The Barrister Bounce Back Plan

May 13, 2020

Update 30 June 2020: the sustained protests against police violence has created a natural experiment that has produced the heartening if unsurprising finding that the proper use of masks really does significantly reduce the transmission of SARS-COV-2.  While our Barrister Bounce Back Phase One plan already required all customers to wear masks while in the store, seeing the absence of widespread new infections from people spending time together in close quarters outdoors while wearing masks makes us more confident in opening up for in-person shopping with a firm mask requirement at our Board Game Barrister stores and patio dining outside at our Oak & Shield Gaming Pub.

Our updated criteria are:

  • A Positive Test Rate at 5% or below for 7 days; and
  • 500 or fewer New Daily Infections

The hospital availability remains unchanged.

I had started mulling over these changes about 10 days ago when the presumed spike in protest-related infections wasn’t happening and it seemed like “Hey, this seems safe enough to reopen.  Let’s collect some evidence and see if our criteria should incorporate new data.”  As I write this, however, infections are definitely trending in the wrong direction.  My hope is that this turns out to be a blip, and we’ll return to the level we were at a month ago, even 2 weeks ago.  If that becomes the case, I’m optimistically hoping we’ll reopen for limited in-store shopping in the first half of July.

Caveat – testing: there is mitigation and there is suppression.  I am a proponent of suppression (New Zealand, I love you), but I have to accept the reality that nowhere near enough of our country believes similarly.  Therefore, I am considering successfully meeting mitigation standards as being “safe enough” for our staff.  My understanding of current thinking on testing is that Wisconsin should be doing close to 20,000 tests per day to be able to achieve mitigation (and 70,000 tests per day for suppression).  We’ve been doing between 5,000 and 12,000 tests per day recently.  I believe that both the Positive Test Rate and the New Daily Infections could be acceptable at even higher numbers if the number of tests performed reaches the mitigation standard. [This caveat is informed by the Harvard Global Health Institute’s State Testing Targets data available at https://globalepidemics.org/july-6-2020-state-testing-targets/]

Update 07 June 2020: we’re seeing a significant sustained downward trend on the Positive Test Rate.  This implies we’re finding more of the new transmissions that are occurring.  If this can be maintained, we feel that the New Daily infections criteria can be a higher number.  Because it seems that we might see sustained Positive Test Rates in the 2-3% range – far below the 7% our original plan foresaw – it is appropriate to add these scenarios to our plan.

  • A Positive Test Rate at 3% or below for 7 days increases the threshold for New Daily Infections to 250.
  • A Positive Test Rate at 2% or below for 7 days increases the threshold for New Daily Infections to 400.

TL;DR: We will open for in-store shopping and in-restaurant dining when these criteria are met:

  • 10% or greater availability of hospital beds (IBA) in the Southeast region as reported by
    https://www.dhs.wisconsin.gov/covid-19/hosp-data.htm
    If our area hospitals are running too-near capacity, we will do our part in reducing the opportunity for the virus to infect new humans.
  • Positive Test Rate is 7% or below for 7 consecutive days as measured by https://coronavirus.jhu.edu/testing/individual-states/wisconsin
    A low positive rate tells us that our testing capacity is great enough that our reported numbers are believable.
  • 100 or fewer New Daily Infections statewide for 10 consecutive days as measured by
    https://www.dhs.wisconsin.gov/covid-19/cases.htm
    A believable count of confirmed infections gives a read on how much virus is circulating between humans in our area.  This is our best proxy for the level of risk we’re asking our staff and customers to endure.  Setting a bright line puts our cards on the table and let’s us plan in a real way.

Until these three criteria are met, our Board Game Barrister retail stores will be open for online ordering with curbside pickup and our Oak & Shield Gaming Pub will be open for take-out and delivery.  They will not be open for in-store shopping on in-restaurant dining.


[Note: I wrote this before the Wisconsin Supreme Court ruling on May 13 that invalidated Emergency Order #28 which extended the Safer At Home period to May 26.  There is truly nothing I loathe more in this world than someone who says No to a proposed solution without proposing an alternative solution in its stead.  Growing up listening to Rush, the lyrics of Neil Peart are again ringing in my mind’s ear, “If you choose not to decide, you still have made a choice.”  Negating a considered decision to act only to replace it with unconsidered inaction is reprehensible and an abdication of leadership.  The more of this negation that occurs and results in real harm, the more I fear the post-pandemic reckoning will include not words and ideas, but torches and pitchforks.]


The goal of Wisconsin’s Safer at Home and similar measures around the country is to reduce transmission of the SARS-COV-2 virus that causes the Covid-19 disease by creating fewer opportunities for the virus to jump from human to human.  We believe that having an understanding of how much virus is being transmitted through the population is a critical prerequisite in determining our risk of becoming infected and of transmitting infection.

The risk will never be essentially-zero in the near-term – we’ll have a vaccine for that – so the question is how much risk of infection we are willing to take for ourselves, our staff, and our customers by increasing our exposure to other potentially-virus-carrying humans knowing that the disease may be present in one human without symptoms and yet prove serious or even fatal to the human they unknowingly pass it to through otherwise innocuous contact.

The Badger Bounce Back Plan informed by Wisconsin’s Covid-19 Infection Model laid out some reasonable criteria based on measurable evidence to hit pre-defined triggers for different degrees of increased human-to-human interaction that would allow different types of businesses to resume operation in varying limited forms.  I am a fan of evidence and I am a fan of plans.  Because I am a lifelong businessperson and business is based around predictability, it should be no surprise that I am a big fan of plans based on evidence precisely because they allow for predictability.  That brings us to May 11 and Wisconsin Emergency Order #36 (EO36), which surprisingly, unpredictably, and without notice allowed certain types of businesses to resume limited operation.  Note that EO36 applies to some, but not all, of our locations.

EO36  caught me off-guard, and I wanted to understand why.  I look at the data every. single. day. and I certainly wasn’t saying to myself “hey, looks like things have been on a downward trend for 7-days, only 7 more to go” or “Wow, it’s been 10 days of decreased cases, looks like we’ll be opening soon.”  A reasonably informed person who is paying attention should be able to follow along with an evidence-based plan, making predictions and changing course accordingly.  That didn’t happen.  And, while it seriously pisses me off as an example of the illusion of a plan (and the corollary of reduced trust in the ones fabricating said illusion), it also means that we have to use our own informed judgement to assess what level of infection risk we are willing to expose our staff and customers to.

And why was I surprised?  Because it turns out that the Badger Bounce Back Plan has what I consider to be a meaningful loophole in its analysis.  It looks at symptomatic cases that result in Emergency Room visits, and it looks at testing capacity as inferred by the rate of positive tests – but it does not look at actual positive tests for Covid-19.  Moreover, it doesn’t set hard benchmarks for what it does measure, instead merely looking for “decreases”.  (Don’t get me wrong, decreases in what the Badger Bounce Back is measuring are preferable to increases – but what matters is the actual level, because a decreased level of infection can still be a too-high level of infection.)  Here are the two criteria where the Badger Bounce Back Plan falls short.

Symptoms (2nd entry).  Downward trajectory of Covid-like syndromic cases reported reported in a 14-day period.

As written, this seems fine.  However, the devil is in the detail of “cases reported”, which becomes implemented as “Daily number of emergency department visits with suspected Covid-19 related concerns” in graph below (emphasis mine).  While the ER is one way to report a case, it is not the only way to report a case.  I believe that actual positive tests from all sources is the relevant data point to monitor.

Cases.  The Cases criteria in the Badger Bounce Back Plan is defined as a “Downward trajectory of positive tests as a percentage of total tests within a 14-day period.”

There are two elements of this that fail to actually account for risk.

A) “downward trajectory” is necessary but insufficient.  Fewer of a too-high number is still too-high.

B) “positive tests as a percentage of total tests” is insufficient because arithmetic tells us the percentage itself will fall if the number of tests dramatically increases (which it did, yay!)

The percentage of positive cases is important.  Let me say it again.  The percentage of positive cases is important.  And we want it to be low.  3-12% is what the World Health Organization has found to be the positive test rate from countries that do a lot of testing.  My understanding is that a sufficiently low positive rate tells us that we’re testing enough people to believe that we’ve actually identified all the positive symptomatic cases out there (eventually, maybe, we’ll be able to then estimate how many asymptomatic cases are also out there).  If the positive rate is not sufficiently low, we instead know that we aren’t doing enough testing and thus cannot believe the infection numbers are a valid indication of the amount of virus present in the community of humans tested..

The Cases criteria in the Badger Bounce Back Plan would be more appropriately named “testing capacity”, meaning that we have enough testing capacity in Wisconsin to detect new cases.  This is very different than the number of new cases actually decreasing (not yet).  The good news is that we’ve been seeing positive test rates in the 10% range for several weeks, and this has recently dropped to 7%, which does imply that our testing capacity in Wisconsin is up to scratch.  While this doesn’t mean our cases are dropping, it does mean we’ll be able to believe the drop in cases when we see it.  A necessary but insufficient step in the process, as it were.  For those keeping score at home, every day has had more confirmed infections than the previous day going back to April 13 (through the last update on the graph, May 10).

Ergo my surprise at EO36 which I thought was premised on a decrease in all cases, not merely an indication of our testing capacity or a decrease in ER Covid cases.

Because what I care about is the actual number of new cases.  That informs how many people are walking around spreading virus, and thus how likely my staff are to encounter the virus while they’re working at one of our stores.  Because the Badger Bounce Back Plan doesn’t address total new cases, and total new cases is what we really need to assess risk, I’ll have to roll out my own, as it were.  Let’s call it the Barrister Bounce Back Plan.

Criteria 1 – Hospital Utilization
The percentage of Immediate Bed Availability (Total IBA Beds as a percentage of Total Beds) should be greater than 10% as reported by the Wisconsin Department of Health Services Hospital Capabilities for the Southeast region.  As of this writing, there are 4,398 Total Beds in the Southeast region per this report; 440 or more IBA Beds meet this criteria.
https://www.dhs.wisconsin.gov/covid-19/hosp-data.htm

Criteria 2 – Testing Capacity
The positive test rate should be no higher than 7% on any of the prior 10 days according to the “blue line” labeled Percentage Positive Tests (7-day moving average) on the Johns Hopkins Rate of Positive Tests Over Time graph for Wisconsin.  This demonstrates that we have adequate testing capacity to believe the New Confirmed Infection numbers in Criteria 3.  
https://coronavirus.jhu.edu/testing/individual-states/wisconsin

Criteria 3 – New Confirmed Infections.
There are not higher than 100* new confirmed infections state-wide for each of the prior 10 days according to the “light blue line” labeled “Newly reported cases” on the Wisconsin Department of Health Services Covid-19: Wisconsin Cases -> Cumulative total and newly reported Covid-19 cases by date confirmed graph.
https://www.dhs.wisconsin.gov/covid-19/cases.htm
If the WI DHS make county-level data available in a daily updated report or graph (not the cumulative or raw downloadable data currently available), this will change to fewer than 10 new confirmed infections in Milwaukee County for each of the prior 10 days.

Until these three criteria are met, our Board Game Barrister retail stores will be open for online ordering with curbside pickup and our Oak & Shield Gaming Pub will be open for take-out and delivery.  They will not be open for in-store shopping on in-restaurant dining.

When these three criteria are met, we will open for Phase 1 of our Barrister Bounce Back Plan.

What’s in Phase 1?

We have extensive planning documents for Phase 1 at the stores and the restaurant.  There will be separate posts as those plans become executed (and, if we’re good, walk-through videos) going over what you can expect when we re-open for in-store shopping and in-restaurant dining.  There will be changes, and we’ll be asking you to participate in many of those changes. Some of the easy things I can say here are:

  • We’ll open for in-store shopping and in-restaurant dining.
  • We’ll have a 6′ bubble marked around the checkout counter, and we’ll have fewer checkout stations.
  • Contactless payments will be encouraged and facilitated as much as possible.
  • Our order online / pickup-in-store service will remain, with a pickup area near the door for those who want to minimize the time and distance spent inside the store.
  • There will be caps on the number of customers allowed inside the store.
  • All employees will wear face masks and face shields (yes, we’re providing them).
  • There will be hand sanitizer stations at convenient locations.
  • All customers will be required to wear masks for the safety of our staff.
    • “No shoes, no shirt, no mask – no entry.”  (I’m hoping this doesn’t end up being viewed as too-snarky for the times, because I appreciate the call-back to the signs of my youth.)
  • We have an extensive 2-hour cleaning schedule with logs for all common surfaces in the store.
    • To minimize common surfaces, there are areas of the store that will be off-limits to customers.
    • Customer-touch items such as pens or check presenters will be quarantined for 3 days or sanitized after each use.
  • The event spaces and game libraries will be unavailable at the stores.
    • Access to the game library at the Oak & Shield will be limited, and games will be quarantined for 3 days before being returned to the library.
  • Our hours will change,.
    • Stores will include a seniors-only shopping hour at the start of the day.
    • Schedules will be written to reduce overlap of employees working.
    • The Oak & Shield will close for one hour on Saturday afternoons to allow for cleaning and a shift change.
  • The Oak & Shield in-restaurant seating will be reduced and reservations will be required.
  • In the event that an employee tests positive for Covid-19, their location will close for 3 days (and, of course, they’ll stay home until they are well).
  • In the event that a customer tests positive for Covid-19 and we somehow learn of it… I don’t know.  A customer who was wearing a mask and was in-and-out, that doesn’t feel too risky given all the other precautions.  A customer who was hanging around for an hour, used the bathroom, meandered through the whole store and touched a bunch of things… that’s a lot more virus that could be hanging around.  Not the same as if an employee working a full shift was spreading virus, but certainly more than the first scenario.  In practice, I don’t think we’re likely to be told of a customer infection.  But if we are – then with what we’ve thought through as of today – I don’t know what the right response is.  I hope you enjoy this insight into the state and extent of our thinking.

What if things go the other way after we’ve moved into Phase 1?

The great thing about having a plan is that we can re-use said plan.  We’ll move back to Phase Zero (no in-store shopping or in-restaurant dining; online ordering with curbside pickup at the stores, take-out and delivery at the pub) if one or more of these things occur (using the same data sources as above).

  • Hospital Utilization: IBA beds drop below 5% for any 3 of the last 5 days.
  • Testing Capacity: positive test rate is greater than 10% for any 5 of the last 10 days, or above 12% for any 3 of the last 5 days.
  • New Confirmed Cases: newly reported infections are greater than 110 for any 5 of the last 10 days, or above 150 for any 3 of the last 5 days.

Philosophically, I want to allow for a bit of a cushion in case things get a little bit worse for a little while so long as they return to being in-bounds, while being less-forgiving of what might be the start of a larger spike.  Any epidemiological statisticians in the audience who want to give more informed buffer numbers, I’m all ears.


Source material and related links

Wisconsin’s Covid-19 Model
https://www.dhs.wisconsin.gov/covid-19/model.htm

Badger Bounce Back Plan
https://www.dhs.wisconsin.gov/covid-19/prepare.htm

Badger Bounce Back – full plan
https://www.dhs.wisconsin.gov/publications/p02653a.pdf

Wisconsin Department of Health Services Cumulative total and newly reported Covid-19 cases by date confirmed
https://www.dhs.wisconsin.gov/covid-19/cases.htm

Wisconsin Department of Health Services Hospital Capabilities (filter for Southeast region)
https://www.dhs.wisconsin.gov/covid-19/hosp-data.htm

Johns Hopkins: All State View of Week to Week Change in Percentage of Positive Tests
https://coronavirus.jhu.edu/testing/individual-states/wisconsin

Johns Hoplins State Comparison of Positivity Rates
https://coronavirus.jhu.edu/testing/testing-positivity

Institute for Health Metrics and Evaluation at the University of Washington (IHME) Covid-19 Projections for Wisconsin
https://covid19.healthdata.org/united-states-of-america/wisconsin
I was going to use this data as the basis for our Criteria 3 because I felt its 3-day averaging allowed us to use a shorter number of lookback days (7-days instead of 10).  But, IHME’s data appears to update more slowly, which is fine for their use but not for ours.  We’ll use the Wisconsin DHS data instead.

Harvard Global Health Institute estimated targets for testing required by each state for mitigation and suppression of the COVID-19.  https://globalepidemics.org/july-6-2020-state-testing-targets/


* Why 100 new cases (or 10 for Milwaukee County)?
First, let me say that there is substantial benefit in having a bright line, even if there ends up being some variance around that line.  The key is the ability to plan, and watching a trend towards that bright line let’s us say “hey, we dropped to xx cases today, if we keep it up we’ll be opening in yy days.”  Ok, how did I arrive at 100 / 10?

In April, the IHME tried to make a best-guess for guidance to states on when they could reduce social distancing.  (The IHME has removed this because their new model has more specific guidance for states to base their decisions.  I won’t be surprised to learn, later, if there were other reasons as well.)  That guidance was that states could begin removing restrictions when new cases were not more than 1 new case per 1-million people per day, i.e. 6 new cases per day statewide, or 1 new case per day in Milwaukee County.  Those are unbelievable numbers – in that I don’t believe that we’ll get transmission that low without a vaccine.  So, I looked at how many new cases we had when the Safer At Home order was issued on March 25.  104 cases.  We want to be better than that, and 100 is an enticingly round number.  I don’t have near as much reasoning the Milwaukee County number, other than 10 is another round number and I want some slop because I’d prefer a more independent source than the Facebook Covid-19 Resource Page.

Gordon L.

Gordon, a.k.a. "the" Barrister, or simply "G", opened the first Board Game Barrister store in 2005 in a fit of delusion: that he could both finish up law school and start a retail business - one has worked out, the other less so. Gordon loves working in the toy industry, which is filled with wonderful people who love what they do; he loves working with his fellow Barristers; and he loves the community of people that has grown around the playing of games together at the Board Game Barrister stores. If Gordon were a Dungeons & Dragons character, he would be a lawful goody-two-shoes silicon wizard with a skills in perl and complex compound sentences. Gordon loves many games, but when forced to choose he went with China, Power Grid, Lost Cities, and Empire Builder.

Share!
Enjoy the article so far? Recommend it to your friends and peers.