#predictor 2020

How Do Erin’s Miscarriages Factor Into Her Data?

Very, very, very good Ask.  I was actually hoping someone would Ask this.  Erin had 3 Miscarriages in 2014, followed by 4 Live Births—Carson (2015), Brooklyn (2016), Everly (2018), and Holland (2019).  Erin + Chad have one Child Spacing data point for each live birth, obviously…  But what about the miscarriages?

Ideally, we’d have enough information to assign a ‘Theoretical’ Spacing to each loss, too.  To do so, we’d need to know a Due Date, Loss Date, or both.  Sadly, in this case, we only have the information for one of Erin’s three miscarriages—the first one.  For Erin’s 1st Miscarriage, we know that she was due on or about October 15, 2014—i.e., a week before Whitney (w/ Bradley).  (Whitney was due on October 22, 2014.)  As for the Loss Date…  Erin found out she was pregnant in February 2014.  (Erin + Chad married on November 2, 2013.  Erin discovered she was pregnant 3 Months Later.)  Unfortunately, trouble arose at the very first ultrasound, which “showed only an empty sac.”  It’s possible that it was simply too early to visualize the sac, so Erin’s OB–GYN did “several ultrasounds,” plus  blood tests, probably over a period of weeks.  Ultimately, Erin did miscarry, and underwent a D&C.  Apparently, this all occurred by March 17, 2014, when Chad announced the loss on their personal blog…  From all this, we can determine a rough loss date range of mid–February to March 17.  (Duggar Data uses March 1, 2014 as an estimate.  If I knew the exact date of the D&C, that’s the date I’d use.)  That allow us to calculate a spacing for Miscarriage 1, as follows—

Child Spacing for Erin’s 1st Miscarriage = Projected DOB (i.e., Due Date + Erin’s Typical Delivery vs. Due Date [–27 Days]) — Erin + Chad Wedding Date

… which works out to 320 Days.

Since we don’t have Loss Dates or Due Dates for Miscarriages 2–3, there is no way to calculate Theoretical Child Spacings.  Instead, the Predictor assumes a loss at 6 Weeks (42 Days), and reduces the next child spacing by that much.  It isn’t a perfect system, but it’s the best we can do with limited data.  In the case of Erin + Chad, that next spacing is Carson’s.  It’s calculated as follows—

Child Spacing for Carson = Carson’s DOB — 1st Miscarriage Loss Date — 42 Days for Miscarriage #2 — 42 Days for Miscarriage #3

… which works out to 355 Days.

Since there aren’t any intervening miscarriages, the rest of Erin + Chad’s child spacings are simply the difference between the two DOBs—

Child Spacing for Brooklyn = Brooklyn’s DOB — Carson’s DOB

Child Spacing for Everly = Everly’s DOB — Brooklyn’s DOB

Child Spacing for Holland = Holland’s DOB — Everly’s DOB

Ultimately, the Paines’ Child Spacings work out to be as follows—

Miscarriage #1   320 Days

Carson   355 Days

Brooklyn   450 Days

Everly   601 Days

Holland   606 Days

For purposes of the Bates 2nd Child Multiplier, Miscarriage #1 and Carson are both treated as a 1st Spacing, while Brooklyn is treated as a 2nd Spacing.  For purposes of Procreative Pace, all spacings other than those deemed to be 1st Spacings (Miscarriage #1 and Carson) are averaged.  So…  Erin + Chad’s pace is 1 Child Every 552 Days, as or right now.  

When to Expect:  The Next Pregnancy Announcements

As of March 26, 2020, Predictor 2020 expects that the next couple to publicly announce a pregnancy will be…

The Duggars   Jill + Derick Dillard  (June 10, 2020)

The Bateses   Michaela + Brandon Keilen  (April 2, 2020)

The Maxwell   Chris + Anna Marie Maxwell  (December 21, 2020)

The Kellers   Nathan + Nurie Keller  (February 10, 2021)

The Bontragers   Carson + Carolina Bontrager  (May 27, 2020)

Wait.  Explain the C–Section Delay Again, Please?

Yeah, I should have explained that better.  Let me try again...

For quite awhile now, the Predictor has increased a couple’s expected spacing by 93 Days after a C–Section.  That’s based on a medical journal study I found, which showed that, on average, it takes 93 Days Longer to conceive after a C–Section (vs. a vaginal birth).  That study also showed that this 93 Day delay will go away if a woman who once had a C–Section later has a successful VBAC—so it’s the woman’s most recent delivery that drives things, not the fact that she once upon a time had a C–Section.

After a C–Section, the Predictor continues to impose the 93–Day delay, when it forecasts the next pregnancy.  The question arises, however, in how to treat the actual, post–Cesarean spacing, when that woman’s next child is born.  Prior to this, the Predictor just treated that spacing like any other...  It didn’t modify it, at all.  This is problematic, however, since there is a good chance that that spacing was longer than it otherwise would have been, due to a C–Section immediately prior.  Basically...  That spacing is a bad representation of the woman’s ‘natural’ pace, since it’s delayed by a condition that may not exist in later pregnancies—a prior Cesarean delivery.  With this in mind, Predictor 2020 modifies any post–Cesarean spacing by –93 Days, to prevent delay caused by the C–Section from impacting the woman’s average pace too much.  So, in summary—

After a woman has a C–Section, the Predictor delays the next Pregnancy Announcement / Birth by 93 Days, on the theory that that C–Section may slow her down due to a longer recovery time, or whatever.

Then, when a woman actually gives birth after a C–Section, the Predictor subtracts 93 Days from the actual spacing, so that the couple’s pace will not be skewed low due to delays solely attributable to a prior C–Section.  This modified spacing is what is used to calculate their Procreative Pace, going forward.

If the woman’s post–Cesarean delivery was a successful VBAC, then the Predictor forecasts her pregnancy like normal...  Next Child’s DOB = The Latest Child’s DOB + Procreative Pace.  If, however, she had another C–Section, she would again be delayed by 93 Days.  Her Next Child’s DOB = Latest Child’s DOB + Procreative Pace + 93 Days.

[ Referenced Ask ]

No problem!  Hilariously, that Post was actually already outdate, when posted.  Joy–Announced announced Pregnancy #2* that day, which shifted the Duggar Family’s Child 2 Multiplier ever–so–slightly, from 1.7755x to 1.7015x.  Jinger + Jeremy’s Pregnancy #2 prediction relies on the multiplier (along with their own Marriage–to–Child 1 Spacing), so their Child 1–to–Child 2 Spacing—and thus, Vuolo #2′s Predicted DOB—also shifted a bit.

Factoring in Forsyth Pregnancy #2*...  Here’s what Predictor 2020 expects for Vuolo #2, right now—

Child Spacing   1,057 Days  (vs. 1,103 Days)

Baby News   November 24, 2020    

Due Date   June 10, 2021   

Birthdate   June 9, 2021

No, I don’t think so.  You’re right...  Some women have children at Age 45.  (It’s a bit telling, though, that you were only able to come up with one example of a birth at Age 45, even among this group of proven–to–be–fertile, not–preventing pregnancy, wanting–more–children women.)  But, that’s not the point...  Age 40 is the default cut–off because, based on what I’ve read, it’s reasonable to think that many women are still fertile to that age.  I don’t think it’s reasonable to say that most women are fertile to Age 45.  Because most women aren’t.

Check out this opinion from the American College of Obstetrics & Gynecology, if you’re curious for some of my sources.

Also, just so you know, the Predictor wouldn’t stop functioning if a couple was to unexpectedly get pregnant again after their projected ‘cut–off.’  If just stops expecting them to announce.  If they announce, it treats that pregnancy like all their others!

There has to be some kind of cut–off, and that might result in underestimating some women’s fertility.  Hell, there’s a case of a woman giving birth at Age 66.  But, all things considered, I think Age 40 is a good cut–off.  Shifting to Age 45 would, IMO, overestimate fertility far more than Age 40 underestimates it.

Procreative Pace (or Average Child Spacing) is one of the most important data points for the Predictor, since it drives Pregnancy Announcements—and, thus, Projected DOBs and ESOQ.  Here's the relevant data, as of March 20, 2020...

Working Pace for Individual Couples—

(Note—If a ‘+’ is listed after the pace, that means they are ‘late’ to announce their next pregnancy, and the Predictor is preemptively factoring in the delay, starting with their next next pregnancy.  As a result, late couples’ Procreative Paces will increase by ~1 Day every few days, depending on how many data points they currently have.)

The Duggars—

Joy + Austin   477 Days

Joe + Kendra   512 Days  (Note—Only one data point.)

Josiah + Lauren   622 Days  (Note—Based on Child 2 Multiplier.)

Jessa + Ben   649 Days

John + Abbie   732 (731.6) Days  (Note—Based on Child 2 Multiplier.)

Josh + Anna   732 (732.0) Days

Jill + Derick   919 Days+

Jinger + Jeremy   1,057 Days  (Note—Based on Child 2 Multiplier.)  

The Bateses—

Carlin + Evan   423 Days  (Note—Based on Child 2 Multiplier.)

Josie + Kelton   483 Days  (Note—Based on Child 2 Multiplier.)

Tori + Bobby   508 Days  (Note—Only one data point.)

Erin + Chad   552 Days

Alyssa + John   648 Days+

Zach + Whitney   691 Days   

The Maxwells—

John + Chelsy   510 Days  (Note—Only one data point.)

Chris + Anna Marie   557 Days

Nathan + Melanie   632 Days   

Joseph + Elissa   787 Days

The Kellers—

Esther + John   569 Days

Priscilla + David   653 Days

Anna + Josh   732 Days

The Bontragers—

Chelsy + John   510 Days  (Note—Only one data point.)

Josh + Cassidy   699 Days  (Note—Only one data point.)

Carson + Lina   807 Days+  (Note—Based on Child 2 Multiplier.)

Mitchell + Bryn   947 Days  (Note—Based on Child 2 Multiplier.)

... and, just for fun, let’s see how the Family Data stacks up—

Family Data—Average for Child 1

Bates Family   312 Days  (10.3 M)

Duggar Family   374 Days  (12.3 M)

Bontrager Family   396 Days  (13.0 M)

Keller Family   408 Days  (13.4 M)

Maxwell Family   458 Days  (15.0 M)  (Note—Nathan + Mel had a lot of fertility issues, at first, and their 1st Spacing isn’t included.  Including it would skew their Family Average massively to 803 Days.  Outlier!)

Family Data—Average for Child 2+

Bates Family   591 Days  (19.4 M)

Bontrager Family   605 Days  (19.9 M)  (Note—Only one data point.)

Maxwell Family   620 Days  (20.4 M)

Keller Family   627 Days  (20.6 M)

Duggar Family   650 Days  (21.4 M)

Family Data—Child 2 Multiplier  (vs. Child 1)

Keller Family   1.3083x

Maxwell Family   1.3323x

Bates Family   1.6834x

Duggar Family   1.7015x

How Did The Vuolos 2nd Pregnancy Announcement Shift So Much?

ICYMI—There’s an entirely new Predictor!  So, that is the main reason.  As for exactly what changed with Predictor 2020 that caused this…  It’s the rules for predicting Child 1–to–Child 2.

Felicity arrived 621 Days into the Vuolo marriage.  As such, the Prior Predictor set the Vuolos Procreative Pace at 621 Days, the average of all their spacings. (Air quotes around “average,” since it’s really not an average.  It’s the only data point, thus far.)  Note that Felicity isn’t treated as a Honeymoon Baby, and thus their spacing was just set at 621 Days, not at 1.5 * 621 Days.  At that 621–Day pace, Jinger + Jeremy are already late to announce Pregnancy #2.

Predictor 2020 does two things differently, here—

Did Away w/ All Special Rules for Honeymoon Babies.  All 1 Child Couples’ future projected pace is calculated the same way.  No special rule to ‘slow down’ a couple after a Honeymoon Baby.  (It is so hard to define “Honeymoon Baby,” anyway, so that’s what motivated me to do away with those special rules.)

Uses Data to Estimate The Slow Down From Child 1 to Child 2.  The New Predictor sets the Procreative Pace of a 1 Child Couple with a very simple formula:  Marriage–to–Child 1 Spacing * a Data–Based Multiplier, which is the average slow down from Child 1 to Child 2, as seen in other couples in the same family (Duggar).

As of March 2020, your Average Duggar Couple’s Child 1–to–Child 2 Spacing is 1.78x Longer that their Marriage–to–Firstborn Spacing. Thus, Predictor 2020 set the Vuolos pace at 621 Days (Marriage–to–Felicity) * 1.775 (Duggar Multiplier)  = 1,103 Days.  As such, Predictor 2020 doesn’t expected Vuolo #2 to arrive until Felicity is 1,103 Days Old…  At that pace, Jinger + Jeremy aren’t even overdue to announce, yet.  They’ll be on–time if they announce in January 2021.

It’s New, Improved, and Better Than Ever  (Part 2)

The Predictor is Duggar Data’s favorite forecasting tool.  It’s a massive Excel workbook, written by yours truly, that tracks all of the data—i.e., dates—about relationships and pregnancies, crunches that data, and then plugs all of the data into an algorithm to predict future events.  Predictor 1.0 hit the scene almost 2 Years Ago...  Since then, I’ve improved it several times, but Predictor 2020 is by far the most significant revision.  Honestly, it’s a total rebuild, which incorporates more data than ever before, tweaks the algorithm to hopefully be more accurate and realistic, and allow for future expansion.  (Now, I can easily add new families, or make specific predictions for the 4th Generation!)

Here’s a quick overview of Predictor 2020—what’s the same, what’s different.  Part 1 focused on what’s the same.  Now, we’ll talk about what’s new.

What’s New + Improved—

Factoring in Miscarriage Data—

For the first time, Predictor 2020 incorporates data from known miscarriages or stillbirths.  I’ll admit that, in the past, I’ve been hesitant to add this feature.  But, after Joy + Austin lost Annabell, I decided it was critical not to ignore this data.

Miscarriages will impact Child Spacing, as follows—

Losses w/ Known Due Date + Loss Date   If the baby’s due date a date of loss are both known (e.g., Asa, Annabell), or if one is known and the other can be reasonably estimated (e.g., Erin’s 1st Miscarriage, Whitney’s 2018 Miscarriage, Melanie Maxwell’s 2009 Miscarriage), Duggar Data assigns a Child Spacing to the loss itself, representing how closely that child would have been spaced after the prior child, prior loss, or marriage.  Here’s the formula:  Child Spacing for Miscarriage = (Due Date + Average Timing of Birth vs. Due Date) – Prior Event (Birth, Loss, or Wedding).  Also, the next child’s spacing is counted from the Loss Date, not the Prior Event.

Losses w/ Minimal Data   When we know a couple suffered a loss, but we don’t have any real data to go on—e.g., Anna’s 2009 / 2010 Miscarriage—Duggar Data factors in that data by slightly decreasing the spacing for the subsequent child.  Specifically, the next child’s spacing is reduced by 6 Weeks (42 Days), by the Gestational Age at Prior Miscarriage, if known.  I’m using 6 Weeks as the default, since most losses occur by Week 6.

Example #1   Josiah + Lauren suffered a loss (Asa), then welcomed Baby Bella.  Asa was due on June 5, 2019, and Lauren miscarried on October 4, 2018.  This falls within “Losses w/ Known Due Date + Loss Date.”  As a result, Si + Lauren have two child spacing data points labeled as ‘Marriage–to–Firstborn’—

331 Days for Asa   (Asa’s Due Date [6–5–19] + Average Delivery vs. Due Date [9 Days Based on Lauren’s Data; Prior to Bella’s Birth, Family Data Was Used] – Prior Event [Josiah + Lauren’s Wedding, 6–30–18]) 

400 Days for Bella   (Bella’s DOB [11–8–19] – Prior Loss Date [10–4–18])

Example #2   Anna (Keller) Duggar suffered a miscarriage after Mackynzie, but we don’t know when, etc.  The Predictor factors in this loss by slightly reducing the spacing for Anna’s next–born child (Michael).  We don’t know how far along Anna was when she miscarried, so it assumes 6 Weeks.  Michael’s Spacing will thus be 573 Days—i.e., Michael’s DOB –Mack’s DOB (615 Days) – 42 Days.

New Method of Predicting The Future Pace of 1–Child Couples—

One Child Couples are a conundrum, because they have a single data point to use...  But at the same time, it’s unlikely that they’ll maintain their Marriage–to–Firstborn Spacing for the duration of the woman’s fertility.  An earlier Predictor addressed this with a new rule...  If a couple’s first child arrived relatively quick, then Child 2′s DOB would be set at Child 1′s DOB + (1.5 * Marriage–to–Child 1 Spacing).  This dealt with the issue of grossly overestimating the future pace of 1 Child Couples, but it always bothered me that the1.5 Multiplier wasn’t data–based...  It was arbitrarily–selected, based on what I thought was ‘reasonable.’

Finally, Predictor 2020 fixes the problem!  One Child Couples’ next spacing will still be based on Marriage–to–Firstborn Spacing and a multiplier; however, that multiplier is now data–based.  I’ve calculated each family’s typical ‘slow–down’ from Child 1 to Child 2, and will use that as the multiplier.  So, the new formula is...  Child #2′s DOB = Child #1′s DOB + (Average Slow Down [Family’s Data] * Marriage–to–Firstborn Spacing).

Example   Per the data, a typical Bates Couple’s Child 1–to–Child 2 Spacing is 1.68x longer than their Marriage–to–Firstborn Spacing.  Based on this, Carlin + Evan’s—who had Layla only 251 Days into their marriage—are due to welcome Stewart #2 1.68 * 251 Days = 423 Days after Layla.

Avoiding Skew from Marriage–to–Firstborn Data—

Under the Old Predictor, a couple’s Procreative Pace was calculated by averaging all their child spacings, including Marriage–to–Firstborn. This was slightly problematic for couples with only a few children, as the Marriage–to–Firstborn spacing seemed to skew the average much lower than it otherwise would be. (For almost all couples, Marriage–to–Firstborn is by far their shortest spacing).

Predictor 2020 eliminates this problem. Marriage–to–Firstborn will no longer with averaged in with the other spacings. A couple’s Procreative Pace will be set at the average of all their other spacing. When a couple has only one child, the multiplier is used to set a tentative pace. When they actually announce Pregnancy #2, the estimated pace (based on the due date for that pregnancy) will substitute in (and later be updated again when the baby actually arrives). Thereafter, each subsequent spacing is averaged in, thus possibly shifting the pace over time.

Simpler Way of Dealing w/ Late Events—

Prior Predictors had this nifty feature called “dynamic adjustment,” which was intended to preemptively increase the couple’s baseline spacing based on the lateness of their next pregnancy.  This was cool, but also really confusing...  I’ll admit that even I sort of forgot how it worked.

Now, I’m going back to basics.  When an event is late, it will simply be pushed forward by a certain amount of Standard Deviations (SDs).  Previously, this was done by increments of 1 (e.g., 1 SD, 2 SDs, 3 SDs).  Now, it will be done by 0.5 increments (0.5, 1.0, 1.5, 2.0).  Additionally, the Predictor will still adjust to take into account this longer spacing; however, the ‘dynamic adjustment’ only kicks in for the next pregnancy.  So, now I can clearly state how ‘late’ a pregnancy is; none of this confusing ‘dynamic’ stuff.  Hurrah!

A Rule for Deciding When A Late Event Won’t Ever Happen—

Many of you have asked if there’s any point at which I’d assume, say, Jana will never court of the Keilens will never have children.  I’ve thought it over, and the New Predictor includes this as a feature.

Here’s the rule:  If the event is 5+ SDs and 5+ Years overdue, the Predictor will assume that, actually, it’s never going to happen.  Also, just FYI, this applies to courtship and pregnancy announcements.  If someone’s really late to give birth, the Predictor won’t give up and decide the baby isn’t ever coming.  Likewise, it won’t assume that a long engagement means the wedding is off.

Example   The Maxwells have a 38–Year–Old daughter named Sarah.  The data indicates that she was due to court in 2006, and is currently 9.1 SDs and over a decade late to court.  At this point, the Predictor thinks she’ll remain single.  By contrast, it hasn’t yet given up on Jana courting or Michaela + Brandon having children.  Jana is >5 Years, but only 3.5 SDs, Late.  And the Keilens are 46 SDs Late, but haven’t yet been married for 5 Years.   

Not Double–Counting The C–Section Delay—

Awhile back, the Predictor began assuming that a C–Section would slow down a couple’s ordinary pace by 93 Days.  While I added time to a couple’s pace to account for C–Section, I didn’t do anything to prevent the resulting spacings—which might be ~93 Days Longer than they otherwise would have been—from skewing the family data.

This isn’t a huge change...  After a C–Section, I will continue adding 93 Days to the affected couple’s expected pace, and now will also reduce the subsequent Child Spacing by 93 Days.  This is intended to remove the delay caused by the C–Section from the family data, on the theory that that spacing would’ve been slightly quicker but–for the C–Section.

Example   Jill had a C–Section with Israel.  Sam arrived 824 Days Later.  Sam’s Child Spacing won’t be 824 for Predictor 2020, but 731 (or 824 – 93); however, since he arrived by C–Section also, the Predictor will continue to add 93 Days to the Dillard’s expected pace.    

New Predictor Families (and Room to Expand)—

The Predictor has expanded!  Predictor Families now include—

The Duggars 

The Bateses

The Maxwells

The Kellers  (Quiverfull Only)

The Bontragers

Sierra + Mark

Bethany + Dav

Additionally, because of the way the algorithm is now written, it is super simple to add new families or couples.  I could even have the Predictor run using all of the data from all families, rather than family–specific data.  Nifty!