# When will @egreg reach 654321

And the winner is: yo'! Congratulations!

As prize yo' selected a scout duck with neckerchief:

(it is already available in the development version of the tikzducks package and will be included in the next release)

In this little game you have to guess when egreg will reach a reputation of 654321 (or higher) on TeX.Stackexchange.

## Rules

1. Put your guess in an answer (not comment) to this question before 15 January, 2018 (23:59 UTC).

2. The entry with the closest prediction (either + or -) will win.

3. Please add your time as precisely as possible (up to seconds). Missing values for hours/minutes/seconds will be replaced by 0.

4. If two entries state the exact same date and time, the one posted first will be considered.

5. Unless otherwise stated, all answers are assumed to be UTC (if you answer in some other time, it would be nice if you would already include a translation into UTC). The frame of references is Earth, exposing @egreg to time dilation (putting him on a rocket to Alpha Centauri, tuning his bike...) will not change anything.

6. Only one answer per user.

7. Contestants must have been a member of TeX.SE before 26 December, 2017.

8. @egreg is not allowed to participate (sorry :), but of course you can add an answer for fun!

9. I am excluded as well, my answer below is just for fun.

10. Time lords in possession of a working Tardis are also excluded.

11. Participants are disqualified if they offer a bounty specifically for @egreg - this means if someone offers a bounty to attract attention to a question and @egreg answers it, that's fine, offering a bounty to reward an existing answer from @egreg is also fine, you just cannot win this contest.

12. the winner will only be announced a day after @egreg reached this reputation, just in case some serial upvoting or similar will be corrected.

1. Don't change your voting behaviour because of this contest - don't try any serial voting for @egreg. This would not help, as @egreg is reputation capped most days anyway and serial voting is automatically reversed by the site!
1. Any legal action is precluded.

## Prize

The winner may wish an addition to the tikzducks package. For example a new accessories, some kind of hat, clothing, a duck based on you or your avatar ... (I reserve the right to reject a wish because of moral scruples).

## Currently submitted entries

Format is YYYY-MM-DD hh:mm:ss

• 2018-03-14 15:09:26 RobtAll
• 2018-04-01 19:00:00 Tobia Tesan
• 2018-04-20 11:10:28 (corey979)
• 2018-05-01 11:23:47 Sigur
• 2018-05-04 16:04:04 CarLaTeX
• 2018-05-05 17:05:00 marmot
• 2018-05-06 03:14:15 (samcarter)
• 2018-05-06 19:04:03 mrclary
• 2018-05-07 07:05:18 BMWurm
• 2018-05-08 08:15:32 Paulo Cereda
• 2018-05-12 12:34:56 Ulrike Fischer
• 2018-05-13 07:12:00 (swarm intelligence)
• 2018-05-13 14:55:25 thewaywewalk
• 2018-05-17 10:52:00 Alan Munn
• 2018-05-17 14:58:36 anderstood
• 2018-05-18 08:05:08 yo'
• 2018-05-19 15:44:38 egreg crossed the 654321 threshold
• 2018-05-22 23:55:42 koppor
• 2018-05-23 12:00:00 Werner
• 2018-05-23 22:59:39 Axel Sommerfeldt
• 2018-05-25 18:57:00 Sverre
• 2018-06-01 00:00:00 Raoul Kessels
• 2018-09-08 08:13:32 A Feldman
• 2019-01-01 00:00:00 David Carlisle
• ????-??-?? ??-??-?? (egreg)

After excluding the first and last two entries the mean value is MJD 58251.3 -> let's see if swarm intelligence really works :)

(please tell me if I missed anything obvious in the rules section -- this is my first contest)

• How about shifting the target to 654323, which is a prime number? – user121799 Dec 26 '17 at 20:43
• @marmot Good idea, however as this prediction started in chat some while ago, I'd like to stick with the nice number sequence. – samcarter is at topanswers.xyz Dec 26 '17 at 20:45
• I'm ROFL, great idea! I'll do some calculations and answer :) – CarLaTeX Dec 26 '17 at 20:45
• @marmot I prefer 654321, too: the goal is nearer :) – CarLaTeX Dec 26 '17 at 20:46
• I think this will encourage serial voting when he gets close to the value. No matter what it will encourage voting for the person not the answers (the fact that the answers are independently worth voting for is irrelevant.) – Alan Munn Dec 26 '17 at 21:47
• @AlanMunn serial voting will not work as egreg is rep capped most days. – samcarter is at topanswers.xyz Dec 26 '17 at 21:50
• But quickly after obtaining 654321 he will reach 666666. Sodom and Gomorrah! ;-) – Przemysław Scherwentke Dec 27 '17 at 4:59
• @PrzemysławScherwentke In the early ages of the site, it must have survived 66666, so I am optimistic about 666666 :) – samcarter is at topanswers.xyz Dec 27 '17 at 12:14
• I think that a rule shall be added that participants cannot reward @egreg a bounty after January 15th. – yo' Dec 27 '17 at 14:10
• @yo' I think this is too general. Imagine a bounty to attract attention to a question, egreg answers and even if the OP does not award the bounty to egreg, the system will auto-award half of the bounty. That's why I formulated rule #11 that no targeted bounty is allowed. – samcarter is at topanswers.xyz Dec 27 '17 at 17:42
• re rule "14, Any legal action is precluded.", does this imply that only illegal actions are permitted? – barbara beeton Dec 27 '17 at 20:43
• @samcarter -- oh, i was just being a nuisance. <grin> i know this really means that no lawsuits are permitted in case of disagreement. the judge's decision is final. i think the concept is great (if a bit silly). but that's the ethos of this group. – barbara beeton Dec 27 '17 at 21:37
• ooh I can sue people if I don't win? Interesting... – Paulo Cereda Dec 28 '17 at 12:18
• @PauloCereda no :) well, maybe. I don't know which laws apply to ducks :) – samcarter is at topanswers.xyz Dec 28 '17 at 12:52
• A "ducky defendant" indeed. – A Feldman Jan 1 '18 at 19:18

18-05-18 08:05:08 UTC

• Thank you for participating! – samcarter is at topanswers.xyz Dec 27 '17 at 17:45
• Please let us know which is your request about tikzducks! – CarLaTeX May 20 '18 at 16:46
• Congratulations!!! – Raoul Kessels May 21 '18 at 11:17
• @CarLaTeX If you follow the chat, it's obvious. If not, you'll have to wait until the release :-) – yo' May 21 '18 at 15:59
• @yo' I follow the chat :):):) – CarLaTeX May 21 '18 at 16:44

Like Renaissance mathematicians did, I'll give my guess in encrypted form so as not to influence other participants. I expressed my guess in the format YYYY-MM-DD and issued

echo ZZZZZZ YYYY-MM-DD | shasum -a 256


where ZZZZZZ is my old landline phone number (six digits) getting the fingerprint

8ab685f7a789f83d8c79f770e8c085be847a99b2c0feb574ca81e06771f354d1


This is to make it useless to type in all dates in order to get a matching SHA256 checksum.

When the contest will be finished, I'll reveal my guess, so anybody can check the matching of the data with the checksum. The old phone number cannot be used any longer, so it will be irrelevant information when revealed.

### UPDATE 2018-05-19

It turns out I was a bit optimistic in my estimate:

655401 2018-05-13


Congratulations to yo' for being the winner!

• Thank you for joining the fun! Great idea to encrypt your guess! – samcarter is at topanswers.xyz Dec 26 '17 at 21:34
• WHERE ARE MY RAINBOW TABLES? Time for some duck computations. :) – Paulo Cereda Dec 26 '17 at 22:20
• @egreg Uh, you're right, that wasn't very nice of me. I'm sorry. Keep in mind to not underestimate modern computing power next time. I should've posted your guess hashed with the phone number reversed instead: 8c327d8fd72f490cdb0b6b363e6105693f1c23bcc191fb13951e90ec83579e8c. Figuring this out was fun. I didn't expect the actual cracking to take 13 seconds, to be honest. – NieDzejkob Dec 28 '17 at 12:19
• @NieDzejkob Cracking it knowing precisely the format is of course very easy. – egreg Dec 28 '17 at 13:18
• @NieDzejkob: thankfully phone numbers and dates have a very restrict range of possible symbols. :) – Paulo Cereda Dec 28 '17 at 14:41
• Your intelligence gives exactly the same result as the swarm intelligence... – Paul Gaborit Feb 24 '18 at 9:08
• Since people can no longer participate. Will you post your answer? Or will you wait until you have reached the goal? – Thorbjørn E. K. Christensen Apr 4 '18 at 11:57

I think he will never get to the target in 2018 due to several thousand questions about % at end of line being reclassified as duplicates and deleted resulting in a loss of most of his Rep.

If I win I think the tikzducks should be re-implemented in picture mode.

• Thank you for participating! Oh, the football is even in colour - fancy! – samcarter is at topanswers.xyz Dec 27 '17 at 12:12
• @samcarter it's not a football it's a ham and pineapple pizza, in honour of egreg! – David Carlisle Dec 27 '17 at 12:16
• Oh, please excuse my inexperience in high art :) The pizza of course makes perfect sense! – samcarter is at topanswers.xyz Dec 27 '17 at 12:18
• @samcarter How could you miss this answer: tex.stackexchange.com/a/390382/101651 to the tikzduck question! – CarLaTeX Dec 27 '17 at 12:21
• @CarLaTeX :) Somehow I only remembered the picture mode football duck from this answer - my subconsciousness must have erased the pineapple pizza - too traumatic to remember! – samcarter is at topanswers.xyz Dec 27 '17 at 12:33
• Somebody want's to be number 1? I would totally use that duck everywhere! – Thorbjørn E. K. Christensen Dec 30 '17 at 23:33

Based on this query and linear extrapolation, it should be

May 13th, 2018, 14:55:25

• Thank you for participating! Oh, let's see if science is right! – samcarter is at topanswers.xyz Dec 27 '17 at 17:46
• What has science ever done for us? /sent from my smartphone :) – Paulo Cereda Dec 27 '17 at 18:42

2018/05/05 05:05 pm

... because that's when the marmots awake from hibernation ;-)

• ooh advanced hibernation techniques from the marmot society! :) – Paulo Cereda Dec 27 '17 at 17:00
• Thank you for participating! A really good reason to choose the date :) I am just wondering how you added this answer during hibernation :) – samcarter is at topanswers.xyz Dec 27 '17 at 17:45
• @samcarter: ooh marmot mysteries – Paulo Cereda Dec 27 '17 at 17:52
• @samcarter Hibernating for 8 months can be boring, so I play with LaTeX once in a while ;-) – user121799 Dec 27 '17 at 18:52
• @marmot Oh, that explains it :) – samcarter is at topanswers.xyz Dec 27 '17 at 18:56

May 23, 2018 at around 12:00 noon.

This is the cumulative, annual reputation of @egreg, post 2011 (the first year of joining):

• 2011: 54,140
• 2012: (+ 92,699; +71.2%) 146,839
• 2013: (+ 98,357; +6.1%) 245,196
• 2014: (+ 97,496; -0.9%) 342,692
• 2015: (+ 96,697; -0.8%) 439,389
• 2016: (+ 92,547; -4.3%) 531,936
• 2017: (+ 88,897; -3.9%) 620,833

I used the top-ranked 2D function suggestion via ZunZun's function finder:

y = c * atan(exp((x-a)/b)) + Offset

Fitting target of lowest sum of squared absolute error = 2.0467080344324614E+05

a =  2.0128116742956618E+03
b =  8.0055160074372509E+00
c =  1.5744218139080477E+06
Offset = -1.0100039377450938E+06


Solving for y = 654321 yielded a solution of x = 2017.3933394731732, or halfway into day 144 of 2018. This lies somewhere between May 23rd and 24th (2018).

• Wow, very scientific, and we discover he has a decreasing trend during the last years! – CarLaTeX Jan 2 '18 at 6:32
• Did you consider the loss of reputation due to deletion of user P. in February 2017? – user31729 Jan 2 '18 at 8:03
• Thank you for participating! I am crossing my fingers, science has to proof right! – samcarter is at topanswers.xyz Jan 2 '18 at 9:45
• @ChristianHupfer: No. The minimum rep-per-day since 2012 has never seemed to that low. – Werner Jan 2 '18 at 17:01
• @ChristianHupfer who is/was user P. (never heard of him, but if he can lower egregs reputation with deleting his account, he must've been busy around here)? – Skillmon likes topanswers.xyz Jan 7 '18 at 13:43
• @Skillmon: A user named Papiro ;-) He was an avid voter, but had not much reputation profile (about 3k, if I remember correctly( though, as such, his votes were lost. My reputation was cut by roughly 5500 then – user31729 Jan 7 '18 at 13:45

# 1525767332

Since I already know I am going to win, you can start working on my prize: I want a EG duck to rival our DPC duck. :)

• Thank you for participating! Quack! – samcarter is at topanswers.xyz Dec 26 '17 at 22:28
• @samcarter: thanks for the great game! Quack! <3 – Paulo Cereda Dec 26 '17 at 22:29
• You can have your prize right away: \documentclass{standalone} \usepackage{tikzducks} \definecolor{egreg}{RGB}{171,196,163} \begin{document} \begin{tikzpicture} \duck[] \fill[egreg] (current bounding box.north west) rectangle (current bounding box.south east); \end{tikzpicture} \end{document} – samcarter is at topanswers.xyz Dec 26 '17 at 22:42
• @samcarter the motorbike is missing. – Ulrike Fischer Dec 26 '17 at 22:53
• @UlrikeFischer The bike will be delivered upon winning the contest :) – samcarter is at topanswers.xyz Dec 26 '17 at 22:59

My guess: April 1st, 2018, 19:00:00 UTC.

Just in case some well thought out April fools' joke gains the guy a reputation boost.

The majority seems to be very sure that it will happen in May, but since I am in Chemistry and lower temperature gives lower reaction rates, the winter should slow down the daily reputation increase ;-). So I will place my guess a bit ahead:

2018-06-01 00:00:00

• Thank you for participating! Interesting idea! On the other hand winter is not ideal for bike tours - so maybe reputation gain is actually faster during winter? – samcarter is at topanswers.xyz Dec 28 '17 at 19:41
• Good point. We shall see! – Raoul Kessels Dec 29 '17 at 9:18
• I wish I were a chemist, so I'd dedicate my research to making things explode all the time. :) ZE GOOGLES, ZEY DO NOTHING! – Paulo Cereda Dec 29 '17 at 9:18

My guess is:

2018-05-23 22:59:39 UTC

(assuming egrep will earn an average rep. of 234 per day)

• Thank you for participating! – samcarter is at topanswers.xyz Dec 28 '17 at 23:08
• Is "egrep" a Freudian slip? :-) – Mico Dec 29 '17 at 8:19
• @Mico: worst: it's a command line tool! :) – Paulo Cereda Dec 29 '17 at 9:17

I stored egreg's last 90 days rep (from this page source), then used Mathematica commands:

fit = FindFit[data, a*t + b, {a, b}, t]
Show[ListLinePlot[data, PlotStyle -> Red],
Plot[a*t + b /. fit, {t, 0, Length@data}, PlotStyle -> Dashed]]


That gives the following plot (data in red, fit in dashed blue)

and date:

n = -Length@data + (654321 - b)/a /. fit;
DatePlus[DateObject[{2018, 1, 5}, TimeZone -> 0], n]


2018-05-17 14:58:36

Also, 2 StandardDeviation[Differences[data]]/Sqrt[n] returns about 6, so, assuming the daily rep follows a normal distribution, the 95% confidence interval is plus or minus 6 days.

Saturday, 12 May (2018 naturally), 12:34:56

If I win I want a vampire duck

May the royal flush magic do its work!

As approximately stated in https://chat.stackexchange.com/transcript/41?m=40710281#40710281 my guess is

Sunday, 6 May 2018, 03:14:15 am (UTC)

My guess is:

May 17, 2018 10:52:00

Since I don't know if there is some limit for daily reputation I am not able to try to predict.

So I'll answer just for fun (and of course based on some other people's vote).

2018-05-01 11:23:47

May the Force be with egreg!

Friday, 4 May 2018, 04:04:04 pm (UTC)

If I win, I would like to have a girlfriend, or, better, a female-duck-friend, for Herr Professor Paulinho van Duck.

• Nice choice! Thanks for participating! – samcarter is at topanswers.xyz Dec 26 '17 at 21:15
• @samcarter Yay! :):):) I thought that he could earn more reputation at the beginning of the year rather than in the Summer... So the per day average is higher... – CarLaTeX Dec 26 '17 at 21:18

Obviously 2018-03-14 15:09:26, Using 09 instead of 92.

• Thank you for participating! – samcarter is at topanswers.xyz Dec 30 '17 at 1:02
• If you round instead of truncating, that becomes 15:09:27. I hope that won't make you loose the first prize. – anderstood Jan 9 '18 at 17:34
• @anderstood What, are there more digits to pi ? I can only count to 8 digits, because I need my other two fingers for typing. – user139954 Jan 9 '18 at 17:52

Using anderstood's data from here or here, I used Mathematica's FindFormula:

data2 = Transpose@{Range[Length@data], data};
ff = FindFormula[data2, x, 10]


Plot[Evaluate@{ff[[2 ;; 10]], 654321}, {x, 1500, 1650}, Frame -> True]


pred = Table[x /. FindRoot[ff[[i]] == 654321, {x, 1560}], {i, 2, 10}]
ListPlot[pred, Frame -> True, PlotRange -> All, DataRange -> {2, 10}]


time = Mean@pred - Length@data


to get

DatePlus[DateObject[{2018, 1, 5}, TimeZone -> 0], time]


Fri 20 Apr 2018 11:10:28 GMT

• *2018 (comments must be at least 15 characters in length) – Jason B. Jan 5 '18 at 19:21
• @JasonB. Yup, thanks. – corey979 Jan 5 '18 at 19:32
• I get 20 Apr 2018 11:10:23 with my answer, but using all data points instead of the last 90. Not too different! – anderstood Jan 5 '18 at 19:48
• Thank you for participating and welcome to TeX.Stackexchange! Unfortunately you do not fulfil rule #7, so I added your answer in () -> you're more than welcome to see how good your prediction is, but some other participant will be declared winner. (we can of course negotiate some consolation prize in case you have the best forecast :) – samcarter is at topanswers.xyz Jan 5 '18 at 20:25
• @samcarter No, no, I didn't read the rules carefully, just jumped into computations. So I'll just see what will the answer be. In case my guess turns out to be not-so-bad, an honorary mention would be enough :) – corey979 Jan 5 '18 at 21:05
• @corey979 We'll see when the time comes :) – samcarter is at topanswers.xyz Jan 5 '18 at 21:29
• Very nice to see how people are working hard to try to answer a so simple question.... lol Very cool your post. – Sigur Jan 8 '18 at 20:08

Using anderstood's data, I noticed that egreg consistently required about 120 days from Jan 4 to increase is reputation by 32289 for each of the first 3 years; last year was an outlier at 128 days. You see, one has to consider seasonal effects ;)

I'm going with 2018-05-06 19:04:03, 121.79 days after Jan 5 00:00:00

• Nice! I had not seen any periodic seasonal effects, even though I checked (including after smoothening). – anderstood Jan 6 '18 at 1:27
• Thank you for participating! Just a few hours from my own guess :) – samcarter is at topanswers.xyz Jan 6 '18 at 17:26
• The residual fit here does not display seasonal effects (image from this answer). – anderstood Jan 9 '18 at 17:35

My guess is: 2018-09-08 08:13:32

2018-05-22 23:55:42

This date diff between beginning of JabCon and that date is 106 if one includes start and end date. I plan to have published JabRef 4.2 on that date. 100 days is the normal time one allows to settle down (like governments). 100 + 4 + 2 = 106. 23:55 is 5 minutes before twelve.

The scientific-thesis-template contains two (out of 16) questions, where egreg took part (http://tex.stackexchange.com/a/50638/9075, How to display cover page alone in the "two page view" mode). 2+16 are the seconds missing to midnight. Thus, 60 - (2+16) forms the seconds.

• It would be much more fun if you would have found some far fetched connection between your advertisement and your guess :( – samcarter is at topanswers.xyz Jan 9 '18 at 16:10
• Mea culpa. I updated the answer. – koppor Jan 9 '18 at 22:24
• Great explanation! Good luck for your guess and thank you for participating! – samcarter is at topanswers.xyz Jan 9 '18 at 22:36
• Looking at the other answers so far, there seemed to be a strong consensus for it happening in May, so I went with that.

• The 7th wasn't yet taken, so I thought I'd pick that one.

• For the time only a palindrome could ever be correct, so it picked itself:

 (20)   18-05-07 07:05:18    (,02 seconds)


May 25, 2018, 18:57 UTC. (extra characters here)

• Thank you for participating! – samcarter is at topanswers.xyz Jan 15 '18 at 19:22
• And when will Sverre reach 9999? – anderstood Jan 15 '18 at 23:11
• @anderstood Can't you please estimate that with your math skills as you did for egreg? :) – Sverre Jan 16 '18 at 8:46