Robby posted this to Facebook on March 15th, and just updated section 2 and 3 with new information, making this I think currently one of the best guides to how to respond to this whole situation.

If you live in the US, I recommend that you self-quarantine immediately (to the extent that's possible for you) to avoid exposure to COVID-19, the new coronavirus disease. I'll explain why below, then give tips on how to reduce exposure and what to do if you get sick.

Quarantine isn't all-or-nothing, and every little bit helps. Even if you expect to catch COVID-19, you're likely to get sicker if you're exposed to more viral load early on.

(Paul Bohm says "pretty much any viral/bacterial dose study shows that result". Divia Eden: "As I understand it, the virus replicating is an exponential process, and antibody production is an exponential process too. So an early difference in load should make it easier for the latter to outpace the former.")

So even if you can't (e.g.) work from home right now, I would still recommend taking large measures to reduce your exposure.

I expect things to get increasingly dangerous over the next few weeks, so I'd especially recommend taking action now to reduce the amount you're likely to be exposed between end-of-March and end-of-May. By default, risk of exposure doubles every few days until the pandemic peaks.

Disclaimer: I am not a doctor or an epidemiologist.

1. Why I think you should self-quarantine

[Section omitted from the crosspost, since most people appear to understand the importance of self-quarantine by now. The section can still be found here]

2. Ways to reduce risk now:

A. Avoid people.

The CDC believes that COVID-19 is mainly spread person-to-person, "between people who are in close contact with one another (within about 6 feet) through respiratory droplets produced when an infected person coughs or sneezes. These droplets can land in the mouths or noses of people who are nearby or possibly be inhaled into the lungs."

Given that asymptomatic transmission seems common, talking is also probably a common vector — you'd be surprised how much spit flies when people talk.

So my top recommendation is to avoid being near other people (unless you're self-quarantining together), especially in crowded or indoor spaces.

(Added Mar. 16, from Michael Keenan: "Starting at midnight, all non-essential work and movement outside homes is banned in Alameda county and most other bay area counties. Grocery stores and pharmacies will remain open, and restaurants may still do takeout food (and presumably delivery), for those who are still using that." Similar measures may be taken elsewhere in the US soon; be prepared for the possibility.)

B. If you do need to be around people, wear something over your mouth and nose.

Be sure not to touch/adjust the mask/covering (or your face) while wearing it, except to take it off (and throw it away or sanitize it) when you're done using it.

I've heard some people claim masks aren't useful, but this is wrong, provided you're careful to wear them correctly.

(Added: Emily Neff Berlin says in the comments, "Also, please do not buy masks, unless you are in a high risk demographic. They will reduce your risk of exposure some. But they are in short supply and health care/public health institutions need them more than most individuals, as part of work that will ultimately benefit large numbers of people.")

If you don't have surgical masks, home-made masks or scarves can also be effective. This study shows that home-made masks (made from tea cloth) were less effective than surgical masks, but still useful.

(Added Mar. 28: This SSC post reviews the literature and concludes, "If the shortage ends, and wearing a mask is cost-free, I agree with the guidelines from China, Hong Kong, and Japan – consider wearing a mask in high-risk situations like subways or crowded buildings. Wearing masks will not make you invincible, and if you risk compensate even a little it might do more harm than good. Realistically you should be avoiding high-risk situations like subways and crowded buildings as much as you possibly can. But if you have to go in them, yes, most likely a mask will help.")

C. Don’t put coronavirus in your face.

From Patrick LaVictoire:

Like most respiratory diseases, coronavirus is transmitted in water droplets coughed or sneezed out by sick people. If the droplets dry out completely, the virus is destroyed.
They dry out in air pretty quickly- if you’re more than ten feet away, you’re unlikely to inhale them. But on surfaces including metal, glass, and plastic, they can persist for hours or even days. (That’s where the bleach wipes come in; clean your sink handles, especially, in case someone comes in with the virus and washes it off.)
The biggest danger, then [if you're avoiding social contact with others], is that you will touch a surface with coronaviruses on it, and then (before washing your hands thoroughly) touch your mouth, nose, or eyes. Once the virus is in one of those, you’re likely to get sick.
So the number one tip is to learn, now, to stop touching your face with your hands. If your eye itches, at least use your sleeve. Carry a handkerchief in case you need to wipe your nose or mouth. Etc.

PBS says (added Mar. 30):

What’s the best surface for killing viruses? Our skin. In the cases of both flu and cold-causing viruses, infectious particles on our hands are usually gone after 20 minutes.
Between its pH and its porous nature, our body’s natural barrier to the word does a great job at killing viruses, Greatorex explained. “Our hands are quite antimicrobial themselves,” she said. “They have their own bacteria that live on them — no matter how clean you are — and they don’t actually harbour viruses that well.”

So I could imagine it being quite helpful to switch from touching your mucous membranes once every few minutes to doing so once every few hours.

I hear that surface transmission wasn't the primary way SARS spread, so I'm not sure how worth worrying about this is. A lot of people have noted that coronavirus can survive on surfaces up to 9 days in certain conditions (or even longer if it's cold), but Will Eden says: "This paper is the source of the 9 day figure. However that’s the maximum length on any surface under ideal conditions. In some cases it only lasts hours. And it doesn’t establish whether any of that is infectious!"

Regardless, my other three main recommendations are: spend less time touching surfaces that a large number of other people touch (added: or talk near); try to minimize how much you touch your eyes, nose, or mouth; and wash your hands more often, using full medical hand-washing protocol (image; mnemonic; animated version). Note that for COVID-19, hand sanitizer isn't a good replacement for hand-washing.

D. Have emergency supplies.

I would recommend stockpiling at least a month of non-perishable food. 2-3 months is even better. Err on the side of buying food that you like; idiosyncratic favorite foods are less likely to be sold out on Amazon right now, and it's important to make your self-quarantine pleasant so you're less likely to want to take unnecessary risks later.

From residents of the Event Horizon house (https://docs.google.com/…/1Rn6WrTat0Cb6gfatWv-TcSBToh3…/edit): "Most dedicated disaster prep food is already sold out, but you could get a bunch of canned foods (e.g. beans, fish, vegetables, fruits) and dry foods (e.g. rice, pasta, dried fruit, peanut butter) that you like and can eat."

I would similarly stock up on other essential items, especially medications you need.

E. Print out copies of your health records.

From Event Horizon: "If hospital infrastructure is overwhelmed, you may not be able to access this information. The most important information to have on hand is your history of major health problems and surgeries, immunization records, and medications you are currently taking. Here is a template for such a health record: https://docs.google.com/…/1fCCTcwrULo5w4vnymhziCVdctM…/edit…"

F. Disinfect surfaces.

(Especially commonly touched surfaces like door handles and light switches.)

The WHO notes, "In China, human-to-human transmission of the COVID-19 virus is largely occurring in families [...] most clusters (78%-85%) have occurred in families." Divia Eden comments: "One of my takeaways was that delivery of packages was unlikely to be a major vector, in China anyway."

That said, packages may still be the main risk source for you if relying on packages for food; and the risk may rise in contexts where containment measures have failed and the virus is more widespread globally. So if it's not too much trouble for you, or if you're unusually at risk, it probably makes sense to sanitize packages too.

The easy version of this is letting packages sit in direct sunlight for a while before opening them. Heat and UV radiation are good for killing coronavirus. (Added Mar. 28: https://www.nejm.org/doi/full/10.1056/NEJMc2004973 reports, "SARS-CoV-2 was more stable on plastic and stainless steel than on copper and cardboard, and viable virus was detected up to 72 hours after application to these surfaces (Figure 1A), although the virus titer was greatly reduced[....] On copper, no viable SARS-CoV-2 was measured after 4 hours[....] On cardboard, no viable SARS-CoV-2 was measured after 24 hours[.]")

Fortunately, the hard version is only slightly harder, since coronaviruses are easy to kill with a wide variety of cleaning agents. From https://en.wikipedia.org/wiki/Viral_envelope, discussing enveloped viruses in general: "The lipid bilayer envelope of these viruses is relatively sensitive to desiccation, heat, and detergents, therefore these viruses are easier to sterilize than non-enveloped viruses, have limited survival outside host environments, and typically transfer directly from host to host."

Good options for cleaning agents include 70% isopropyl alcohol (NOTE: higher or lower percentages may not work) or the items on this (CDC-referenced) list: https://www.americanchemistry.com/Novel-Coronavirus-Fightin….

G. Buy copper tape from Amazon (https://smile.amazon.com/…/ref=ppx_yo_dt_b_search_asin_title). and use it to cover things you commonly touch, like your cell phone and door handles.

Solid copper doesn't kill viruses on contact, but it makes it hard for viruses to survive on a timescale of hours (https://www.lesswrong.com/…/coronavirus-justified-practical…).

Added Mar. 28: Try to minimize wrinkles in the tape, so that taped surfaces remain easy to disinfect. Robert Miles suggests: "Tape should be cut to the right size, and you need a tool which is finger shaped, harder than a finger but softer than copper - wood or plastic is good, I use the rounded end of a swiss army knife. Running that along seams and small creases makes them mostly disappear". Be careful of papercuts.

H. Probably stop taking NSAIDs like ibuprofen.

NSAIDs are immunosuppressive, and there's some suggestive evidence that they might make cases worse.

Samira Jeimy: "In Germany and France, ICU physicians have noticed that the common thread amongst young patients needing #COVID-19 related ICU admission is that they had been using NSAIDS (Advil, Motrin, Aleve, Aspirin)." She cites https://www.archyde.com/against-covid-19-the-minister-of-h…/, and there's a plausible mechanism (https://www.thelancet.com/…/la…/PIIS2213-2600(20)30116-8.pdf), but obviously the causation might be going the other way and people might be seeing patterns where they don't exist.

The WHO (and some thoughtful lit reviewers like Elizabeth van Nostrand) are very skeptical that the 'NSAIDs are dangerous' thing will turn out to be true once properly studied. Others find it more plausible. I'm not sure how worried to be about NSAIDs, but I expect it to be easy for most people to switch from NSAIDs to acetaminophen and other drugs, so that's what I recommend.

I. Consume 2,000-6,000 IU of Vitamin D daily, in the morning.

"Since COVID usually kills via pneumonia, and insufficient vitamin D appears to be a surprisingly large risk factor in respiratory infection, it’s probably pretty important to keep vitamin D levels sufficient (which in most people means supplementing it specifically, esp if there’s any quarantine that affects food)." (https://www.lesswrong.com/…/coronavirus-justified-practical…)

J. Run an air purifier.

Sarah Constantin: "In a retrospective study in Utah hospitals, high concentrations of particulate matter in the air were associated with slightly but significantly higher (OR = 1.004) rates of admission to the emergency room for pneumonia, and slightly (OR = 1.02) but significantly higher rates of pneumonia mortality.[12] Having air filters in the home may be slightly protective."

Also, quit smoking.

K. Stay healthy: eat well, sleep well, get exercise.

3. Things to do if you get sick:

(Mar. 28 update: I largely rewrote this section today. This section remains relatively speculative and subject to change. COVID-19 is a new disease that we're still in the early stages of understanding, and as Sarah Constantin notes, there aren't many good studies on how best to treat flu- or pneumonia-like illnesses from home. Still, phlegm and flu wait for no man, and we're forced to make the best guesses we can with the available evidence.)

A. Prepare in advance.

Things to buy right now for if you get sick: Pedialyte or gatorade powder, over-the-counter inhalers, a humidifier, acetaminophen, mucinex/guaifenesin, pseudoephedrine, zinc lozenges, oral thermometers, and a finger pulse oximeter. Maybe hydroxychloroquine or chloroquine and/or a home oxygen concentrator, if you can find it online or get it prescribed by your doctor. See below for details.

While still healthy, take your temperature (orally) and use your finger pulse oximeter. Do this multiple times per day over several days, to make sure the devices are working and to get a sense of your baseline numbers.

Figure out who can help take care of you if you get sick.

B. Understand how COVID-19 usually presents and progresses, so you can make an informed guess about how likely you are to have it.

Direct testing can obviously help, but depending on where you are and what symptoms you have, it may be very difficult to get tested in the US, and the test results may come back too late to be useful. It's still worth getting tested if circumstances allow, e.g., to clarify treatment priorities, aid in containment, or check for other infections. Per Sarah Constantin:

It’s important to get tested by a doctor if you have a respiratory illness you think is COVID-19. Even if you can’t get access to a COVID-19 test, you might also have a different bacterial or viral infection (either instead of or in addition to COVID-19) which is treatable by antivirals or antibiotics. It’s very common for viral pneumonia to be complicated by an opportunistic bacterial infection, and killing the bacteria can help improve outcomes.

As of Feb. 24, the WHO listed COVID-19 as causing these symptoms with the following frequencies:

fever (87.9%), dry cough (67.7%), fatigue (38.1%), sputum production (33.4%), shortness of breath (18.6%), sore throat (13.9%), headache (13.6%), myalgia [muscle pain] or arthralgia [joint pain] (14.8%), chills (11.4%), nausea or vomiting (5.0%), nasal congestion (4.8%), diarrhea (3.7%), and hemoptysis [coughing up blood] (0.9%), and conjunctival congestion [inflammation in the white of the eye] (0.8%).

(Also, up to 50% of infections might be asymptomatic, while still being contagious.)

The WHO report continues:

People with COVID-19 generally develop signs and symptoms, including mild respiratory symptoms and fever, on an average of 5-6 days after infection. [...] Using available preliminary data, the median time from onset to clinical recovery for mild cases is approximately 2 weeks and is 3-6 weeks for patients with severe or critical disease. Preliminary data suggests that the time period from onset to the development of severe disease, including hypoxia [serious lack of oxygen], is 1 week.

From Vox:

'Covid-19 disease usually begins with mild fever, dry cough, sore throat and malaise,' writes Megan Murray, a professor of epidemiology at Harvard Medical School, in an FAQ for the Abundance Foundation. 'Unlike the coronavirus infections that cause the common cold, it is not usually associated with a runny nose.' These symptoms emerge five or six days after infection on average, but can show up in as little as a day or as much as two weeks after exposure.

From Business Insider, a day-by-day progression of typical cases:

Day 1: Patients run a fever. They may also experience fatigue, muscle pain, and a dry cough. A small minority may have had diarrhea or nausea one to two days before.
Day 5: Patients may have difficulty breathing — especially if they are older or have a preexisting health condition.
Day 7: This is how long it takes, on average, before patients are admitted to a hospital, according to the Wuhan University study.
Day 8: At this point, patients with severe cases (15%, according to the Chinese CDC) develop acute respiratory distress syndrome, an illness that occurs when fluid builds up the lungs. ARDS is often fatal.
Day 10: If patients have worsening symptoms, this is the time in the disease's progression when they're most likely to be admitted to the ICU. These patients probably have more abdominal pain and appetite loss than patients with milder cases. Only a small fraction die: The current fatality rate hovers at about 2%.
Day 17: On average, people who recover from the virus are discharged from the hospital after 2 1/2 weeks.

Elizabeth van Nostrand has been looking into what the earliest symptoms of COVID-19 infection tend to be. She writes:

Most academic/medical papers start with the person’s first contact with the medical system, which is too late, so I looked at social media and news reports. These are obviously going to be biased towards people with symptoms severe enough to be interesting, but not so severe as to die. I also restricted myself to test-confirmed cases, which because I was also looking at mostly American sources biases things towards severe cases. And I’m counting on people to represent themselves honestly. So there’s a lot going against this sample.
In total I found 11 cases, plus two notes from doctors doing front line work. [...] From this very small and biased sample:
36% of people started with a cough on their first day (55% if you count two people who had very mild symptoms on day 1 and developed a cough on day 2)
64% started with a fever.
18% of people started with both on the same day.
18% started with neither symptom (but developed a cough on day 2)
78% eventually developed a cough
91% eventually developed a fever. The only person who didn’t eventually develop a fever I think might be a false positive, because his symptoms were very weird.
27% had digestive symptoms (mostly nausea)
1/3 of recovered people had been hospitalized. [...]
You may have heard that 80% of cases are mild. Keep in mind that that paper defined mild to include mild pneumonia, which I would classify as at least moderately severe.

From Sky News:

The British Association of Otorhinolaryngology (ENT UK) say asymptomatic patients - ones who do not have a fever or a cough - could show a loss of smell or taste as symptoms after contracting coronavirus. [...] 'In [many] young patients, they do not have any significant symptoms such as the cough and fever, but they may have just the loss of sense of smell and taste, which suggests that these viruses are lodging in the nose.' [...] Dr Nathalie MacDermott, clinical lecturer at King's College London said infections that normally occur through the 'nose or the back of the throat" often lead to a loss in sense of smell and taste, but cautioned that research around the new symptoms for COVID-19 isn't yet widespread in the medical community.'

Stat reports that anywhere from 30% to 67% of COVID-19 patients temporarily lose their sense of smell.

Anecdotally, one of the handful of people I know who likely has COVID-19 (has the relevant symptoms, and is living with someone who tested positive) reports: "given the symptoms I'm seeing, it's a perfect match. Including the on and off nature of it, the chest pains, the shortness of breath, the loss of smell. (I just tried smelling a bottle of rubbing alcohol... smells like nothing! I'm having trouble telling whether my clothes need to be laundered.)" On Mar. 27 I decided to ask five other friends with plausible (but unconfirmed) COVID-19 cases about this; one other said they had extremely diminished sense of smell/taste, two said they did not, and two weren't sure.

One study found higher rates of digestive issues (103 out of 204 cases), broadly construed. 40% of patients experienced lack of appetite, 17% diarrhea, 2% vomiting, 1% abdominal pain. "Moreover, the digestive symptoms among the larger group grew more severe as the severity of COVID-19 increased[.]"

COVID-19 can also present with rashes on the limbs. And the virus can apparently infiltrate cerebrospinal fluid and can cause neurological symptoms.

So... apparently COVID-19 symptoms include "approximately everything".

COVID-19's presentation seems to vary a lot. Typically, if you see symptoms at all, you'll see respiratory symptoms such as coughing, chest tightness, or shortness of breath; and these are the symptoms to keep the closest eye on, since COVID-19 deaths are usually via respiratory failure. But I've also heard accounts of pretty debilitating-sounding fatigue or intense/long-lasting fever. E.g. (from an unconfirmed case):

[...] IT CAN LAST FOR AGES: Onto what I didn’t expect (or hadn’t been prepared for) symptoms wise. The main thing was the sheer length of it – though again, in my own case it was over very rapidly. Isabel wasn’t so lucky. Fevers are nasty, but this was a fever that ground on for days without really breaking or spiking. The duration of C19 seems to be one of the unusual things about it – in Isabel’s case it peaked around day 8 (the worst of the fever), then again on day 10 (the worst of the breathing issues, which luckily were still mild). Apparently for hospitalised survivors in Wuhan the median release day was Day 24. Be prepared for a long haul.
IT EBBS AND FLOWS: This is something I really wish I’d been expecting – for Isabel, C19 didn’t follow a predictable pattern of worsening, peaking and recovery. There was a period of initial but manageable illness, then a rapid decline into being entirely bedridden, then two separate peaks of illness with improvements in between, and only now a sustained recovery. It took her until Day 11 to say “I’m feeling a bit better”, and as an observer there were clear rallies and declines along the way, which made everything particularly nerve-racking: you never quite know if things have improved, and declines are especially frightening.
THE BEST-KNOWN SYMPTOMS MIGHT NOT BE THE WORST: If you’re like me, when you think of Coronavirus you think fever, dry cough and changes in breathing, which are the signature symptoms. Isabel had all of those, but for her the worst effects were exhaustion, nausea, aches and dehydration. She was weakened by it to a degree I’ve never seen outside very old and ailing relatives, well beyond any flu either of us have had. She needed a constant supply of painkillers and water (which she was often too weak to actually lift to her lips) and had zero appetite. In a non-pandemic – or if she was on her own – she would definitely have been in hospital. As it was, fortunately her breathing was never bad enough to need that.

The "ebbs and flows" aspect of COVID-19 seems very common in anecdotal reports. E.g., from another unconfirmed case:

3/7 - I attended a party with at least one known person that tested positive, and >5 others that reported similar symptoms on an almost identical timeline
3/9 - Day 1 - Mild fever (99.5) for a few hours. Chills. Headache. It was gone by the evening
3/10-11 - Days 2-3 - Muscle pain, fatigue - both pretty mild
3/12-14 - Days 4-6 - I felt better, no symptoms at all. This led me to believe it was just a head cold
3/15-16 - Days 7-8 - I had a tightness in my chest and some difficulty breathing. This is when I started suspecting it might not be a cold
3/17 - Day 9 - Cough started. Mild, but dry. Throat is irritated.
3/18 - Day 10 - No more cough, no more chest tightness, no trouble breathing. Throat is slightly irritated but feels like I’m getting better
3/19-20 - Days 11-12 - Uh oh. Fever comes and goes throughout the day, much harder to breath, cough is back and is worse. The shortness of breath was sudden and very unpleasant.
3/21 - Day 13 - Today. Right now I feel better than I did all week. No fever, no difficulty breathing, and no cough. But I’ve been tricked before so I’m being patient and giving it a few more days to be sure. I don’t think I’m quite out of the woods yet. [...]
Well, it definitely wasn't over when I wrote this post on day 13. That night had another wave of chest pain and shortness of breath and days 14 and 15 felt like a low grade asthma attack that wouldn't go away.
Days 16 and 17 - The dry cough turned into a productive cough and the combination of mucinex and hot tea gave me some relief.
Today, day 18, I finally feel mostly ok. I'm not coughing and I can breathe!
Unfortunately, the lab I got tested with was shut down by the FDA so I won't be getting my test result: https://www.nytimes.com/…/coronavirus-home-testing-swab-kit… [...]

Some notes:
1. Don’t trust the 'lull'. This progression seems to be pretty common - one slightly feverish day, then back to normal, and then the second wave. Or third [and fourth], in my case. I wish I had been more aware of this - I exposed some people on days 4-6. So far none of them are showing symptoms, but I would have acted differently if I had known this was a typical covid pattern.
2. It is really useful to have a pulse oximeter. Get one on amazon for <20$. This let me check my vitals even in the middle of the worst part of not being able to breathe and see that my oxygen levels were normal (97-99). I would've made my way to a hospital if it had dropped <90[.]

Symptoms like runny noses are useful to track because their presence provides some evidence that you don't have COVID-19. Colds tend to cause sneezing, whereas COVID-19 and flu don't cause sneezing. Colds also cause runny or stuffy nose, which are rare in COVID-19 (and somewhat uncommon in the flu). And the flu has a sudden onset, whereas COVID-19 and colds usually have gradual onsets.

That said, since COVID-19 is especially dangerous and can present in a variety of ways (and it's always possible you have COVID-19 and another illness), you should be careful not to prematurely rule out the possibility that you have COVID-19.

C. Take zinc immediately if you start feeling any cold-, flu-, or COVID-19-like symptoms.

From virologist James Robb:

Zinc lozenges "have been proven to be effective in blocking coronavirus (and most other viruses) from multiplying in your throat and nasopharynx. Use as directed several times each day when you begin to feel any "cold-like" symptoms beginning. It is best to lie down and let the lozenge dissolve in the back of your throat and nasopharynx. Cold-Eeze lozenges is one brand available, but there are other brands available." (Sources: this + this)

The discussion on LessWrong suggests that ordinary zinc lozenges may be ineffective for various reasons, and that one of the only products that seems likely to be effective is 'Life Extension Enhanced Zinc Lozenges' (which are currently sold out on Amazon).

One option that maybe makes intuitive sense is mixing liquid ionic zinc with water and using it like mouthwash when you start feeling ill? I'm not sure. Be warned, however, that zinc nasal sprays have been found to permanently remove people's sense of smell, and regular use of zinc in other forms may also damage your sense of taste or smell over time.

From Elizabeth van Nostrand:

Take zinc at the first symptom. [...] The method of action is zinc ions attaching to your throat. So pills are useless. It has to be a lozenge. [...] Chris Masterjohn claims only zinc-gluconate and -acetate work. I don't know if this is true, but the only studies I found used gluconate and acetate, so it seems wise to prefer them. [...] The terrible taste and loss of sense of smell says it’s working, unless you already lost your sense of smell to COVID.

D. Start monitoring your oxygen immediately if you develop a fever or experience significant chest tightness or difficulty breathing.

From residents of the Event Horizon house:

Since COVID-19 is a disease with significant respiratory involvement, and that's the pathway by which it usually kills, you'll want to go to the hospital if you’re having trouble breathing, and your blood oxygen levels are consistently below ~90-94% (for people at sea level), while you are using the pulse ox as directed. (Brief dips should generally not worry you, due to normal fluctuations and the imperfection of the measurement.)

(Also from Event Horizon: "[F]or adults, you’re generally considered to have a fever if you have an orally measured body temperature of 100.4°F / 38 °C).")

Eli Morningstar adds:

Advice to avoid freaking yourself out with a home pulse oximeter (avoid user error):
Keep your hand below your heart when measuring
Don't wear nail polish on the finger you're using
Hold something warm first, ie don't have cold hands

Some people are also buying home oxygen concentrators in case they experience serious symptoms while hospitals are overloaded. Connor Flexman discusses home use of both oximeters and concentrators here:

Tl;dr: Not knowing much about this, my current policy is to go to a hospital if PaO2 drops below ~92% and my hospital isn’t completely overrun, unless my PaO2 is naturally low or some other extenuating circumstance. If I was forced to use an oxygen concentrator outside of a hospital, I would target a ~~94-96% PaO2 range, trying very hard to make sure I didn’t hit 99%
If you do have COVID and shortness of breath, when do you go to a hospital?
Hopefully you already have a pulse oximeter as Julia Wise recommends. But sources say anywhere between 90 and 95% PaO2 is the threshold for hospitalization (WHO says <= 93% is classified as severe, ctrl+f “O2”), while other sources say you should threshold on trouble breathing and shortness of breath, not the actual PaO2 number.
It seems to me that using “trouble breathing” as the indicator would track the lung blockages and thus immune response relatively well, while O2 as an indicator would track the danger metric directly (if in fact the primary source of death is insufficient oxygen; if anyone knows this, would be useful).
The benefit of looking at trouble breathing is that it’s an advance indicator. Usually people progress from oxygen therapy to ventilators relatively quickly. If you have naturally low PaO2, your O2 might drop under threshold (say, 93%) in the early stages with mild trouble breathing, but you wouldn’t have much of a dangerous immune response until later. In this case, you’d have wanted to use difficulty breathing as your indicator instead of PaO2.
That being said, having low oxygen seems pretty bad for you, both by common sense and science. For example, 92% or lower is associated with increased morbidity in pneumonia patients; <90% is increased with 36% increased morbidity. Since it’s hard to measure even moderate effects due to the treatment-correlated-with-severity issue, my guess is that there’s some general bodily harm from reduced oxygen even at levels like 95%, though I don’t know how much. So at some PaO2 threshold, I think you want to be supplementing oxygen even if your breathing doesn’t feel that difficult.
Unfortunately, it seems like you can’t supplement oxygen at 95%, because over-oxygenating causes neuronal damage. Standard targets appear to be 94-98% or 92-96%. This study says it seems bad to set your target range during oxygen therapy to greater than 92-96%, because one inevitably exceeds the upper target occasionally. This review/musing muses that it’s a difficult problem, evidence for hyperoxaemia being pretty bad is “comparatively strong”, but not strong enough to warrant especially conservative oxygen titration. Because of these numbers, I think 92-93% is a reasonable threshold to self-hospitalize, since anything above this means they probably shouldn’t be oxygenating you anyways.
If hospitals are overloaded and you have to do oxygen therapy yourself (really try not to do this), I think the targets above are still reasonable, subject to your ability to titrate well with the machine. If you have lots of trouble, of course be conservative. However, you may be able to do better than hospitals: the first study above says that “even in a research setting in the intensive care unit, in which patients receiving mechanical ventilation are closely monitored, most patients who were randomized to an SpO2 target of 90–92% and were receiving supplementary oxygen did not have their inspired oxygen reduced if the SpO2 was 99% or 100%.” So—seems like you could easily do better monitoring than this if you were oxygenating at home. This is why I would probably shoot for 94-96% myself.

I want to emphasize that you may need to go to the hospital on very short notice. Serious COVID-19 cases appear to not-infrequently progress from 'little or no care needed' to 'serious oxygen therapy needed immediately' on very short timescales. E.g., from ProPublica:

I have patients in their early 40s and, yeah, I was kind of shocked. I’m seeing people who look relatively healthy with a minimal health history, and they are completely wiped out, like they’ve been hit by a truck. This is knocking out what should be perfectly fit, healthy people. Patients will be on minimal support, on a little bit of oxygen, and then all of a sudden, they go into complete respiratory arrest, shut down and can’t breathe at all.

The second week of symptom presentation seems like it may be particularly dangerous. Please monitor symptoms carefully, even if your symptoms were relatively mild in the first week. From the CDC:

Some reports suggest the potential for clinical deterioration during the second week of illness. In one report, among patients with confirmed COVID-19 and pneumonia, just over half of patients developed dyspnea [labored breathing] a median of 8 days after illness onset (range: 5–13 days). In another report, the mean time from illness onset to hospital admission with pneumonia was 9 days. Acute respiratory distress syndrome (ARDS) developed in 17–29% of hospitalized patients, and secondary infection developed in 10%. In one report, the median time from symptom onset to ARDS was 8 days.

Depending on how serious your symptoms are, and on how crowded nearby hospitals are at the time, you may want to go to the hospital even if your breathing is OK. If possible, talk to a doctor over the phone/video to get their recommendations. The above is meant to be advice about when you definitely should receive hospital-level care; but it's not necessarily smart to wait until the last moment.

All that said, for several reasons, it's very often a bad idea to go to a hospital during a pandemic when it isn't absolutely necessary:

You might not have COVID-19. Going to a hospital may then expose you to COVID-19, on top of your other illness. (Catching COVID-19 is much more dangerous for people who have another illness.)
You might have COVID-19 and end up exposed to a lot more copies of the virus early in your disease progression, which is likely to seriously worsen your symptoms. Discussion of the importance of viral load here: https://www.facebook.com/robert.wiblin/posts/887350766835. But Nick Tarleton says: "[I] would naively imagine that once someone is solidly {infected + mounting an immune response}, more external viral load would be a pretty small factor."
Hospitals can provide superior care for very severe cases -- and they have the crucial advantage of being able to care at all for patients who require mechanical ventilation. But for more mild symptom management, an overtaxed hospital may provide worse care than you could receive at home, since medical staff's limited attention needs to mostly go to patients with the most serious cases.

Similar advice from Elizabeth van Nostrand: https://www.lesswrong.com/…/what-should-we-do-once-infected…

I suggest monitoring how busy the hospitals nearest to you seem to be as the pandemic progresses, and calling ahead if possible to get advice and information about the availability of ICU beds, etc.

In a shortage, also check whether your local hospital knows that they can split ventilator tubes (and double the air flow) to double their effective number of ventilators. Ventilator splitters can be 3D printed easily if the hospital doesn't have any on hand.

E. Take care of yourself.

From Duncan Sabien:

Symptom mitigation seems extremely likely to be extremely important, especially those related to the respiratory tract. Early prophylactic measures (zinc lozenges, things to soothe throat and prevent coughing), Tylenol (NOT NSAIDs). Lots of fluids with lots of electrolytes (Pedialyte if you can get it, Gatorade if you can't). Temperature regulation.
The biggest thing I'm worried about is permanent post-recovery respiratory problems, and it seems to me that most of those are caused by symptomatic damage rather than by the direct actions of the virus itself.

I believe that most (though not all!) sources these days suspect that fevers are good by default, and you shouldn't try to lower your fever unless it gets dangerously high. Regarding what counts as "dangerously high", Elizabeth van Nostrand says: "I was taught 103ºF, some people say 104ºF, that’s probably the range."

From Event Horizon:

[If] your fever is dangerously high (103°F / 39.4°C or higher) [then you may need to] lower your body temperature manually. Ways to treat fever safely: acetaminophen/paracetamol (do not exceed recommended dose; do NOT use NSAIDs); Drinking more fluids; Take a bath (but NOT an ice bath or a cold bath); Place cool washcloths (NOT icepacks) under the armpits and in the groin area.
[...] Have something to help you rehydrate if you are sick and losing fluids. Gatorade Powder (https://smile.amazon.com/Gatorade-Thirst-Quencher-Powder…/…/) is a good option because you can buy it in bulk, unlike Pedialyte. Also, liking the things that you drink might ensure you keep doing it. Here (https://med.virginia.edu/…/Homemade-Oral-Rehydration-Soluti…) is a guide to creating oral rehydration solution (used to treat fatal diarrhea, also useful for vomiting and general dehydration) from a variety of bases and household ingredients."

From Divia Eden: "I would get mucinex and a humidifier if possible." See the nurse recommendations here. Elizabeth van Nostrand says: "Take pseudoephedrine if you have unpleasant sinus pressure." (Note: not phenylephrine / Sudafed PE, which is apparently useless.)

Stay warm, take it easy, and get lots of rest.

F. Consider taking more extreme measures if you're experiencing serious COVID-19 symptoms, or if you're at-risk (i.e., you're older or have other health issues) or have the resources to take bonus precautions.

As noted above, moderate-to-severe cases may benefit from a home oxygen concentrator. Oxygen concentrators have been used in a much larger number of COVID-19 cases to date than mechanical ventilators, and unlike ventilators, are easy to operate.

If your case is severe but there aren't hospital beds available, postural drainage strikes me as something that might possibly help.

Speculative and risky, but: If you can, you may want to acquire chloroquine or hydroxychloroquine. Be sure to read the entire write-up at https://docs.google.com/…/160RKDODAa-MTORfAqbuc25V8WDkLjqj…/.

Specifically for people in high-risk groups, Anjali Gopal suggests: "can you sign [...] up for clinical trials for antiviral drugs in nearby research hospitals? Some research coming out around this has started to show efficacy of remdesivir, chloroquine, and hydroxychloroquine wrt covid, though of course, this is still all very speculative and needs to be verified by RCTs. (A lot of these trials are often only open for people in high risk groups[.)]" Elizabeth van Nostrand: "One place I have found to find RCTs is https://clinicaltrials.gov/ . Normally I would verify the details of how to join but I assume they are busy right now."