"Head Count" - June 12, 2020

A place to discuss the weekly Wall Street Journal Crossword Puzzle Contest, starting every Thursday around 4:00 p.m. Eastern time. Please do not post any answers or hints before the contest deadline which is midnight Sunday Eastern time.
Rate the difficulty of the meta
Locked
Scooby
Posts: 70
Joined: Tue Apr 23, 2019 2:30 pm

Re: "Head Count" - June 12, 2020

#341

Post by Scooby » Mon Jun 15, 2020 10:00 pm

Give this one to Mike. Pretty clever meta, I didn't even think anything remotely to his train of thought. When Mike and Matt compete, i lose...sigh. Always next friday.

User avatar
MarkL
Posts: 385
Joined: Fri Apr 12, 2019 8:44 am
Location: Tampa Bay, FL

#342

Post by MarkL » Mon Jun 15, 2020 10:18 pm

Bob cruise director wrote:
Mon Jun 15, 2020 5:48 pm
BethA wrote:
Mon Jun 15, 2020 4:47 pm
It turns out I was on Beach Forty on Thursday. While going through the mechanics, I wrote the translated numbers both as numerals and spelled out as words. When TFSES didn’t spell anything, the next thing I thought of was adding the numbers up to get Forty. I did check out grid square 40, to see if there was anything interesting there, just an S. Thought FORTY must be the answer, but did not get that resoundingly strong CLICK. Why 40? Other than it’s a number.

So Friday morning, I decided I’d better check out the letters in the individual numbers, too. And found TALLY, which all added up! Fortunately I almost always let my initial answer percolate for awhile, not in a rush to send it in. So I did make it to Tally Beach in time.

On a different topic, someone earlier in the thread wondered about solve success rates across different constructors. I don’t have that, but am interested, too! Just in reviewing my own WSJ history, Gaffney has stumped me almost twice as many times as Shenk. And for 2020, the only ones I’ve missed have all been by Gaffney. Wrong wavelength trouble at times!
If I get a chance and/or inspiration, I will add the constructors to the spreadsheet unless Barbara or someone else is more inspired and beats me to it.
I'll just TALLY my fails and get a count that way. I suspect a 3:2 Gaffney-Schenk constant.
'tis... A lovely day for a Guinness!

User avatar
MarkL
Posts: 385
Joined: Fri Apr 12, 2019 8:44 am
Location: Tampa Bay, FL

#343

Post by MarkL » Mon Jun 15, 2020 10:23 pm

Bob cruise director wrote:
Mon Jun 15, 2020 6:59 pm
Here is the results going back to April 12, 2019 (more to come later). I only counted Mike and Matt. And I used Mike Millers correct percentage numbers.

Matt Gaffney 61 contests with 76.5% correct answers
Mike Shenk 61 contests with 77.7% correct answers

Below is the week to week chart of correct vs time. The individual weeks are all over the place but a linear trend line over the past four years shows slight improvement

I will work on priors to that another time

And yes, we are a bunch of nerds.
Wow!
Last edited by MarkL on Mon Jun 15, 2020 11:41 pm, edited 2 times in total.
'tis... A lovely day for a Guinness!

User avatar
ky-mike
Posts: 298
Joined: Thu Apr 11, 2019 4:22 pm
Location: Louisville Ky

#344

Post by ky-mike » Mon Jun 15, 2020 11:00 pm

Bob cruise director wrote:
Mon Jun 15, 2020 6:59 pm
Here is the results going back to April 12, 2019 (more to come later). I only counted Mike and Matt. And I used Mike Millers correct percentage numbers.

Matt Gaffney 61 contests with 76.5% correct answers
Mike Shenk 61 contests with 77.7% correct answers


I may be a bit slow tonight, but aren’t there about 61 total weeks from April 2019 through today? Not sure how each of them had 61 puzzles during this time frame unless I am missing out on some secret puzzles.

User avatar
Bob cruise director
Cruise Director
Posts: 1862
Joined: Thu Apr 11, 2019 2:38 pm
Location: any golf course that will take my money

#345

Post by Bob cruise director » Mon Jun 15, 2020 11:20 pm

ky-mike wrote:
Mon Jun 15, 2020 11:00 pm
Bob cruise director wrote:
Mon Jun 15, 2020 6:59 pm
Here is the results going back to April 12, 2019 (more to come later). I only counted Mike and Matt. And I used Mike Millers correct percentage numbers.

Matt Gaffney 61 contests with 76.5% correct answers
Mike Shenk 61 contests with 77.7% correct answers


I may be a bit slow tonight, but aren’t there about 61 total weeks from April 2019 through today? Not sure how each of them had 61 puzzles during this time frame unless I am missing out on some secret puzzles.
I had my IF statement wrong
It is corrected below
Bob Stevens
Cruise Director

User avatar
hcbirker
Posts: 607
Joined: Thu Apr 11, 2019 7:24 pm
Location: Studio City, CA

#346

Post by hcbirker » Mon Jun 15, 2020 11:35 pm

yourpalsal wrote:
Mon Jun 15, 2020 7:34 pm
REMINDER: Zoom meetup for members of the XWord Muggles Forum today at 5p PT/8pm ET. Hoping for a healthy TALLY!

YourPalSal’s Beach & Ship Zoom Cafe
Jun 15, 2020 05:00 PM Pacific Time (US and Canada)
Join Zoom Meeting:
https://us02web.zoom.us/j/86538717219
How did your Zoom meeting go? I was going to join but right before my son announced they had bought their first house and I was totally engulfed by text messages and updates to the situation. Hope it went well!
Heidi

User avatar
Joe Ross
Posts: 1035
Joined: Fri Apr 12, 2019 4:46 am
Location: Cincinnati

#347

Post by Joe Ross » Tue Jun 16, 2020 3:42 am

DrTom wrote:
Mon Jun 15, 2020 10:00 pm
We are all going to have to step it up a notch to match Joe' Ross's diagram though, I cannot look at it without chuckling
NOT mine!

XKCD.COM

User avatar
Joe Ross
Posts: 1035
Joined: Fri Apr 12, 2019 4:46 am
Location: Cincinnati

#348

Post by Joe Ross » Tue Jun 16, 2020 4:04 am

Cindy N wrote:
Mon Jun 15, 2020 12:57 pm
I generally do my solving "online" - with a combination of the online version of the puzzle, copying the grid into Paint to highlight what I've found and then into a Word doc to list everything out. I can color code, highlight, add things in and take them out without having to have my messy scrawl all over the page.

I had the Word doc up on the left side of my screen, with the five entries listed, and was going back and forth between that and the tab with the completed puzzle open.
Come over to the dark side...

User avatar
BethA
Posts: 264
Joined: Tue Apr 09, 2019 6:44 pm
Location: Beaver, PA

#349

Post by BethA » Tue Jun 16, 2020 7:38 am

Bob cruise director wrote:
Mon Jun 15, 2020 6:59 pm
Here is the results going back to April 12, 2019 (more to come later). I only counted Mike and Matt. And I used Mike Millers correct percentage numbers.

Matt Gaffney 61 contests with 76.5% correct answers
Mike Shenk 61 contests with 77.7% correct answers

Below is the week to week chart of correct vs time. The individual weeks are all over the place but a linear trend line over the past four years shows slight improvement

I will work on priors to that another time

And yes, we are a bunch of nerds.
Wow, Bob! Thanks for the quick response. Nerds rule!

Those percentages seem very close, and maybe surprising to those who feel they struggle more with one constructor than the other. But we must balance each other out. For example, I noticed Tony S wrote: “KAS 1 last week, KAS 5 this week”, but I had just the opposite ratings. (Never got last week’s, but got this week’s fairly quickly).

Interesting! Thx again.

User avatar
BethA
Posts: 264
Joined: Tue Apr 09, 2019 6:44 pm
Location: Beaver, PA

#350

Post by BethA » Tue Jun 16, 2020 8:44 am

Bob cruise director wrote:
Mon Jun 15, 2020 9:04 pm
Historical count

First a big thanks to Joe Ross and Steve B for putting the past crosswords in the blog. Next Thanks to Barbara Koehler for generating the spreadsheet and populating it that I built off of. And thanks to Brian Mac for setting up the blog

Prior to April 29, 2016 we don't have the data of number of entries and percentage correct so all the data below is subsequent to that date

From Easiest to hardest we have (just a week to week average and not any kind of weighted average)

Patrick Berry with 13 contests and 86.4% correct percentage
Peter Gordon with 6 contests and 82.1% correct percentage
Mike Shenk/Marie Kelly with 87 contests and 76.9% correct percentage
and Matt Gaffney with 105 contests and 71.2% correct percentage
Sorry I had skipped over this post before. It looks like there IS a little more difference in solve percentages! Will have to ponder further. Gotta run!

michaelm
Posts: 64
Joined: Fri Jan 24, 2020 3:21 pm

#351

Post by michaelm » Tue Jun 16, 2020 8:47 am

Bob cruise director wrote:
Mon Jun 15, 2020 6:59 pm
Here is the results going back to April 12, 2019 (more to come later). I only counted Mike and Matt. And I used Mike Millers correct percentage numbers.

Matt Gaffney 61 contests with 76.5% correct answers
Mike Shenk 61 contests with 77.7% correct answers

Below is the week to week chart of correct vs time. The individual weeks are all over the place but a linear trend line over the past four years shows slight improvement

I will work on priors to that another time

And yes, we are a bunch of nerds.
Appears to be good healthy competition between and among the constructors.
Thanks Mr. Director and all who contribute to keeping the stats!

User avatar
mheberlingx100
Posts: 105
Joined: Sun Jul 14, 2019 11:39 am

#352

Post by mheberlingx100 » Tue Jun 16, 2020 9:43 am

BethA wrote:
Tue Jun 16, 2020 8:44 am
Bob cruise director wrote:
Mon Jun 15, 2020 9:04 pm
Historical count

First a big thanks to Joe Ross and Steve B for putting the past crosswords in the blog. Next Thanks to Barbara Koehler for generating the spreadsheet and populating it that I built off of. And thanks to Brian Mac for setting up the blog

Prior to April 29, 2016 we don't have the data of number of entries and percentage correct so all the data below is subsequent to that date

From Easiest to hardest we have (just a week to week average and not any kind of weighted average)

Patrick Berry with 13 contests and 86.4% correct percentage
Peter Gordon with 6 contests and 82.1% correct percentage
Mike Shenk/Marie Kelly with 87 contests and 76.9% correct percentage
and Matt Gaffney with 105 contests and 71.2% correct percentage
Sorry I had skipped over this post before. It looks like there IS a little more difference in solve percentages! Will have to ponder further. Gotta run!
I don’t think the solve percentages adequately measures the difficulty of the puzzles. There are lots of solvers (me included) who don’t submit an answer unless they have some level of confidence in the solution.

Can a measure be constructed based on absolute number of correct solvers? I am assuming that the number of solvers has grown over time, so that would have to be taken into account. Any statisticians out there with time on your hands?

User avatar
TPS
Posts: 590
Joined: Sun Mar 15, 2020 2:19 pm
Location: Florida

#353

Post by TPS » Tue Jun 16, 2020 10:02 am

mheberlingx100 wrote:
Tue Jun 16, 2020 9:43 am
I don’t think the solve percentages adequately measures the difficulty of the puzzles. There are lots of solvers (me included) who don’t submit an answer unless they have some level of confidence in the solution.
This is probably true - I know I also do not submit unless I am 75%+ sure I have the right answer and even if I have the right answer I don’t submit unless I get it 95%+ on my own.

User avatar
Bob cruise director
Cruise Director
Posts: 1862
Joined: Thu Apr 11, 2019 2:38 pm
Location: any golf course that will take my money

#354

Post by Bob cruise director » Tue Jun 16, 2020 11:30 am

Eric Porter wrote:
Mon Jun 15, 2020 9:48 pm
For this puzzle, I saw COUNT in the title, then DOS in 17A & SIEZE in 64A, then remembered 61D. Taking French in high school helped.I thought that with 5 theme answers, a 5 letter word, and 5 languages that people would have an easier time with it. I didn't even think to sum the numbers and went directly to TALLY. The hardest Shenk puzzle for me this year was Variety Show, which I did get, but slowly. If I subscribed to MGWCC, I think I'd do much better on Gaffney's WSJ puzzles.

Bob, have you run any statistics on how many people correctly solve the puzzles from the different creators?
see my post 335 and the chart on 326
Bob Stevens
Cruise Director

User avatar
Bob cruise director
Cruise Director
Posts: 1862
Joined: Thu Apr 11, 2019 2:38 pm
Location: any golf course that will take my money

#355

Post by Bob cruise director » Tue Jun 16, 2020 11:34 am

mheberlingx100 wrote:
Tue Jun 16, 2020 9:43 am
BethA wrote:
Tue Jun 16, 2020 8:44 am
Bob cruise director wrote:
Mon Jun 15, 2020 9:04 pm
Historical count

First a big thanks to Joe Ross and Steve B for putting the past crosswords in the blog. Next Thanks to Barbara Koehler for generating the spreadsheet and populating it that I built off of. And thanks to Brian Mac for setting up the blog

Prior to April 29, 2016 we don't have the data of number of entries and percentage correct so all the data below is subsequent to that date

From Easiest to hardest we have (just a week to week average and not any kind of weighted average)

Patrick Berry with 13 contests and 86.4% correct percentage
Peter Gordon with 6 contests and 82.1% correct percentage
Mike Shenk/Marie Kelly with 87 contests and 76.9% correct percentage
and Matt Gaffney with 105 contests and 71.2% correct percentage
Sorry I had skipped over this post before. It looks like there IS a little more difference in solve percentages! Will have to ponder further. Gotta run!
I don’t think the solve percentages adequately measures the difficulty of the puzzles. There are lots of solvers (me included) who don’t submit an answer unless they have some level of confidence in the solution.

Can a measure be constructed based on absolute number of correct solvers? I am assuming that the number of solvers has grown over time, so that would have to be taken into account. Any statisticians out there with time on your hands?
I agree with your comment on solve percentages as an absolute criteria of difficulty. I am not sure what is a good criterion for the reason that you gave. I think that wrong answers either come, as we saw this week, from sound logic that was flawed leading to an incorrect answer or if you have a finite number of answers such as last week with a president, some people will just throw in a WAG.

I will post a few charts in a separate message.
Last edited by Bob cruise director on Tue Jun 16, 2020 6:21 pm, edited 1 time in total.
Bob Stevens
Cruise Director

MikeMillerwsj
Posts: 92
Joined: Tue Apr 09, 2019 2:46 pm

#356

Post by MikeMillerwsj » Tue Jun 16, 2020 11:45 am

This was another tricky one--we had 1009 entries with about 2/3 correct. Lots of solvers were allured by five-letter numbers (just because of the title "Head Count"? or another reason?): FORTY (76), THREE (32), EIGHT (32) and SEVEN (29). Plus three for SIXTY and one for FIFTY. (Bonus question: Is that all the five-letter numbers?)

Congrats to this week's winner: Andrew Bradburn of North Hollywood, Calif.!

User avatar
Bob cruise director
Cruise Director
Posts: 1862
Joined: Thu Apr 11, 2019 2:38 pm
Location: any golf course that will take my money

#357

Post by Bob cruise director » Tue Jun 16, 2020 11:45 am

Here are some charts. The good about Excel is that you can plot anything against anything once you have the data in the files. The question is whether the information is of any value. I will let you decide if this or anything else is more than idle curiosity.
Attachments
Submissions per week.jpg
Correct vs On shore.jpg
On Shore percentage vs time.jpg
On shore number vs time.jpg
Bob Stevens
Cruise Director

MikeMillerwsj
Posts: 92
Joined: Tue Apr 09, 2019 2:46 pm

#358

Post by MikeMillerwsj » Tue Jun 16, 2020 11:46 am

PS I love the spreadsheet analysis underway here! Big thanks to our Muggle data miners.

User avatar
Bob cruise director
Cruise Director
Posts: 1862
Joined: Thu Apr 11, 2019 2:38 pm
Location: any golf course that will take my money

#359

Post by Bob cruise director » Tue Jun 16, 2020 11:48 am

MikeMillerwsj wrote:
Tue Jun 16, 2020 11:45 am
This was another tricky one--we had 1009 entries with about 2/3 correct. Lots of solvers were allured by five-letter numbers (just because of the title "Head Count"? or another reason?): FORTY (76), THREE (32), EIGHT (32) and SEVEN (29). Plus three for SIXTY and one for FIFTY. (Bonus question: Is that all the five-letter numbers?)

Congrats to this week's winner: Andrew Bradburn of North Hollywood, Calif.!
I hope that Matt, Mike et al. are getting a good laugh at all the data on these metas being pumped out. All of dubious value probably
Bob Stevens
Cruise Director

User avatar
BrianMac
Site Admin
Posts: 548
Joined: Mon Apr 08, 2019 11:45 pm
Location: Connecticut

#360

Post by BrianMac » Tue Jun 16, 2020 12:31 pm

hcbirker wrote:
Mon Jun 15, 2020 11:35 pm
How did your Zoom meeting go? I was going to join but right before my son announced they had bought their first house and I was totally engulfed by text messages and updates to the situation. Hope it went well!
It was great! Nice to put some faces (and voices) to names. There was only a handful of us, but we had a nice talk about everything from solving techniques, to crossword tournaments, crossword constructing and the difference between Gaffney and Shenk. Thanks to Sal for setting it up. He mentioned that he might try to set up another one next week, which I hope he does. Hopefully we can get some more people to join...don't be shy!!

Locked