4 ms·
A basketball game can’t last longer than the lifetime of the people involved. Therefore the number of points is finite. Therefore there is a max value beyond
by syki 5y ago
A basketball game can’t last longer than the lifetime of the people involved. Therefore the number of points is finite. Therefore there is a max value beyond which no game can score beyond. If you wish to enlarge this max value take a fixed multiple of the number of particles in the observable universe as a max value.
In a similar way the amount of money that can be created in the next X years without completely devaluing the currency is finite. Once people lose all trust in the currency it is meaningless how much one has. Therefore the amount of money (buying power) one can have is limited.
- neilwilson 5y agoAnd that's the other key point. Although there is an infinite amount of basketball points available, you can only award them when games are played. And then only at a value that people have agreed is reasonable according to the agreed rules of the games decided by the game's authorities. In monetary terms, the spending stops automatically when there is no longer anything to buy at a price worth paying. The argument is never "we can't afford it", the argument is always "it's too expensive".
- syki 5y agoI think you don’t understand. There isn’t an infinite amount of points available in basketball in a similar there isn’t an infinite amount of money available. Each particle in the observable universe can contain only a finite amount of information and therefore there can’t be an infinite amount of money or basketball points because each requires “information”. It’s all constrained by the finite number of particles in the observable universe.
- dredmorbius 5y agoAgain: scoring events != awardable points. There is no bound on points. https://news.ycombinator.com/item?id=29663864 https://news.ycombinator.com/item?id=29663864
- syki 5y agoYou can’t store a value that requires more storage than is available in the observable universe.
- dredmorbius 5y agoRed herring in this context.
- dredmorbius 5y agoThe number of potentially awardable points is infinite. They literally come from the set of positive integers. The potential awardable events --- baskets sunk in a game, multiplied by their value (1, 2, or 3 points per basket) --- is finite and bounded. But should there ever be a means to score more baskets, the points will be there. You seem to be confusing awardable events (baskets) with points (numbers from the set of all positive integers). It's the latter which are infinite. Put another way, even if the total number of baskets scored in all games for all time has some finite maximum, by changing the awarded points per basket the total points in a game could increase. Rather than 1, 2, and 3 points we might award 2, 4, and 6. Or 10, 20, and 30. Or 1*10^100, 2*10^100, and 3*10^100. There are always more points available to match to baskets and their value. There is no "point store" that can be depleted. (I'm not arguing that awarding a googol points per free-throw is desirable or would improve the game. I'm saying it's possible to set the scoring to do this. Pinball games seem to follow a similar scoring logic, for comparison.)
- syki 5y agoThere are a finite number of particles in the observable universe and each particle can store a finite amount of information. Therefore anything that requires information or anything that requires matter or energy will have a finite amount of availability. The point store as you put it is the observable universe.
- dredmorbius 5y agoEstimates are that there are 10^78 -- 10^82 atoms in the observable universe. 10^100 would be a number greater than either of these values. That's the named large number googol. A googolplex is 10^10^100 ... that is, ten raised to the power of ten raised to the power of a googol. There are notations for describing yet larger numbers. One of these is Knuth's up-arrow notation, in which an upward-pointing arrow (↑) indicates iterated exponentiation.[1] 2↑4 = 2 * (2 * (2 * 2)) = 2^4 = 2^16 2↑↑4 = 2↑(2↑(2↑2) = 2^16 = 65,536 2↑↑↑4 is harder to describe, especially with the typographic limitations of HN. It is 2↑↑(2↑↑(2↑↑2)), which is 2^2, itself raised to the power of 2, and so forth, 2^65,536 times. (The Wikipedia article represents this more clearly, though omitting ... some detail...) The number is immensely large. It is immensely larger than the number of atoms, or subatomic particles (roughly 1 electron and 3 quarks per atom), or Planck lengths (~10^53), or Planck volumes (~10^53^3), in the observable universe. And yet it can be represented. And it's hardly the largest representable number.[2] The up-arrow notation itself can be more concisely represented as ↑^n, where n is the number of arrows: ↑^2 = ↑↑, ↑^3 = ↑↑↑, etc. And all of these are simply representations of integer numbers. There are infinitely many integers, and given a need, a sufficiently concise set of expressions can be constructed to describe these. Infinity is like, really big, man.[3] There is no bound on points. There is no bound on monetary units which can be created. There is a bound on what those can represent, but that is a separate argument. Money != value. Money is a unit of measurement of value. ________________________________ Notes: 1. See generally: https://en.wikipedia.org/wiki/Knuth%27s_up-arrow_notation https://en.wikipedia.org/wiki/Knuth%27s_up-arrow_notation 2. See generally: https://en.wikipedia.org/wiki/Large_numbers https://en.wikipedia.org/wiki/Large_numbers 3. I won't even get to different sizes of infinities, though that's a fun exercise. See for example Cantor's Diagonal Argument: https://en.wikipedia.org/wiki/Cantor%27s_diagonal_argument https://en.wikipedia.org/wiki/Cantor%27s_diagonal_argument