What proportion of months contain parts of six calendar weeks?

What proportion of months have that annoying property that, on an old-fashioned paper calendar, the 23rd and 30th, or 24th and 31st, have to be scrunched up into a single box? Or, on a computerized calendar, six rows are necessary? Or, if we don’t want to refer to a particular calendar format, that it contains parts of six (Sunday-to-Saturday) calendar weeks?

For example, consider a 30-day month that starts on a Saturday, like September 2012, which is the next example of this phenomenon:

Sun Mon Tue Wed Thu Fri Sat
2 3 4 5 6 7 8
9 10 11 12 13 14 15
16 17 18 19 20 21 22
23 24 25 26 27 28 29

31-day months that start on Friday or Saturday also have this property. To identify this month we can quickly code up Zeller’s congruence for the day of the week:

zeller = function(m, d, y){
if(m K = y %% 100;
J = y - 100*K;
h = (d + floor(13*(m+1)/5) + K + floor(K/4) + floor(J/4) - 2*J) %% 7;

This returns 0 for Saturday, 1 for Sunday, …, 6 for Friday.

Then put the lengths of the months of the year in a vector:

lengths = c(31,28,31,30,31,30,31,31,30,31,30,31);

(You may object “what about leap year!” — but that doesn’t concern us, as February, even in leap year, can never require six rows.)

The sixweeks function returns TRUE if a month contains parts of six (Sunday to Saturday) calendar weeks, and FALSE otherwise:

sixweeks = function(days, first){
((days == 30) && (first == 0)) || ((days == 31) && (first == 0)) || ((days == 31) && (first == 6))}

Now the Gregorian calendar has a period of 400 years. So we just run over some 400-year period and run sixweeks on every week. The result is a vector containing the number of each month which fall within parts of six calendar weeks, in that 400-year cycle.

counts = rep(0, 12);

for(y in 2000:2399){
for(m in 1:12){
first = zeller(m, 1, y);
days = lengths[m];
counts[m] = counts[m] + sixweeks(days, first);

The output is (cleaned up a bit, and with the month names inserted):

Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
116 0 112 60 116 56 116 112 56 116 56 116

These numbers add up to 1032 (run sum(counts)). So in each 400-year period, 1032 months out of 4800, or exactly 21.5%, include parts of six calendar weeks. 116 of these months are January, 0 are February, 112 are March, and so on. (If you believe that a calendar week is Monday to Sunday, because you take the dictates of the ISO too seriously or because you’re European, it’s not hard to adapt the sixweeks function to that; instead of 1032 you get 1028.)

Could we have predicted this number without the need for computation? We can come pretty close. Seven out of every 12 months is a 31-day month; of those about two-sevenths should start on a Friday or a Saturday. Similarly, four out of every 12 months is a 30-day months, and one-seventh of those should start on a Saturday. So the probability that a randomly chosen month contains parts of six calendar weeks ought to be quite close to
{7 \over 12} \times {2 \over 7} + {4 \over 12} \times {1 \over 7} = {18 \over 84} \approx 0.214
and indeed we come pretty close!  In fact this post is historically backwards. I did this calculation first and then went to the computer and wrote the code to check it.


One thought on “What proportion of months contain parts of six calendar weeks?

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s