My friends birthday is Nov. 26, and it is Thanksgiving this year. He wants to know how often the 26th has been on thanksgiving from his birth in 1960 til now. I tried to look, but it won’t show anything earlier than 2011.
There’s a table on this page where you can count them up. The answer is 9 if I counted correctly.
Counting them is the way to know for sure. But there are only 14 possible calendars and 2 of them have Thanksgiving on 11/26, so you can ballpark it at 8.5 over 60 years, which is pretty close to 9.
This answer is correct, but your reasoning is wrong (or at least not explained fully. There are 14 different calendars, but 7 of them are 3 times as likely as the other 7. So yes 1/7 of the time is correct.
Yes, you are correct, but my reasoning was not wrong, just not explained fully. One set of 7 is for leap years and the other for non-leap years. But each set of 7 has exactly one where Thanksgiving falls on 11/26, so the probability of a randomly selected year having Thanksgiving on 11/26 is 1/7. This assumes that the calendars of each set are equally likely, which I have not confirmed.
I did a little figuring on this.
In a leap year, Thanksgiving is on 11/26 if New Year’s day is on Wednesday. It’s Thursday for non-leap years.
In the last 2000 years, the fourth Thursday of November fell on the 26th 70 times in leap years (14.37% of leap years), and 220 times in non-leap years (14.54% of non-leap years). In the aggregate that’s 290 times, or 14.5%. That is a little more than 1/7 (14.29%).
The Gregorian Calendar repeats exactly every 400 years. You need 400 years because of the 400 year rule on Leap Years. Each regular year has one day more than 52 weeks. Each Leap Year has 2 extra days. There are 97 Leap Years in 400 years so there are 303 + 2*97 = 497 “extra days” or exactly 71 “extra” weeks in 400 years.
As the calendar repeats exactly in 400 years but 400 is not divisible by 7 some dates are slightly more or less common for holidays that are fixed on a given day of the week.
Of course Thanksgiving wasn’t always on the fourth Thursday in November as it is now. Until 1938 it was traditionally the last Thursday in November as set by Lincoln in 1863. It was set by Presidential proclamation, and that rule might not always have been followed before then. In 1939 Roosevelt declared Thanks giving to be Nov 23 (the fourth and not the last Thursday). In 1940 Roosevelt set in as Nov 21, the third Thursday.
In October 1941 the House passed a bill making Thanksgiving the last Thursday. The Senate amended the bill to make it the fourth Thursday. That’s the bill that was signed into law by Roosevelt.
Perpetual Calendar - https://www.beercoast.com/images/calendars.jpg
Your friend’s is asking about dates in the past 60 years, so the changing date of thanksgiving or some of the weirder quirks of the Julian and Gregorian calendars don’t really matter.
If you check the calendar, 11/26 is a Thursday in years that match year 5 & year 11. Count the number of 5s and 11s since 1960.
I will not dispute your calculations, but I do think that your conclusion is phrased in a misleading way. By concluding “that is a little more than 1/7,” you give the impression that there is some sort of imbalance, that the calendar is predisposed towards having Thanksgiving on the 26th slightly more than on other days. This is not correct.
“A little more than 1/7” is NOT because the calendar is playing favorites with the 26th. It is because you chose to look at a group of 2000 years, and 2000 is not evenly divisible by the cycle that the calendar uses. The calendar goes through a cycle of 28 years, and if you choose any group of years that is divisible by 28, you’ll find that Nov 26 is on Thursday exactly 1/7 of the time.
Today is Oct 29 2020, and it is a Thursday.
52 weeks and a day later, Oct 29 2021 will be on a Friday.
52 weeks and a day later. Oct 29 2022 will be Saturday.
52 weeks and a day later, Oct 29 2023 will be on Sunday.
52 weeks and two days later, Oct 29 2024 will be on Tuesday.
The existence of leap day (Feb 29 2024) is why you can’t depend on this cycle repeating every seven years. But if you allow for four consecutive seven-year groups, you’ll find that the cycle does repeat every 28 years.
For example, here is Nov 26, for the 28 years from 2010 to 2037:
Friday, November 26, 2010
Saturday, November 26, 2011
Monday, November 26, 2012
Tuesday, November 26, 2013
Wednesday, November 26, 2014
Thursday, November 26, 2015
Saturday, November 26, 2016
Sunday, November 26, 2017
Monday, November 26, 2018
Tuesday, November 26, 2019
Thursday, November 26, 2020
Friday, November 26, 2021
Saturday, November 26, 2022
Sunday, November 26, 2023
Tuesday, November 26, 2024
Wednesday, November 26, 2025
Thursday, November 26, 2026
Friday, November 26, 2027
Sunday, November 26, 2028
Monday, November 26, 2029
Tuesday, November 26, 2030
Wednesday, November 26, 2031
Friday, November 26, 2032
Saturday, November 26, 2033
Sunday, November 26, 2034
Monday, November 26, 2035
Wednesday, November 26, 2036
Thursday, November 26, 2037
In those 28 years, Nov 26 falls on each day of the week exactly four times. The same will be true for any days of the year, and any set of 28 consecutive years.
(One might think that Feb 29 is an exception to the above rule, but it is not; in any 28 years, Feb 29 will be on each day of the week exactly once, and it will not exist in the other 21 years. A true exception to this rule, however, would be those century years that aren’t leap years. If the 28 years include 1900 or 2100, for example, the “7*4=28” cycle doesn’t happen.)
Confirmed. The seven leap years are equally likely, and the seven non-leap years are equally likely. Any set of 28 (or multiple of 28) consecutive years (that does not contain a non-leap century year) will contain exactly one of each leap year, and exactly three of each non-leap year.
Your second sentence is true. Your first sentence doesn’t follow from the first and is incorrect. Let me rephrase from before:
The calendar repeats exactly in 400 years but 400 is not divisible by 14 let alone 7. That means that some of the 14 types of years are slightly more common than others. For example in any 400 year period (under the Gregorian calendar), January 1st falls on the different days of the week
Su Mo Tu We Th Fr Sa
58 56 58 57 57 58 56
This is not even maximally “flat” with a Monday Jan 1st occurring two times fewer than some other days.
OldGuy does seem to be correct that the calendar repeats every 400 years. The calendars for 2020 and 2420 are identical, for example.
But I can’t figure out why this is so. My brain has been convinced of the 28-year cycle for a very long time. I will need to research this further at home tonight. It is possible that I am wrong, but I suspect that the 400-year cycle has some hard-to-find details that make it similar to a 2800-year cycle, or something like that.
All that really matters is what day of the week is the 26th of November. Who cares about the rest of the calendar. The fact that 7 is relatively prime to whatever cycle of calendar you choose to look at, (4 years or 400 years) means that the all days will cycle over time so each day has equal chance of being a given day of the week.
The 28 year cycle is gives all of the combinations of position in the 4 year leap year cycle and all days of the week for a particular day (say Jan 1 or Nov 26). This is because 28 is the least common multiple of 7 and 4. This ignores the fact that leap years are skipped every century but not on multiples of 400 (hence the 400 year cycle) If you want to include all possible combinations of starting day of the week and position in the 400 year cycle you end up with a 2800 year cycle.
Every year that is exactly divisible by four is a leap year, except for years that are exactly divisible by 100, but these centurial years are leap years if they are exactly divisible by 400. For example, the years 1700, 1800, and 1900 are not leap years, but the years 1600 and 2000 are.
This works out to a 400 year cycle, due to the centurial leap year rules.
I think I got it. Consider:
A non-leap year is 52 weeks and 1 day. Because of that extra day, any given calendar day this year will usually have advanced in the week by one day when next year comes around.
A leap year is 52 weeks and 2 days. Therefore, if Feb 29 intervenes, next year’s date will have advanced in the week by two days.
ANY given set of 400 years will have 300 normal non-leap years, and also 3 years which are non-leap because they are divisible by 100 but not by 400. This leaves exactly 97 leap years in any set of 400 consecutive years.
During the 303 non-leap years, the days will advance once per year. During the 97 leap years, a date will advance by two days. In the course of 400 years, a date will have advanced by a total of 303+(2*97) days, which is 497 days.
497 days is exactly 71 weeks, putting any given day on the same day of the week as 400 years earlier.
I do not think this will work for any period shorter than 400 years, because the number of leap years will be variable. Depending on the exact start and end points, the number of years divisible by 100 can vary, and the number of years divisible by 4 can certainly vary.
Perhaps an easier way to think about it is that a block of 4 years with 1 leap year is 1461 days. 400 years of such blocks would be 146100 days, which does not divide by 7. But in a 400 year (inclusive) span with the Gregorian leap year rule, you have two extra leap days, and 146102 does divide by 7.
Wow. Thank you all.
No sorry. First 146102 is not evenly divisible by 7. Second in 400 years you have 3 fewer leap years so 146100-3 = 146097 which is 7*20871. That’s why the calendar repeats every 400 years and why dates are not exactly even on days of the weeks in the long run.
The calendar repeats exactly every 400 years but not over any shorter period. Somewhere I read that the leap year should omitted every 4000 years, but does anyone expect we will still be using this calendar in the year 4000? Looking at the actual length of the day it appears to me that that skip should happen every 3200 years. Looking a little deeper it seems that the leap year ought to be omitted every 128 years instead of approx. every 133. This leads to an amusing fact. If our arithmetic was hexadecimal, that number dec. 128 would be hex. 80 and means we could omit a leap year at a round number of years. Psychologically, hex 80 is like dec 50 since 8 is half the base.