Covid 8/25/22: What We Owe

post by Zvi · 2022-08-25T14:40:01.532Z · LW · GW · 3 comments

Contents

  Executive Summary
  The Numbers
  Predictions
  Deaths
  Cases
  Physical World Modeling
    UK’s A&E Crisis
  In Other News
  Monkeypox
  Not Covid
  Bad News
  Modest Proposals
  Good News, Everyone
None
3 comments

A lot of people owe a lot of things. Some of those debts get forgiven, or placed upon others. Other times, new debts and obligations are introduced whether we like it or not. There was a bunch of both going around this week on several fronts. Many systems are not well functioning, which is nothing new, and the fixes look to make things worse, which is also nothing new.

The Covid situation itself is mostly quiet. We have settled into the new normal, with even people whose entire brand is crazy Covid paranoia finding ways to accept that.

It leaves us room to worry about all the other ways things are going wrong. I suspect this week is a new low for percentage of post that is Covid focused.

Executive Summary

1. UK emergency room system is collapsing.
2. Biden uses public purse to pay off student loans.
3. Signs that monkeypox is peaking or will soon.

Let’s run the numbers.

The Numbers

Predictions

Prediction from last week: 560k cases (-8%) and 3,200 deaths (+0%).

Results: 567k cases (-7%) and 3,183 deaths (-1%).

Prediction for next week: 542k cases (-5%) and 3,030 deaths (-5%).

I got lucky this week to bullseye both numbers, so I don’t expect much to happen this week. With everything quiet, small continued declines matching trend seems most likely.

Deaths

Cases

Physical World Modeling

U.S. on verge of being first in the world to approve and roll out a BA.5-specific booster.

The sources I saw for this are saying they are ‘impressed by CDC & FDA’ because we were first to do this. I non-respectfully disagree. We are essentially first to do this because we effectively have bullied everyone else into waiting for our approval on such matters. Impressive would have been production starting a week after we had the sequence. Acceptable would have been three weeks. This is after it’s too late to do much good. I am not impressed.

Claims about Covid testing, for those who are treating Covid testing as a serious attempt at robustly avoiding exposing others like it was 2020.

Covid test expiration dates are being extended over time as more data comes in. Chances are tests still work long after they say they expire.

Dr. David Nabarro, WHO special envoy on Covid-19, says lockdowns were an effective way for governments to buy time to prepare their health systems at the start of the pandemic, but should not have become a default response to outbreaks.

…

“We just couldn’t understand why so many countries were see-sawing between what was sometimes called freedom and lockdown,” he said. “That is not the way. You keep your number of cases down by having really good surveillance and detection and isolation.

“I just thought the narrative was wrong. Why on earth are people always going on about lockdown as though that is the best thing a government can do when, in my judgment, it was a failure of narrative and a failure of policy?
…

“You might want to restrict movement a lot at the beginning because you are trying to reduce the spread of the thing, but the key thing is not to do it for too long,” he said.

…

“There is a middle path between zero COVID and letting the virus just do what it wants. That is, keeping it at bay through good quality public health services but, most importantly, with people and the state being on the same side.

Things that could have been brought to our attention when everyone was being locked down. The WHO head of pandemic response thinks extended lockdown procedures are bad policy, and thought that at the time? So much so that they couldn’t understand what was happening?

And he’s telling us about this now?

It is important not to jump on people when they admit they made mistakes. This is not admission of one’s own mistakes. This is saying that you knew others were making a massive mistake, you were in position to help with that, and you said nothing until it was too late.

If you are the WHO head of pandemic response, and you think many nations are making huge mistakes in their pandemic response that will cause trillions of economic damage for minimal gain, then I think this very much counts as things that could have been brought to my attention yesterday.

Also, I don’t know what that last line means. The same side of what? By doing what? The ‘middle path’ requires a plausible physical path to working.

Zeynep reports generally that reporting on scientific papers has somehow gotten even worse, to the point where she feels the need to read every paper to know what it says.

She doesn’t have good solutions. Reading the whole paper is an expensive option at best, and can only be used sparingly, although reading the abstract often 80-20s that and is relatively cheap. Another compromise is if the author has a Twitter thread explaining key findings, that tends to be quite good and free of misrepresentation.

I try to only read the whole paper if I’m looking for a detailed understanding and willing to write an entire post analyzing what was found.

My only skepticism is the claim that things are getting worse on these fronts, as opposed to always having been awful, although perhaps in slightly different ways.

I would not be wearing a mask to a psychiatric appointment, but still, wow.

Data increasing chance that Covid-19 likely came from the animal markets.

Costs of emergency healthcare collapse in UK as high as 500 lives/week?

Introducing the concept of ‘variant pressure’ which measures the extent to which new variants are taking over at a given time which is predictive of rising case numbers.

This seems like a promising approach. It leads to common sense conclusions while being more legible.

UK’s A&E Crisis

The thread is long and full of insight, so I’m going to pull from quite a lot of it.

It is plausible that many failures of various sorts could be causing excess deaths in various places, and there is always room for improvement, but it is worth noticing that there is no overall special crisis here. The effects overall look a lot like zero. He notices that the Iberian heat wave likely is the main cause in Spain and Portugal.

Link to excess morality report.

Link to the study in question. I haven’t examined it or especially its controls.

John Burn-Murdoch @jburnmurdochYou may have heard almost 30,000 people waited 12 hours in English A&Es in July, but that figure actually only refers to the wait *after* initial assessment Thanks to @Rebeccasmt’s reporting, we know that if you include all time spent waiting, 100,000 people waited 12+ hours

August 23rd 2022116 Retweets361 Likes

The obvious next question is why. Are we overwhelmed with Covid patients? No.

This also seems like a strong case against claims of extensive Long Covid, at least in terms of its impact on acute conditions.

The problem is that the system has no capacity – almost no available beds.

He goes through various general failings of NHS, including bad IT and too little attention to prevention, but these are presumably not new. He then gets to this:

This gets a lot of play, and 2% missing staff is a big problem, but this thread directly points out no increase in hospital visits generally. That doesn’t mean there aren’t more missed sick days, Covid in 2022 has that profile, but I’d posit that a lot of this is likely that everyone involved is overworked and at their breaking points. It’s no surprise they need more sick days. I doubt a third of all days out are Covid.

My summary would be that NHS services have declined in capacity, this is causing long waits, and those long waits are getting people killed with a plausible 500/week point estimate. This is due primarily to delays in services rather than denial of services, so it is mostly an inefficient failure to provide services, which is even worse.

If the true limiting factor is indeed a shortage of beds, it seems like there are ways available to make discharges faster and that should be the priority here.

This post offers additional perspective that boils down to the whole system falling apart, without much explanation of why that happened.

In Other News

Zeynep thread and article on Long Covid. I like the framing of ‘chronic conditions have for a long time not gotten the attention they need, whether or not they have anything to do with Covid.’

“The Trump administration pressured the FDA to authorize unproven treatments for Covid-19 and the first Covid-19 vaccines on an accelerated timeline, according to a House report released Wednesday.”

That saved a lot of lives. If they hadn’t been blocked they would have saved more.

They were not lucky with the vaccine. They had very good reasons to be confident.

So… that’s good. Very good. They know that’s good, right?

Whereas that thing where public health elites pressured Pfizer to slow its protocols? Which killed a lot of people. They know that was bad, right?

Leana Wen, M.D. changes mind, makes sane decisions for self and kids.

She is not saying oops for her previous insanity, such as treating ‘unvaccinated’ kids as being at high risk:

She is proud of her growth:

Note that she explicitly says that masks harmed her child’s language development.

I am so happy to see an advocate of crazy paranoia acknowledge the costs of paranoia.

The whole thread is not Leana saying ‘I was too paranoid and I was wrong.’ She is not only not apologizing to those who acted sane earlier, she is implicitly saying they were blameworthy for doing it too quickly, and only in the last few months should they have been willing to ‘see other families indoors’ for playdates. No mistakes were made here.

It is more ‘I was correctly paranoid before, and I want to explain carefully that I am still paranoid but also decided not to entirely sacrifice the childhoods of my kids, but it’s totally fine if you want to sacrifice your kids’ childhoods this way. And if things were slightly less overdetermined I’d totally be joining you.’

Yeah. Don’t be that guy (better with Gif in original):

Good resource: Chart of potential objections schools might have to installing air filtration units. Not reproducing here due to all the tiny print, there’s a lot here.

2 min video of Scott Gottlieb hitting the CDC for its actions early in the pandemic, nothing you don’t already know.

You mean the most awesome stage of the pandemic, right?

I notice that the couple still wasn’t wearing masks. Right-o.

Monkeypox

Note of correction from last week: When I estimated an 0.03% case fatality rate I forgot to discount recent cases. In a rapidly growing pandemic, this can make a big difference, potentially bringing this answer as high as 0.1%, with the likely answer somewhere in between. My apologies, and thanks to the commenter who caught it.

Monkeypox in decline in New York and Los Angeles. Tail risks have passed.

Even better, appointments for vaccinations are no longer being snapped up instantly.

New York Times proposes masking children in school to protect against monkeypox.

This appears to be real.

Needless to say…

1. Masks don’t prevent monkeypox.
2. Approximately zero kids get monkeypox (17 so far, 0.1% of total cases).
3. Approximately zero people in a school have monkeypox.
4. There is no evidence of transmission in a school.
5. Again: Masks do not meaningfully prevent monkeypox.

For a while, the big risk from Covid-19 has been that we may overreact to it. Is this stupider than every Covid-19 reaction? I think so, because there’s never been this many distinct reasons why an intervention has zero benefits.

1.8 million more doses of vaccine to be available. That should capture the bulk of the benefits, even if it doesn’t fully match demand.

However, there is trouble. Caitlin Rivers finds that instead of the change to intradermal administration meaning everyone had up to five times more doses to give out because they’d get to use each vial three to five times (there are some issues getting the full five 1/5th-size-doses from a vial), instead it was used as justification to divide vaccine allocations by five.

In other words, when the pivot to intradermal administration was announced, a jurisdiction that ordered 1,000 vials thought they would be able to deliver 5,000 vaccinations. Instead, the administration is sending 200 vials. (More on this from the Washington Post.)

That’s not all. I have now heard multiple reports from my state and local colleagues that it is very difficult to extract five doses from a single vaccine vial. Average yield is more like 3 or 4 doses.

…

Taken together, state and local health authorities now have up to one-third fewer doses for use in their communities than they were before the move to intradermal administration.

NYT also reports that the distribution of monkeypox shots is not going so great. One aspect of federal deliveries is a per state limit on number of delivery sites. Two senators is unfair enough, logistically let’s accept that California and Texas are large.

There was a briefing from @UKHSA on Saturday, Kai Kupferschmidt has a thread.

This seems to be excellent news. The worry with Monkeypox is not that the current situation is so bad, it is that if many doublings happened after this, and especially if it were to spread in the broader population, that would be quite bad.

This report is saying that the disease is almost entirely staying within the MSM networks, and is already on the decline overall in the UK. These two facts together make it highly unlikely that the pandemic will grow out of hand. Even in the most vulnerable group, growth has already been contained, and vaccinations and prior cases will grow over time.

While there are a (very small) handful of cases reported that may be from heterosexual spread, they are almost certainly not self-sustaining.

Helen Branswell also has a thread of thoughts on the report.

Celine Gounder has a thread tallying some things we don’t know or might worry about if we were looking to worry public-health style. Points seem valid as far as they go but none of it should alter behavior or presents a challenge to the overall picture.

Not Covid

Scott Alexander reviews What We Owe the Future so I don’t have to. I confirmed not having to via a Twitter poll. I enjoyed Scott’s review far more than the book, and the best part of both was the discussion of population ethics. Could be persuaded to write about that.

Some important words of wisdom from Scott:

An academic once asked me if I was writing a book. I said no, I was able to communicate just fine by blogging. He looked at me like I was a moron, and explained that writing a book isn’t about communicating ideas. Writing a book is an excuse to have a public relations campaign.

If you write a book, you can hire a publicist. They can pitch you to talk shows as So-And-So, Author Of An Upcoming Book. Or to journalists looking for news: “How about reporting on how this guy just published a book?” They can make your book’s title trend on Twitter. Fancy people will start talking about you at parties. Ted will ask you to give one of his talks. Senators will invite you to testify before Congress. The book itself can be lorem ipsum text for all anybody cares. It is a ritual object used to power a media blitz that burns a paragraph or so of text into the collective consciousness.

I did have a publisher act super excited about getting a book of mine published a few months ago at a private event, and she later said she owed me a phone call, so maybe Immoral Mazes (or something else, there are a few other options) will happen after all. My guess is no, nor do I especially want to hire a publicist and go on a public relations campaign. That sounds exhausting. The podcasts sound fun, though. Meditations with Zohar, here I come?

David Shor warns us, the future is going to get weird. There are many reasons to expect this. His is heterosexual identification percentage as function of age, orange is women and blue is men.

I asked because a plausible hypothesis was this was young people being young, and rather than declining with age the percentage is increasing for Millennials over time and radically so for Gen-Z, while we don’t see declines for Gen-X or Baby Boomers.

On the IRS improving its enforcement, AIER argues that poor people cheat on their taxes too, and those people have higher marginal value of income, so that cheating is good, actually, because it involves not reporting income and thus freedom to do work, and it means the rich, who often massively cheat, pay their ‘fair share.’

I do not think this is the slam dunk argument the author thinks it is, but the implications are worth pondering. A response is asking how the tax authorities intend to audit the college student who earned a few hundred working for a moving company, paid in cash? Even if they get audited, good luck proving it. Whereas someone who reports no income at all, or omits a huge portion of their income, yeah, I kind of do want them caught. The obvious counterproposal if you think people earning that little money shouldn’t be taxed so much is to say ‘well then, we should increase enforcement and use the savings to reduce tax rates for the poor.’ Which sounds good to me. Your terms are acceptable.

More concerning are objections like ‘the IRS should not need this much money to provide competent service and functioning systems.’ Which is fair enough.

Bad News

In a time where our biggest concern is high inflation, Biden Administration distributes roughly $250 billion from the public purse to regressively reward a group that is mostly part of his party’s core constituency on the basis of their having not repaid their debts. The official name is ‘student loan forgiveness’ combined with a continued pause on all payments. No plans to reform or change the underlying system.

In case it wasn’t obvious, this will increase the consumption of the favored group, while decreasing the consumption of others (who are on average poorer) that were already on average consuming less.

Some basic impact analysis.

I am suspicious that it would take 50-75bp on fed funds to offset a one time 0.3% price increase that will likely be diffused over time somewhat, depending on how long the hike remains in place.

Not to be outdone by either Biden or the EU, it looks like California might be about to throw the entire internet into chaos and drown it in red tape and concerns over ‘the children’? An analysis of the ‘potential harm to children’ must be filed for every ‘feature’ on every website? And a requirement for age verification of every user that might in practice involve turning over tons of personal data to the company that runs PornHub along with a large windfall, and generally giving tons of info to advertisers among other people? I don’t know if these concerns are real but I do notice that California is exactly this stupid.

They’re also banning the sale of new gasoline cars, I solved for the equilibrium and it is not what they are hoping it will be.

I do want to note that every time someone notes some horrible government attempt to murder what’s left of the open internet and kill the few places left one is allowed to Do Things, when they say “and this horrible rule also has First Amendment issues” you are legally required to take a drink, while remembering that this is good and means the First Amendment is trying to protect us.

Once again, TikTok is Chinese spyware.

Instagram, on the other hand, is American.

Also persistent, perhaps this is not the healthiest thing to be doing.

Twitter’s ‘misinformation’ algorithms are often outputting gibberish labels on tweets that are clearly not misinformation. It has always been thus, for social media censorship is at best imprecise. The good news is Zeynep looked into it, and it seems to be happening in an ‘unbiased’ way across the whole spectrum of viewpoints.

From someone whose job is Google threat analysis:

I am torn whether this is a good or bad sign for Google’s threat assessments.

The rest of you who work at Google, seriously, who thought it was acceptable to make the top of my GMail inbox the same background color as unread posts?

Abomination, please fix. Also get rid of the new menu bar on the left that’s taking up a huge amount of space for no reason, any number of other places you can put that without wasting that kind of screen real estate.

MIT to continue gating access to its buildings.

I am on the side of this being quite expensive and terrible. I also think it’s happening for profoundly dumb reasons. Once such ‘security’ measures are in place, suddenly people keep them for no reason because of the status quo bias effects or inertia or what not, when they would never have thought to do them otherwise. For another example, the building I live in won’t let delivery people go upstairs, whereas pre-pandemic they allowed this, and this is stupid and bad.

In the case of MIT, I am going to go out on a limb based on my experiences at Columbia (and also with the New York rationalist group) and say that almost always a person who chooses to physically go to MIT to hang out there is someone that will benefit the world and also MIT by their hanging out at MIT. That the open exchange of people and ideas is rather important, and losing general access to this kind of resource is profoundly sad.

I urge MIT to reconsider.

That doesn’t mean we can get a good estimate of the effect size. Dustin Moskovitz speculates such folks can instead go to Harvard, which is nearby. That’s a fair point, but also if you could go to hang out at Harvard but chose instead to go to MIT I very much want you to go hang out at MIT.

It is a wonderful aspect of the world that these colleges exist, in the middle of cities, and actual anyone can go and hang out and talk to professors and otherwise explore. Almost no one actually does it, but that’s part of why it is so great. Stop removing the world’s cheat codes.

The bad news is it needs saying, and no I will not be giving the context, but yes.

I’d qualify it a bit by saying that if you are not scrupulously truth-seeking, in whatever example for whatever reason, and you know it, then wondering how your statements could be weaponized politically becomes a larger concern.

Our health care workers are not okay.

Report: 14% of physicians drinking, using controlled substances on the job

…

Forty percent of those respondents reported feeling anxiety or dread about going to work, and 49% said they are either at their breaking point or looking for new work because of the stress and trauma they endure on the job. To cope with these feelings, 17% said they consume alcohol or controlled substances at least once a day.

…

The report also revealed that men feel the stigma around seeking mental health treatment more acutely — male healthcare workers are three times less likely to admit they have a problem compared to their female counterparts. A full 58% of them said they are either at their breaking point or are looking for a new job due to work-related mental health struggles, whereas a lower proportion of of female healthcare workers (45%) responded the same way.

Furthermore, the research showed that male healthcare workers are more than 5 times more likely to use their position in healthcare to obtain controlled substances and 4.5 times more likely to consume alcohol or controlled substances while at work compared with their female counterparts.

These are terrible numbers. The implied male rate here is even worse.

Also the overall alcohol situation is terrible. The first question I always ask when I see such numbers is what is the baseline? For example, this says that for ages 35 and older, 52.5% of Americans drink, 21.3% binge drink and 5.2% are heavy drinkers. That’s not on the job, and uses a different scale, but it isn’t obviously that different from a 17% rate of daily use. This says 40% of Americans drink ‘too much.’ It defines that as 7+ drinks per week for a woman or 14+ for a man, which doesn’t seem fair at all but is also at least a drink a day.

Report that Biden Administration continues to move towards heavily restricting the definition of independent contractors in a way similar to California’s AB5 law. Comments on proposal were heavily negative, but they are not binding. No one seems to have commissioned proper academic analysis aimed at measuring the carnage that would ensue, but I am confident it would be not small, and be echoed at the ballot box. I would urge those involved to reconsider.

How is it that the Fourth Amendment won a court case in Ohio and I not only am sad, I am feeling much less secure in my home as a result?

The logic here is mind boggling. Telling you that you cannot take the test at home is fine. Telling you that you can take the test at home if and only if you allow a scan is unconstitutional.

So, no tests at home, then.

The talk is secure test sites only. The obvious reaction should be, perhaps it is insecure test sites only? As in, at home you have protection against search. In most other places, you do not. So you need to take the test where you have no constitutional expectation of privacy, in order to justify the search.

You don’t have to stay here, but you can’t go home.

I expect this result to be overturned both because of the nature of the current Supreme Court and because it is both very expensive and very dumb. But perhaps Big Education will get its way and stomp out some combination of public universities and distance learning, so private interests can seek more rent.

Also, Bloomberg notes that ‘UK Plans for Blackouts in January in Emergency Energy Plan,’ wait what? My ‘UK has an emergency energy plan that involves blackouts in January’ shirt is raising a lot of questions that were answered by the shirt. Ut oh.

Modest Proposals

Cash transfers to families with new infants (paper). Once again, prioritize getting them the money quickly. For low-income families, getting $1,300 a year earlier, roughly 10% of income, increased adult earnings by the infant by ‘at least 1-2 percent,’ which more than pays for itself in tax revenue alone. Huge if true! This would once again emphasizes that early transfers of cash are highly valuable. Not only do we know they would create additional births more efficiently than later transfers, they have this benefit as well. Also, from Dec 2021, Paul Taylor discussing fertility issues more.

The size of the effect here is staggering. I hope there is further investigation on many fronts. I am skeptical that this effect replicates. Also making a note to check for proper controls because birth year effects exist. I was also amused by the proposed selection effect of parents who know to game the system having better outcomes for other reasons.

If it does check out, this complements nicely the car seat studies that show declining birth rates. If liquidity issues damage child prospects this much, then that (1) clearly overrides any safety concerns regarding lack of car seats and (2) means that we get both benefits from the same policy.

Change the patent rules on drugs to ensure 12 years of market exclusivity (WaPo via MR), so that drugs that have long clinical trials and are difficult (and more expensive) to get approval for don’t get smaller windows of profitability. This sounds very right and is going into my Policy Binder.

Annual eye exams are a scam, because obviously.

This is part of the Implicit Coalition theory of why it is impossible to rollback dumb requirements and regulations. First the come or the orthodontist appointments required for teeth alignment, and I object, because I am an optometrist concerned about required appointments for contact lenses. Or a car seat manufacturer concerned about making 11 year olds have to use car seats. Or a union representative worried that someone might try to repeal the Dredge Act, after which someone might think they could repeal the Jones Act, which I mistakenly think would reduce union jobs because I forgot about the power of the Longshorman’s Union – or perhaps I think they would then, after that, come for the Longshorman’s Union. Or a bank worried they might issue new banking licenses, or an oil company worried we might build nuclear power plants. I could be anyone seeking rent (or also, more generally, anyone with the Maze Nature, although that’s more complicated to explain). So I instinctively, without any explicit communications, back up all my coalition partners in all such situations. Then politicians talk about the horrors of ‘special interests’ once again.

Skin in the game.

Concrete suggestions for that badly needed permitting reform. FAST 41 requires that any potential litigant must raise their NEPA issue during the public comment period if they want to sue for it later, allowing it to be taken into account. Sounds excellent. Other suggestions mostly seems solid as well.

Good News, Everyone

We may or may not be in a recession.

If Biden runs for re-election then we get to send Ted Cruz to Bayes’ Hell.

Via The Money Illusion:

The following statement was published in a China news outlet:

China is struggling to overcome young people’s growing reluctance to start families. Over the past year, the government has launched an unprecedented drive to push couples to have more kids, raising the birth limit and introducing a range of policies to support new parents. So far, however, the measures appear to be having almost no effect.

Governments take note, if you want more of something, and you have a limit on how much of it is being produced, you might want to get rid of that limit as a first step.

New tech to turn hydrogen, water and a steam turbine into electricity?

Good advice on dealing with (or, better, not bothering) very busy people.

Obesity is not protective against Diabetes, says a paper designed to explain selection bias. I love that this exists.

Yes, students can use the bathroom, all they have to do is follow this simple procedure.

Things that cost less to produce will continue to cost less to consume, the angry vegans were rendered immobile, and also less people went to Starbucks that day.

Guessing ‘C’ on every question is now enough to pass the New York State regents exam. If you are going to fail to teach kids algebra, ruining their lives by then keeping them in school by force to keep not teaching them does not seem like The Way, nor did the old standards go high enough to plausibly test for functional algebra knowledge.

It is all a thing of beauty. The old classic of ‘translate some % right to a different % effective score so we can say this somehow means 65%’ is classic, but that’s not all. So much new declining standards technology just dropped. I especially love having a threshold below passing where all you have to do is say the word ‘appeal’ as if you are invoking the Defense Production Act and you automatically pass. There’s also ‘ask questions that you can answer in first grade’ as a bonus.

This is in contrast to Virginia, which did lower its reading standards but made the mistake of giving students a real math test and saw their math pass rates decline from 82 percent in 219 to 66 percent this year. Scores do seem to be recovering.

To answer your question, young man: No. No, you are not.

Game from the Risk World Championships, which I am glad is a thing. It is not especially a thing of beauty and I am left with zero desire to explore further. Still.

Chess players performed worse in tournaments when playing remotely, but this effect decreased over time (via MR). This matches my intuitions as a former Magic: The Gathering professional. One has various routines to get maximum focus, and being at home disrupts them. Then you get new ones.

Once again, the greatest charity in the world is Amazon.com and don’t let anyone tell you different.

I mostly hope so as well.

3 comments

Comments sorted by top scores.