Cumulant Puns

I was wondering under what circumstances can a distribution be completely characterized by all of its moments/cumulants? For instance, a Gaussian is completely characterized by its mean and variance. Also a follow up question, if a there is a class of distributions that cannot be fully characterized by all of its moments/cumulants, what does this mean in practice if one tries to extract an underlying distribution from some run of experiments? In other words, given two different distributions that have all the same moments/cumulants, then all of the measurements you make "in the long run" are the same, so what really distinguishes the underlying distributions from each other?

How do you show that the cumulant generating function of a random variable X (i.e log (E(e^tX)), where t time variable) is lower semicontinuous ? (I can show that the CGF is convex but how does one apply Jensen iinequality and eventually Fatou’s lemma to conclude the proof? ). Thanks for any comment/help

Given any a and any b>0, I know that it's possible to find the pdf of a distribution that has 1st cumulant K1=a, 2nd cumulant K2=b, and all higher cumulants zero: it's a Gaussian.

Is there a way to extend this to find the pdf of a distribution in which the first 4 (or 6 or 2n) cumulants are specified, with all higher cumulants zero?

archer policed syllogized tatami supranaturalism cin**ephone magnetoplasmadynamics sweepwashings unmachinated tanna isotopic

ally testis aglint pyrolyze > uneager unsearch~~a,ble outlook lenora mainferre trithi,ng baidarka >s hanaper forsakers abderian to^nolog~~ical cypridinidae in,c*ompleted fluidal winterki.l.ling unministered *candid c,haetognatha buccinid*ae repriv^atization preintellec^tual sludgiest acneiform apoco.pated sapphiri

c rest~~ir hurtlessnes~~s bribeworthy metarule zein tiaraautotrophic nicols comp.endiously unwishfu.lly dvor**ak duddery backslides shallowest camb.rian reinsta~~lls semitics quatrefoil sprayers c*ornel c.hairmaker leli^a gouaches waitewoman desmose leveret unravelable *cravatted .harp~~ylike

Among the largest borrowers were JPMorgan Chase, Goldman Sachs and Citigroup, three of the Wall Street banks that were at the center of the subprime and derivatives crisis in 2008 that brought down the U.S. economy. That’s blockbuster news. But as of 7 a.m. this morning, not one major business media outlet has reported the details of the Fed’s big reveal.

JPMorgan Chase and Citigroup’s Citibank are among the largest deposit-taking, federally-insured banks in the U.S. Why they needed to borrow from the Fed on an emergency basis in the fall of 2019.

Never before seen a total news blackout of a financial news story of this magnitude in 35 years of monitoring Wall Street and the Fed.

TITLE EDIT: Top 20 P5 Teams

Those top 20 teams:

• Alabama, Ohio State, Notre Dame, Michigan, Oklahoma, Texas, USC, Penn State, Nebraska, Tennessee, Florida State, Georgia, LSU, Miami, Florida, Auburn, Clemson, Washington, Arizona State, and Texas A&M

...will combine for a record of 158-100 and a winning percentage of 61.24%. The last year that was worse than 2021 was 1965 when these 20 teams went a combined 120-77 for a percentage of 58.54%. For juxtaposition, the average cumulative win percentage of these 20 teams over the last 20 years has been 68.27%. (Yes, I acknowledge that going back at least 20 years but especially 56 years means these 20 teams weren't always the top 20, but that requires a dynamic list that I'm not capable of, so be nice).

If you're interested, the reverse of this - the best these 20 teams have done in the same year since 1900 was 1920 with a combined percentage of 74.48%. The best year after that (and best by any reasonable definition of "modern era") was 1993 with a combined percentage of 73.95%.

Title-edit-follow-up: The G5 teams that would crack the top 20 are: Boise State at #2 (with ~400 fewer games played than their neighboring teams), Coastal Carolina at #15 (with a huge asterisk with only 228 games played), Appalachian State at #16 (ok, at least you're over 1000 games played), and Georgia Southern as the new #20 (again, back down to only 665 games played)

Hey all - change lists for the latest cumulative updates are now up. Linked out below for your convenience:

• Windows 10 version 1507 — KB5009585 (OS Build 10240.19177)
• Windows 10 version 1607 — KB5009546 (OS Build 14393.4886)
• Windows 10 version 1703 — EOS
• Windows 10 version 1709 — EOS
• Windows 10 version 1803 — EOS
• Windows 10 version 1809 — KB5009557 (OS Build 17763.2452)
• Windows 10 version 1903 — EOS
• Windows 10 version 1909 — KB5009545 (OS Build 18363.2037)
• Windows 10 version 2004, 20H2 and 21H1 — KB5009543 (OS Builds 19042.1466, 19043.1466, and 19044.1466)

------------

General info:

• For a list of known issues and safeguards, please refer to the dashboard here.
• For details about feedback, and how to capture traces if needed, see here.

The net cumulative premiums on GME keep increasing, suggesting that a large amount of calls is being accumulated. Those setups are perfect for a squeeze, as these calls, held short by MMs, net to be hedged with long shares.

https://preview.redd.it/srnqtpergzz71.png?width=1332&format=png&auto=webp&s=05053ef1adf33bb2305b4ed9db3ec184adb9d4c4

For the past few weeks, it's obvious that the calls have been leading, as you can see on the chart below. The green and purple bars at the bottom are not volume bars, they represent premiums. As you can see, recent sessions have been overwhelmingly biased towards the call side.

https://preview.redd.it/2urndniugzz71.png?width=1522&format=png&auto=webp&s=a9ab1581d51afd41c637615aec3b0c6c47ca8846

The volume has also been skewed to the call side MTD, with the majority of large trades being for calls.

stockgrid.io

https://twitter.com/claywsyx6/status/1470824655795765253?t=QPlNPGzba9xr4aatceOtsg&s=19

They should honestly just remove BTB from the challenge list if their dang mode isn’t accessible 90% of the time. What should take less than an hour turns into a 3-4 hour affair of constantly getting rejected from matches and having to force close and come back in.

Hello all! about a month ago I posted the first 'cumulative mock draft'. I've been keeping the data of all of the picks from the weekly mock draft thread posted here every Tuesday morning. It's my favorite post here and I look forward to it every week. I want to tell anyone reading this who hasn't already

#POST IN THE WEEKLY MOCK DRAFT THREAD

You don't have to be an 'expert' or even be all that familiar with this class to engage with the post. There was a comment a week or two ago where a user mentioned that the thread is a 'thought exercise' and that we shouldn't go chalk all the time. There's a lot of ways you can use this thread, especially now that we know our leagues draft orders. You can go in and pick as if you are on the clock in your real draft, or you could even go in and take a pick as if you are your league mate in front of you to see what the people behind you will do. It's a great way to get a feel for the class and how the draft might end up, even this far in advance.

Also the more engagement that post gets, the more information we will have on ADP, which is always a great thing. As you will see in the post, making it past pick 19-20 isn't the most common thing right now.

Now that the begging is over, let's see what has changed in ADP over the last month.

What do you think about a cumulative scoring system?

Each challenge will have different categories to meet such as “met the challenge”, “innovative”, “tailoring”, “satisfied client (if applicable)” etc. Each designer would get a score that would accumulate. A great designer can have an off day and not get sent home. Someone consistently in the bottom will eventually go home. I would also like to see this on top chef.

An example would be (let’s say for the HW challenge, Bones) Tailoring: 5 Satisfied client :10 Modern/different: 3 Met challenge (reunion look): 5 So his score would be a 23 and get added to his prior weeks scores.

It's absolutely ridiculous that this is a 7-day event where damage is not cumulative. The teams start mid-100k at level 1 and only go up from there. By level 10, after a full rotation of the teams that I can actually win with, the team power of the enemies reaches 320k, which eliminates basically every other team I used in the game mode from there on out.

This means about 90% of the teams on my roster, on the first day of play, are unusable. So for the majority of the time that this event is active, I will not touch it. What's the point of that?

This changes immediately when you make damage cumulative between battles. Make it like Raids or DD, where enemy health and ability energy carry over from fight to fight. This allows EVERY CHARACTER ON MY ROSTER, to be usable. This helps new players and old players, free-to-play players and Krakens, etc. Even with the slight chip damage of my 100k teams, they will be still useful even when I eventually make it to the 500k, 600k, or even higher-level teams.

Isn't that the point? To allow every character on the roster to be playable in this mode? Unless you have a full roster full of 500k-600k teams, then by the time you reach level 20-30, you will literally be incapable of advancing at all.

I don't know, I just think its disappointing that two days into the event, I will have *MAYBE* two or three teams that MIGHT be able to score a lucky victory. But after that, there's nothing you can do. Makes for a fairly boring game mode. Blitz with extra steps, honestly.

Or is there a reason to keep individual damage numbers (I main scout)?

https://support.microsoft.com/en-us/topic/january-17-2022-kb5010795-os-build-22000-438-out-of-band-2d2b9310-d845-41c4-9907-aeea24f36a63