What is the COVID-19 Case Fatality Rate (Death Rate)?

Nobody knows for sure what the Case Fatality Rate for Covid-19 is in the general population. (It changes based upon the demographics of the population and the quality of healthcare in the region, and can not be accurately calculated until a large number of people have fully recovered.) Currently the US estimate is hovering around 6% of those infected die.

Estimates from various countries have ranged from 0.07% in Singapore to 6% in the United States. Because many cases may be un-diagnosed this 6% estimate is likely to be high. (In Italy it was 14%.)

But due to the involuntary experiment currently spreading through the prisons of Ohio, along with good reporting by the state, we can generate an estimate of the CFR within the Ohio prison system.

On May 22nd 2020 when I downloaded DRCCOVID-19Information from the https://coronavirus.ohio.gov/ website, the corrections system had 4090 prisoners recovered from COVID-19, with 61 deaths (and 277 inmates still infected).  If we discount the inmates who are still infected (not yet recovered) and only look at those who have recovered (or died) we get a case fatality rate of 61 / (61 + 4090) = 61 / 4151 = .01469525 or 1.469 percent (you can round that up to 1.5%)

Of course, if 200 of those 277 “not yet recovered” are currently on ventilators, this CFR number will go up in the future….but this are the numbers we have to work with today.

So 1.5% of Ohio prisoners who were infected with Covid-19 have died.  (Ohio has 35,464 prisoners in isolation, so approximately 11.7% of them have been infected).

This CFR is 15 times higher than that of the seasonal flu (0.1%).  Of course, because prisoners are not generally representative of the US population (the demographics of prisoners skew young, while oldest adults are more frequently killed by covid-19) this CFR will not be the same as for the general population. But it does give us a ballpark number to think about for a lower limit.

If we take 1.5% CFA times 11.7% infected times the 330 million population of the US, we would get 579,150 deaths. If we let it spread to 60% of the population, that would be 2.97 million deaths.

Of course, other research has estimated the CFR in the US to be hovering around 6%.
But that is likely due to missing many un-diagnosed cases because of lack of testing.
So I would use the 1.5% as a lower bound on the CFR in the general population.

Luckily, most Americans are not housed in prisons and can social distance to more effectively prevent the spread of covid-19, so we have not gotten anywhere near 11.7% of the population infected. Yet.

It is also uncertain if prisoners get better healthcare in prison than the general population, but my guess would be no (at least for the 89% of people who could afford health insurance before the pandemic hit).

Related evidence:

  • A related scientific paper with similar numbers can be found here.
  • According to the Marshal Project, on May 24th 2020, at least 415 inmates have died in the US, with at least 29,000 infected (which is a CFR of 1.43%).


Schwarzenegger voice: It’s not a flu!

Covid-19 is a novel coronavirus, and is not a flu (influenza). It is more contagious than the flu, and more deadly than the 2009-2010 N1H1 swine flu.

For example, the first case of the H1N1 flu in America was detected on April 15th 2009.  86 days later, on June 10th 2009, there were 211 recorded deaths (37,246 cases).

The first case of Covid-19 in America was detected on January 20th 2020. 86 days later (April 15th, 2020) there were 32,900 deaths (654,425 cases).

In fact, the CDC’s best estimates place the total death toll of the H1N1 flu for the entire year (April 12, 2009 to April 10, 2010) in the range of 12,469 deaths (8,868-18,306).

In summary, the (86 day) fatalities of covid-19 is already three times higher than the full year (365 day) fatalities from H1N1, even with the physical distancing and safer at home policies attempting to stop transmission.  This is not #JustAnotherFlu.

Why only 100,000 deaths would be amazing

National officials have stated that we may see 100,000-200,000 deaths from Covid-19. I hope the death toll will be lower, but worry it may be much higher.

From April 12, 2009 to April 10, 2010, CDC estimated there were 60.8 million cases (range: 43.3-89.3 million), 274,304 hospitalizations (range: 195,086-402,719), and 12,469 deaths (range: 8868-18,306) in the United States due to the (H1N1)pdm09 virus. – https://www.cdc.gov/flu/pandemic-resources/2009-h1n1-pandemic.html

These numbers indicate that the H1N1 (Swine flu) had a case fatality ratio (CFR) in the US of between .04 and .009 percent. Note that 1 in 3 adults over 60 had antibodies to H1N1, perhaps from an exposure in their childhood.

The World Health Organization reported that 1 in 5 people (20-27%) were infected by H1N1 in the fist year, and “Based on an estimate of around 200,000 deaths, they said the case fatality ratio was probably less than 0.02 percent.” – https://www.reuters.com/article/us-flu-h1n1-pandemic-idUSBRE90O0T720130125

In England the CFR was calculated to be 0.026 percent.

The contagiousness of H1N1 (R0) is estimated to be between 1.4 and 1.6, which means that every infected person infects 1.4 to 1.6 other people. (As long as this number is above 1.0, the growth rate is exponential.)

The Bad News

Reports from China suggest that with good medical support (i.e. hospitals not overloaded) the case fatality rate for COVID-19 is as low as 0.7 percent. [The CFR in Italy where hospitals are overloaded is currently in the 3.7 – 11 percent range.]

The contagiousness of COVID-19 is estimated to be as high as 2.7 (every infected person infects 2.7 others) with no stay-at-home orders. Seattle is one of the first cities in the US to implement physical distancing procedures and the growth rate there has slowed. One model predicts that the R0 has been reduced to 1.4 by the physical distancing implemented in Washington State, which is much better than 2.7, but still about as infectious as the H1N1 flu.

If COVID-19 infects 60.8 million people in the US, and has a CFR of 0.7 percent, that is 425,600 deaths.  [Of course, if it infects 60 million people the case fatality rate may jump as all of our hospitals will be overwhelmed, because about 10-15% of infections normally result in hospitalization.] So an R0 of 1.4 is not good enough.

As a point of comparison, the CDC estimates that smoking results in more than 480,000 deaths in the USA every year. (1 in 5 deaths annually.)

The only way to prevent this is to reduce the contagiousness with extreme physical distancing. In China they were able to bring the R0 number below 1, allowing the growth rate to fizzle out with severe quarantine methods. South Korea was able to achieve a similar result with ample testing and contact tracing / containment.


Related news stories & Updates

The Privacy Preserving Anti-Coronavirus app Google can and should build

Many people already trust Google with their location data (unless you proactively turned off location history on your android phone and Google maps, Google knows where you have been and how long you spent there.)  Obviously, all of these people are trusting Google to follow it’s privacy policy and not release this information. [Google is not alone here, the major mobile phone carriers also have all this data, and you can’t turn that location tracking off….]

What happens when you start to cough, get a fever, and (can hopefully) get tested for the Coronavirus? If you have a positive result, the local health department should be asking you who you interacted with in the last 5-10 days, and then contact THOSE people to tell them they should self-isolate and get tested if they come down with symptoms.  Presumably you’ll tell them about everybody you remember interacting with, but you might forget that 3 days ago you paid for gas at the counter of the local 7-11, or not know the name of that guy that traded you his Aldi cart for your virus covered quarter so he wouldn’t have to return it to the cart stand to get his own quarter deposit back.

Here is where Google can help everybody. Once you have received a positive diagnosis, your  local health department sends Google your gmail address (google account name).  Google emails you asking for your permission to share your location history for the last 5-10 days with the Health Department.  If you give permission, Google will share that data with the health department, who now knows that you spent 8:10-8:20am last Friday at the 7-11 and 4:30-5pm at Aldi grocery shopping on Tuesday. Having this location history can potentially help the health department, but what they really need is to be able to contact the people you interacted with.

Google can help with this by cross-checking your location history with the location history of everyone else who uses an Android phone (50-70% of the population). People who may have crossed your path could be sent a proactive email letting them know that “From 8:10-8:20am on Friday a person who has subsequently tested positive for coronavirus visited the same 7-11 that you visited from 8:15-8:25am”.  No need to share your name, just the fact that exposure was possible. Those people could be given the option to share their contact information with the local health department by clicking a link. This would allow the local health department to sort by risk, and they might choose to make a followup phone call to those people who spent a considerable amount of time in the same locations as you. (They probably wouldn’t have the manpower to call everyone that was at the same gas station, but if you forgot about the monthly bookclub meeting where 10 people’s phones spent 2 hours near your phone, this could be incredibly useful information for them to have….)

The key privacy preserving idea is that sharing your location data would be optional.
If you chose to share your location history, Google would anonymously notify those people who may have intersected with your travel history, and then THEY would have the choice to share their contact information (or not) with the health department (not you). Most people (who have not turned off location history) already “trust” google with their location history, so proactively cross-checking to find intersections with anonymous notifications should not be seen as a major a privacy violation.

Obviously, it would be scary to receive an email from Google telling you that you might have used the same gas-pump as somebody infected with the coronavirus, and there would certainly be a LOT of false positives (If the cross-checking algorithm was too aggressive, it might pair up everybody who stopped at the same stoplight at the same time in their commute, for example….).  I trust that the UX (User Experience) people (and lawyers) at Google could write the emails in a suitably non-threatening way “Out of an abundance of caution, we wanted to let you know that there is a small possibility that you shared the same location as a person who later tested positive for the coronavirus.”

I argue that a little bit of scaring would be a net social benefit. If you were receiving a daily email with a list of the times and places you were near an infected person, social distancing would go up immediately. If Google is more worried about scaring away their users than the global good, they could make a website that people would have to deliberately visit to find out if any coronavirus positive intersections were found for their location history, but this would remove significant portions of the benefit.

I’m sure that many smart people inside Google have already come up with this idea and there are internal debates about privacy and public relations going on internally between them and the lawyers. We should tell Google that this is a perfect time to “Not be Evil”.



Singapore has created an application like this, but it requires that everybody install the same application to trade encrypted contact information. So either the government makes it a requirement that you run the application to leave your house, or it is not as effective.

MIT has released an app that can log your location history completely privately, but it does not collect contact information from others, and has no way to cross-correlate with other’s tracking history.

Apple & Google are working together on ways for iPhones and Android devices to work natively with contact tracing applications.

COVID-19 / Coronavirus Resources / Reading List

Stuck at home, I’ve spent too much time reading the internet about the Coronavirus. Below are the most useful resources I have found for understanding what is going on and what is likely to come in the future.

If you only have 8 minutes, this video is the best overview I’ve found:


Status Tracking

Why Social Distancing?

What needs to improve?

Actions you should take

About Covid-19 / Coronavirus

Timeline Back to Normal


Older links that are still relevant:


Project Source 11″ LED Integrated Light Failure (Flashing) IPX8011LS-ORB

Two and a half years ago I purchased a 2 pack of LED integrated lights for my house from Lowes. They were the IPX8011LS-ORB units, and cost $50 for the two pack. Recently, the electronics in one failed such that the light would flash on and off quickly (2-4 times a second) instead of saying lit continuously once the electronics puck heated up. Here are some photos of the old/bad unit (click to enlarge).

Some of the model numbers I found in the old unit: C041218 (18W LED Driver Triac dimming)  BG0358-110C036-03 (on the circuit board) HH-03 94V0 E327405 4515 (also on the circuit board.

Luckily, it came with a 3 year warranty and I was able to find my receipt, so I took it to Lowes and they replaced it with one that looked exactly the same on the outside, but had a different item number on the box and a completely different set of electronics on the inside. Photos of the new unit, which has worked fine so far:

Serial/Model numbers/markings I found on the new unit: 41056, F11-1L  L.MAC6-1430-L021-00, KF-FAND-S008-V00 2019.01.07 1400LM (1400 Lumen?), Intertek 4008733,



Ego 21″ mower (LM2100SP) 3rd self propel motor failure & repair report

My Ego 21′′ self propelled electric lawnmower started on its third self propel unit failure back in June of 2019.  For those keeping track, I bought the Ego 21′′ mower back in April 2017, and it’s self propel unit failed in October of 2017 I took it to Home Depot to be repaired (which took 6 weeks) and the repaired self propel unit lasted until July of 2018. Ego customer support was nice enough to send out a replacement mower that time, so my downtime was only 10-15 days, and I was hopeful that the replacement mower might have a better self propel unit in it. Unfortunately, that one started to fail in June of 2019, so it looks like the lifespan of my self propel units are 4 months, 10 months and 11-16 months.

So, back in June 2019, the self propel unit just stopped working much like it had previously. I called Ego customer support and they offered to ship me a replacement mower. We had that all set up, but the next day I went out to push-mow the rest of my lawn, and lo-and-behold, the self propel unit had “reset” and was working again! So I called Ego back and canceled the replacement.   Unfortunately, although the self propel unit had not totally failed, it had not fully recovered, and still had some issues that got gradually worse over time. Specifically, the top speed was reduced, and over the next several months, the power and top speed appeared to keep dropping. Eventually it got to a point where it would not propel the mower up a slight hill without me assisting. Eventually, in January of 2020, the self propel unit failed completely, and did not “reset” itself.

Ego shipped me a replacement mower and I shipped the bad mower (my 2nd) to them, so I am now on my 3rd Ego mower (and year 3 of my 5 year warranty). However, for the first time in 3 self-propel failures the SP unit is different! The new mower (manufacture date October 2019) has a different style of self propel unit when compared to the three units that had failed on it in the past. The old unit had a gearbox on the drive shaft and the motor body stuck upwards at a 90 degree angle. The new unit has the motor body above but parallel (horizontal) to the drive shaft. I don’t know if the old unit had a fan, but the new unit has a fan clearly visible.

Older Self Propel Unit


New Self Propel Unit

Objectively, the new self propel drive unit doesn’t look as impressive as the older unit, but given the number of failures I have had with the old style, I’m excited to have something change (and hopefully improve). From a performance standpoint, the new self propel unit works just as well as the old style, so there is no loss in performance. I just hope that it will have more longevity than the older units.

I suspect that the size of my yard (which is large enough that it takes me two 7.5 AH batteries to mow it in the fall/winter, and up to four 7.5 AH battery charges in the heat of summer) may be the reason the self propel unit’s are failing. I suspect I’m putting a lot more “miles” on the SP unit than most Ego owners, plus they seem to be failing in the heat of the summer. I’m not sure if that is due to heat related problems, or if it’s due to the grass growing more in the summer.

2020 Over The Air (OTA) HDTV Channels in Orlando, Florida

After the FCC auctioned off some of the HDTV spectrum for cell phone (5G) use, a few stations had to change their channel allocations. Here is a list of the over the air (OTA) channels I can pick up from the west side of Orlando.

  • 2_1 – NBC (WESH-DT) VHF-11
    • 2_2 – Me TV
  • 6_1 – CBS (WKMG-DT)
    • 6_2 – Dabl  WKMG-DT
    • 6_3 – Cozi
    • 6_4 – Start TV
    • 6_5 – Decades
  • 9_1 – WFTV-HD  (ABC) (UHF-35)
    • 9_2 – LAFF
    • 9_3 – Escape
  • 10_1 – Diya TV (Indian? classic cinema) [Schedules Direct does not have a lineup yet]
    • 10_2 – Orange (county government TV)
    • 10_3 – Vision  (orange TV)  (WLOQ?)
    • 10_4 – This TV
    • 10_5 – News Net
    • 10_6 – JTV
    • 10_9 – open
  • 15_1 – WDSC-HD (PBS 15 – Daytona Beach)
    • 15_2 – WDSC-ED The Florida Channel
    • 15_3 – WDSC-WV (Worldview)
  • 18_1 – WKCF-DT  CW
    • 18_2 Justice
    • 18_3 estrell
  • 24_1 – WUCF-HD (PBS 24 UHF -34)
    • 24_2 create
    • 24_3 kids
  • 27-2 – 27 – WRDQ
    • 27_2 – Antenna
    • 27_4 – Grit TV
  • 31_1 – T31 WTMO CD (TeleMundo Orlando )
    • 31_2 – WTMO-SD
    • 31_3 – TELEXITOS (Xitos)
  • 35_1 – WOFL-DT (FOX)
    • 35_2 – LIGHT
  • 43_1 WVEN-TV (Univision) UHF-22
    • 43_2 – GET TV
    • 43_3 – Bounce
    • 43_4 – Escape
    • 43_5 – Quest
  • 52_1  TBN-HD (Trinity Broadcasting Network) (WHVL-TV Digital 52, UHF 32)
    • 52_2 Hillsng (Hillsing)
    • 52_3 – Smile
    • 52_4 – Enlace
    • 52_5 – JUCE
  • 55_1 WACX-D1 (SuperChannel )
    • 55_2 – WAXC-D2
    • 55_3 – WACX-D3
    • 55_4 – WACX-D4
    • 55_5 – WACX-D5
    • 55_6 – WACX-D6
    • 55_7 – WACX-D7
    • 55_8 – WACX-D8
    • 55_9 – WACX-D9
  • 56_1 – ION (WOPX)  Physical Channel 48
    • 56_2 – QUBO
    • 56_3 – ION Plus
    • 56_4 – Shop (sponsored Television programming)
    • 56_5 – QVC
    • 56_6 – HSN
  • 65_1 – WRBW-DT   UHF-28 (fox 35 plus, my network TV)
    • 65_2 Movies!
    • 65_3 H&I
    • 65_4 Buzzr
  • 68_1 – WEFS-HD Educational  (Eastern Florida State College) [minimal signal, can’t receive reliably]
    • 68_2 – WEFS-CL (Classic Arts)
    • 68_3 – WEFS-NS (NASA Educational)
    • 68_4 – WEFS-FL (Florida Channel)

Salvage 2013 Nissan Leaf modules – 7 year old range update

Back in January of 2016 I put a set of battery modules harvested from a salvage 2013 Nissan Leaf into my S-10 conversion electric pickup. In march of 2016 I drove the truck for a while to see what its range was. [More than 46 miles, as I got tried of driving. The pack had a capacity of at least 15 kWh at that point in time.]

37.4 miles on trip meter.

Today I drove the truck for 35.8 miles before the low cell warning beeper from the BMS started to alert. After I got home [37.4 miles total], the average cell voltage of the pack was 3.75, while my (one) lowest cell was down at 3.3 volts. As it turned out, that cell must have started the trip out at a lower state of charge / voltage from the other cells, as it was still low when charging finished and I had to manually add charge to it individually. [My BMS does a good job of alerting at high/low voltage conditions, but does not do much for balancing the pack.]

According to my JuiceBox, the pack required 14.74 kWh to recharge, which is a good estimate on the battery pack’s current capacity. [This is almost exactly the same amount of power that I used in the trip in 2016, but I didn’t go as far due to different driving conditions. And I also hit the bottom of (at least one cell’s) state of charge.

The 2016 trip averaged 322 watt-hr/mile. This trip consisted of a lot of stop & go city driving as well as a few lengthier stretches of 49 mph arterial streets, and I wasn’t light on the accelerator. My measured watt-hour / mile (from the wall, including charger losses) was: 394 watt-hr/mile

Assuming that the pack has a 15 kWh capacity, this is 63% of the brand new 24 kWh capacity, which means I lost 37 % of the capacity over 7 years. (Some of that was in the original Nissan Leaf, but most of it was in my s-10 conversion.)

I’ll repeat the test after balancing my cells a bit better and see how things go.

Update: I drove the truck until the low cell beeper came on again. I went a total of 38.5 miles, and recharged the pack with 16.69 kWh (16690 watt-hours). The relatively higher  433 watt/hours per mile number is a result of the weather being a lot cooler so I was running the heater in the truck and more 45 mph roads. Balancing the cells got the usable pack capacity (measured from the wall with charging inefficiencies) to 16.69 kWh (which could have theoretically gotten me to 42 miles at 394 watt-hour/mi or 51 miles at 322 watt-hr/mile)

The main take-away is that at 16.5 kWh, I still have access to 68% of the brand new 24 kWh capacity Leaf pack, which isn’t too shabby for a 7 year old battery.




12 Month Grid-Tie Solar system report

Over a year ago we installed a 10.4 kW grid tie solar system.  You can read about the shakedown period here.  This post will cover all of 2019, the first full year the solar system was operational. [Technically, it covers the period between Dec 12th 2018 through Dec 11th 2019, as Duke Energy bills us mid-month.]

In this 12 month period, we consumed 16,695 kWh of power at our house (a 3 bed 2 bath ranch with all electric utilities + two electric vehicles). [We used 16,796 kWh the prior 12 months, so our usage did not appreciably change due to installing the solar system.]

Of this total electrical usage, our solar system produced 15,252 kWh or 91.4% of our total electrical usage, while we purchased 1,443 kWh from Duke Energy and the electrical grid. [There are no economic benefits to producing more than we use, so the ideal system would hit 99.9% of actual usage. We were aiming for 90% when we designed our system.]

Over the year, we paid Duke Energy $314.49 ($130.80 for required connection charges, and $183.69 for the electricity we imported from the grid, averaging 12.7 cents per kWh.)  This compares to our previous yearly cost for power of $2,211.13, giving a yearly cost savings of $1,896.64.  After the EIC tax credit, our solar system cost us $17,439.20, which gives a payback period of 9.19 years. (I’m deliberately ignoring the interest we could have earned by investing the money we paid for the solar system in the stock market, which counteracts the fact that I’m also ignoring the fact that Duke energy raises their rates every so often.)

As the solar system is expected to have a working lifespan of 15-25 years, any energy it produces after the payoff period will be pure profit. So yes, a solar system does make economic sense, in addition to the environmental, social and political benefits.