season and September is historically an active month. However unlikely, it could theoretically happen. 
• By way of comparison, 4% odds are slightly better than my chances of getting into Harvard, judging by 2025 acceptance data. (Same rationale: unlikely, but could theoretically happen.)

• This is a downward revision to my initial forecast in early June, capturing significantly greater-than-expected EU progress in reducing dependency on Russia for gas than I anticipated from their performance to date.
• At current levels of LNG and especially pipeline gas deliveries, it's looks extremely unlikely that total deliveries for any quarter this year could reach 15 bcm. The combination of EU commitments to end Russian energy dependency, the development of alternative supply infrastructure, and the deterioration of EU-RF relations creates significant barriers to restoring historical trade volumes.
• I accord a slightly higher alternative probability than the crowd (2% versus 1%) to capture the (admittedly small) possibility that a diplomatic breakthrough over Ukraine in the near future could alter the gas picture and enable, in turn, the possibility of 15 bpm in 4Q 2025. Remote, yes; impossible, no.