mersenneforum.org > Data Let's finish primality verification through Mp#49*, M(74 207 281)
 Register FAQ Search Today's Posts Mark Forums Read

2021-10-14, 15:09   #34
kriesel

"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

2×3×13×83 Posts

Quote:
 Originally Posted by Zhangrc It's time to add another term for http://oeis.org/A000043!
Not yet updated. How is that done?

Last fiddled with by kriesel on 2021-10-14 at 15:13

2021-10-14, 15:33   #35
chris2be8

Sep 2009

13×179 Posts

Quote:
Possibly because 74209207 is composite:
Code:
\$ factor 74209207
74209207: 331 224197

2021-10-15, 17:22   #36
kriesel

"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

647410 Posts
Estimates of primality test verification completion

1. Participants may post one estimate of verification completion date per known Mp#i*.
2. Entries are open now for Mp#49 - Mp #51*. Entries for Mp #52* are not open, until after its exponent is known, the primality is redundantly confirmed, and an official Mersenne Research Inc. press release is issued announcing the discovery.
3. Participants are encouraged to give ISO-9001 dates, or ISO 9001 date-time strings if especially ambitious.
4. First estimate per (forum user ID, Mp #i*) combination is the only one that counts. No revisions will be accepted.
5. Estimates posted after a date that is 31 days before the actual date of verification completion of all Mersenne numbers up to the relevant Mp#i*s are invalid.
6. Interpretation of estimate dates and times posted will be as follows:
1. Dates posted will be interpreted as Gregorian calendar
2. If an estimate is given imprecisely such as first half / quarter / whatever of a year or other time period:
1. the midpoint of the specified period is considered to be the estimate date value.
2. In case of a fraction, as may occur if the period contains an odd number of days, date value shall be rounded upward. For example, first half of 2024 would be evaluated as if the estimator had stated roundup (2024-01-01 + 2024-06-30) / 2 = 2024-01-01 + roundup(181/2) = 2024-01-01 + 91 =2024-04-01. (2024 is a leap year.)
3. Date evaluations will be performed with https://www.timeanddate.com/date/duration.html while that URL remains functional. If it ceases to function for an extended period, another functional URL will be sought and posted.
4. If time of day is also included in the estimate, time zone will be presumed to be UTC unless otherwise stated.
5. If time of day is not included in the estimate, time zone will be presumed to be US Central unless otherwise stated.
6. Estimate dates posted without attached times will be treated as yyyy-dd-mm 0:0:0.0 for purposes of computing time differences and other values.
7. Actual dates of completion of verification as posted at https://www.mersenne.org/report_milestones/ shall be definitive. (No room for debate on actual date values or TZ effects.)
8. Verification of the test status of an exponent is by whatever tests are accepted by mersenne.org at the time; matching res64 values for LL or PRP of the same type; PRP with Cert; found factors; future tests if any. (No room for debate on what constitutes verification or does not.)
9. Highest quality estimate is that with maximum value for a given Mp#i* of the function abs ((actual date – post date) / (actual date – estimate date)) and for which post date < (actual date - 30.99999 days ).
10. Closest estimate is that with minimum value for a given Mp#i* of the function abs (actual date – estimate date) and for which post date < (actual date - 30.99999 days).
11. In the unlikely case of a tie on evaluation values, the earliest posted estimate wins. In the case of a tie between estimates posted the same day, earlier post time and order is the first tiebreaker. Second tiebreaker is earlier estimate value. Third tiebreaker is later forum join date.
Attached Files
 Mpnn verification forecasts.pdf (20.0 KB, 64 views)

2021-10-16, 21:11   #37
greenskull
Xebeche

Apr 2019
🌺🙏🌺

1BD16 Posts

Quote:
 Originally Posted by kriesel Stake your claims..
My prediction for Mp#49*, M(74 207 281) was March 08, 2025.

You marked my prediction in your PDF file as 2028-03-08. It is incorrect.
It should be 2025-03-08.

2021-10-16, 21:35   #38
greenskull
Xebeche

Apr 2019
🌺🙏🌺

1BD16 Posts

Quote:
 Originally Posted by kriesel Stake your claims..
Also, you have specified your forecast for Mp#49*, M(74 207 281) with the date 2021-08-26 and link:
https://www.mersenneforum.org/showpo...5&postcount=11

But I did not find such a forecast there.
I have seen the forecast 2025-01-18 in attached PDF there but it dated by 10/13/2021 and not by 2021-08-26.

It seems to me better to be more careful with numbers.

Last fiddled with by greenskull on 2021-10-16 at 21:42

2021-10-17, 01:16   #39
kriesel

"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

145128 Posts

Quote:
 Originally Posted by greenskull My prediction for Mp#49*, M(74 207 281) was March 08, 2025. https://mersenneforum.org/showthread...619#post589619
That typo will be corrected at next posting of the pdf. My spreadsheet is corrected now. I plan to update the pdf in the thread only ~monthly, or less frequently if there's no estimation activity and no milestone reached. An astute reader of the rules I drafted may have already noted that they are constructed to limit the effort required of me going forward.

Quote:
 Originally Posted by greenskull Also, you have specified your forecast for Mp#49*, M(74 207 281) with the date 2021-08-26 and link: https://www.mersenneforum.org/showpo...5&postcount=11 But I did not find such a forecast there. I have seen the forecast 2025-01-18 in attached PDF there but it dated by 10/13/2021 and not by 2021-08-26. It seems to me better to be more careful with numbers.
On that one, you just don't know what you're talking about.

1) https://www.mersenneforum.org/showpo...5&postcount=11 is a reference post originated 2018-12-16, 12:03, that has been regularly updated numerous times over the years following, and (relative to now) was last updated 2021-10-13 at 12:45 (my time zone). The page 5 was added to one of the attachments updated, at a previous update, so page 5 got first posted there 2021-08-26.

2) The post date of an estimate does not change. It is an estimate's first appearance that counts, not any other time it gets reposted or mentioned, or remains in an updated post. (Read the rules, again.)

3) The update that added page 5 and multiple estimate computations was demonstrably after 2021-08-16, at which point only Mp#48* occupied page 5: see the attachment of https://mersenneforum.org/showpost.p...postcount=3436.

4) The date I saw and took note of on the previous version of the pdf on my server before generating the 2021-10-13 update pdf for updating was 2021-08-26.

5) Also https://mersenneforum.org/showpost.p...&postcount=128 posted 2021-10-06 makes explicit reference to the 2021-08-26 update with a link. "Projections for Mp#49*-51* were previously posted 2021-08-26 and you were notified of that."

6) Also, I update https://www.mersenneforum.org/showpo...5&postcount=11 when a milestone is reached. A previous milestone was reaching 57M DC completion, which https://www.mersenne.org/report_milestones/ shows as "2021-08-26 All tests below 57 million verified."

7) Another update would have been triggered by https://www.mersenne.org/report_milestones/ "2021-10-07 All exponents below 105 million tested at least once."

8) The next update would have been triggered by https://www.mersenne.org/report_milestones/ "2021-10-13 All tests below 58 million verified."

9) Whichever minor milestone is reached next, for which I then post an update of https://www.mersenneforum.org/showpo...5&postcount=11, will also change its modified-date. Which will also have nothing to do with when the Mp#49*-Mp#51* estimates were first posted.

10) Unfortunately we can not further confirm that 2021-08-26 date with the wayback machine, because archive.org does not have anything for that URL. It only does top level of mersenneforum.org, not individual posts. Even the data subforum comes up empty: http://web.archive.org/web/202108261...splay.php?f=21

11) Page 5 describes the fit on which the Mp#49*-Mp#51* estimates were computed. That fit is based on data for milestones occurring 2010-2020. Updating milestones data progress of 2021 or later has no effect on that.

12) There is no connection between the spreadsheet cells concerning 2021 milestones reached, and the estimate computations for Mp#49*-Mp#51*. The estimates computed on page 5 of the attachment are not and will not be changed by future updates of the earlier pages.

Please stop posting false claims and faulty reasoning. The 2021-08-26 date corresponds to reality. Your false claim 2021-10-13 does not.

Last fiddled with by kriesel on 2021-10-17 at 01:26

2021-10-17, 08:51   #40
greenskull
Xebeche

Apr 2019
🌺🙏🌺

5×89 Posts

Quote:
 Originally Posted by kriesel On that one, you just don't know what you're talking about.
I know very well what I am talking about :)

Your formula for estimating the quality of a participant's forecast uses the date that the forecast was made right?
And the rating of participant's forecast depends on it. The earlier forecast date gives an advantage.

Quote:
 Originally Posted by kriesel 9. Highest quality estimate is that with maximum value for a given Mp#i* of the function abs ((actual date – post date) / (actual date – estimate date)) and for which post date < (actual date - 30.99999 days ). 10. Closest estimate is that with minimum value for a given Mp#i* of the function abs (actual date – estimate date) and for which post date < (actual date - 30.99999 days).
You have written that your forecast for 2025-01-18 was made on 2021-08-26. But this does not follow from anywhere.
It can't be clearly verified and confirmed.

1) Ok.
2) Ok.
3) There is no forecast for Mp#49*, there is an approximate estimate of the year.
4) Do I just need to trust you, or can I see it somehow?
5) This post links to another post that has your prediction dated 10/13/2021.
6) Duplicates point 5)
7) Not directly related to the question.
8) Not directly related to the question.
9) Any bet must be documented. You cannot appeal to something that cannot be verified or confirmed.
10) ¯\_(ツ)_/¯, no link - no bet.
11) Not directly related to the question.
12) Not directly related to the question.

Quote:
 Originally Posted by kriesel Please stop posting false claims and faulty reasoning. The 2021-08-26 date corresponds to reality. Your false claim 2021-10-13 does not.

What is the fallacy of this claim?

If you cannot clearly confirm this date 2021-08-26, then you have to accept the date of your attached file (10/13/2021) where this your forecast is mentioned and anyone can see it. Then there will be more clarity.

I thank you for correcting my forecast soon. But taking into account the fact that you are easily confusing (changing?) numbers as in my forecast, I cannot agree with the date 2021-08-26 of your forecast, which you cannot confirm in any way.
Sorry :)

Last fiddled with by greenskull on 2021-10-17 at 09:37

2021-10-17, 09:54   #41
greenskull
Xebeche

Apr 2019
🌺🙏🌺

5·89 Posts

And further.
In your rules for evaluating predictions, we cannot adjust our earlier predictions and make new bets.

Quote:
 Originally Posted by kriesel 4. First estimate per (forum user ID, Mp #i*) combination is the only one that counts. No revisions will be accepted.
Taking into account the fact that the deadlines for their execution are measured in years, this is too rigid a limitation.
It may happen that within six months the fact deviates from the predicted trajectory and it becomes obvious that the forecast of one or another participant will not work even close.
Such a participant will have no choice but to yearn for the rest of the time - for several years, without having the opportunity to cheerfully continue to participate in the game.

Earlier, I proposed a slightly different system that allows as many predictions from the participants as they want. And this system also objectively assesses their talent for forecasting.
And this system allows the participants to correct their forecasts on the way to Moment of Truth and make new bets. And the result (Diviner Talent Rate) will take into account all the attempts.

https://www.mersenneforum.org/showth...848#post585848

You once said that it is all for fun.

Quote:
 Originally Posted by kriesel It's all just for fun.
So let's change the system to the one I suggested so that it is really fun for everyone right down to the Moment of Truth.
The 30.99999 days amendment can also be taken into account.

Last fiddled with by greenskull on 2021-10-17 at 10:13

2021-10-17, 21:26   #42
kriesel

"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

2·3·13·83 Posts

Quote:
 Originally Posted by kriesel 4) The date I saw and took note of on the previous version of the pdf on my server before generating the 2021-10-13 update pdf for updating was 2021-08-26. ... 7) Another update would have been triggered by https://www.mersenne.org/report_milestones/ "2021-10-07 All exponents below 105 million tested at least once."
#4 should have read 2021-10-07, not 2021-10-13.

I did not find a cached copy of GIMPS minor milestone progress versus year https://www.mersenneforum.org/showpo...5&postcount=11 in my web browser cache, downloads folder, on my server drive, etc. But I did find a copy of "gimps progress and rate.pdf" from the 2021-08-26 update of that page. It has been copied and renamed to preserve it on my server drive. The milestone progress post will be updated shortly.

The actual posting date is shown by multiple methods to have been 2021-08-26. Greenskull's false claim of 2021-10-13 is refuted.
Attached Thumbnails

Attached Files
 gimps progress and rate-2021-08-26.pdf (58.9 KB, 47 views)

Last fiddled with by kriesel on 2021-10-17 at 21:45 Reason: remove firefox's extra line feeds, correct second attachment, removed and readded 3rd & 4th to preserve order

2021-10-17, 21:28   #43
kriesel

"TF79LL86GIMPS96gpu17"
Mar 2017
US midwest

194A16 Posts
rules addition; updated verification forecasts pdf

Effective as of the time and date of this post, these rules are appended to those posted at https://mersenneforum.org/showpost.p...&postcount=36:

12) Any of the following shall constitute grounds for disqualification of a participant:
1. Making false claims about matters relating to the estimations and failing to promptly correct them after becoming aware of the inaccuracy.
2. Engaging in excessively argumentative, specious, or illogical posting or too-frequent posting regarding estimates.
3. Repeated violation of the stated rules.
4. Dishonesty.
5. Trolling. (See https://en.wikipedia.org/wiki/Internet_troll and https://mersenneforum.org/showpost.p...&postcount=314)
6. Other sociopathic behavior.
13) Disqualification and grounds for it shall be determined in the sole judgment of kriesel. (See the Predict M52 game thread https://mersenneforum.org/showthread.php?t=23892 for precedent that the creator of the game is the final arbiter of rules interpretation and enforcement.)

14) There is no mechanism for appeal of disqualification. Disqualification is final. Disqualification is permanent.

15) Estimates made, without specification for each estimate, of all 3 of the following elements,
1. which Mersenne prime, eg Mp#49*/Mp#50*/Mp#51* (or eventually Mp#52*),
2. estimated date of verification completion, and
3. corresponding stated credible basis,
are invalid and shall be ignored. Estimates made before the time and date of this post are grandfathered in, and may be supplemented by stating the basis for them, within 30 days of original posting of estimate elements. Stating a clear reproducible basis or algorithm is encouraged.
(Examples of credible basis are fits to past milestones, projections based on number of verification completions per day, projections based upon total GHzD required and primenet statistics, other fully documented math based approaches; examples of not credible basis are crystal balls, ouija boards, magic 8balls, dreams, premonitions, psychics)

16) Purpose is to identify a variety of alternate estimation methods, and demonstrate by test over extended periods, which are more effective or less effective.

17) Participation after the time and date of this post constitutes consent to the rules.

(Anyone may of course start their own game, contest, whatever, spending the effort to: develop and post rules; tabulate estimates; upon completion of a verification, compute standings; etc. A different forum thread for such an undertaking is highly recommended.)

Quote:
 Originally Posted by Uncwilly There is not so much fun there.
Amen to that.
Attached Files
 Mpnn verification forecasts.pdf (20.1 KB, 52 views)

 2021-10-17, 21:51 #44 greenskull Xebeche     Apr 2019 🌺🙏🌺 5×89 Posts I don't want to participate in Kriesel's game. Due to boring uninteresting rules and sloppy data handling. But I will stay here and will, from time to time, refine my predictions. If necessary, of course. Also, if someone wants to clarify theirs, then I will take this into account in my quality assessment system. I will apply the assessment I presented earlier here: https://www.mersenneforum.org/showth...848#post585848 A little later, I will summarize all the forecasts here in a table and publish. Last fiddled with by greenskull on 2021-10-17 at 21:55

 Similar Threads Thread Thread Starter Forum Replies Last Post kriesel Marin's Mersenne-aries 136 2021-11-02 04:40 Happy5214 Computer Science & Computational Number Theory 5 2021-03-31 12:36 kriesel News 0 2019-02-24 15:47 CuriousKit PrimeNet 13 2015-06-21 14:49 jasong Software 13 2010-03-10 18:53

All times are UTC. The time now is 07:03.

Sat May 21 07:03:37 UTC 2022 up 37 days, 5:04, 0 users, load averages: 0.95, 1.29, 1.27