Berkshire is overflowing a 32-bit unsigned integer representation. When an unsigned integer exceeds its maximum representable value, it wraps around starting at zero.
If you add 1 to the maximum, you get 0. If you add 2 to the maximum, you get 1. If you add 3 to the maximum, you get 2. If you add 4 to the maximum, you get 3. And so on. The number wraps around to zero (it's modulo 2^32).
They're using a 4-byte (i.e., 32-bit) unsigned integer to represent the price in terms of ten-thousandths of a dollar.
For example, the standard size of an "unsigned int" in C on most modern systems is 32 bits. The maximum value you can express in a 32-bit unsigned integer is 32 one-bits:
1111 1111 1111 1111 1111 1111 1111 1111
This binary number is two to the power of 32, minus 1 (i.e., 2^32 - 1) which is, in decimal:
4,294,967,295
This integer represents $429,496.7295 because the price is specified in ten-thousandths of a dollar.
The Berkshire A-share (BRK.A) exceeded that price.
The B-share (BRK.B) equivalent price is $286.33 (divide A-share price by 1,500).
There was no sense of urgency because the closest to reaching the limit aside from Berkshire is several factors away. Also, Berkshire stock has surged ~30% in the past ~6 months which is far above usual performance.
Those aspects together probably allowed it to be deprioritized until it would be closer to being necessary (a classic tale).
It is rather funny to imagine Warren Buffett sitting there mumbling about stock splits and the incompetence of it all, and how in his day, you could buy a share for however much cash you wanted to without worrying about integers and bits and code.
Yeah, normally companies carry out a stock split way before the price reaches this level because having such a high price makes trading and holding their shares a bit of a pain. Berkshire Hathaway is just unusual in that Warren Buffett considers this an advantage. Also, I guess it's not the kind of company that usually has wild fluctuations in its share price - Buffett usually seems to go for boring businesses with steady returns.
5 comments
[ 3.1 ms ] story [ 25.1 ms ] threadhttps://old.reddit.com/r/brkb/comments/n6bjws/integer_overfl...
Berkshire is overflowing a 32-bit unsigned integer representation. When an unsigned integer exceeds its maximum representable value, it wraps around starting at zero.
If you add 1 to the maximum, you get 0. If you add 2 to the maximum, you get 1. If you add 3 to the maximum, you get 2. If you add 4 to the maximum, you get 3. And so on. The number wraps around to zero (it's modulo 2^32).
They're using a 4-byte (i.e., 32-bit) unsigned integer to represent the price in terms of ten-thousandths of a dollar.
For example, the standard size of an "unsigned int" in C on most modern systems is 32 bits. The maximum value you can express in a 32-bit unsigned integer is 32 one-bits:
1111 1111 1111 1111 1111 1111 1111 1111
This binary number is two to the power of 32, minus 1 (i.e., 2^32 - 1) which is, in decimal:
4,294,967,295
This integer represents $429,496.7295 because the price is specified in ten-thousandths of a dollar.
The Berkshire A-share (BRK.A) exceeded that price.
The B-share (BRK.B) equivalent price is $286.33 (divide A-share price by 1,500).
Those aspects together probably allowed it to be deprioritized until it would be closer to being necessary (a classic tale).
It is rather funny to imagine Warren Buffett sitting there mumbling about stock splits and the incompetence of it all, and how in his day, you could buy a share for however much cash you wanted to without worrying about integers and bits and code.