WEBVTT 00:00.120 --> 00:06.420 Hello again! In this video, we are going to start looking at random numbers. A random number is one 00:06.420 --> 00:08.670 which appears as though it has been chosen by chance. 00:09.480 --> 00:14.850 Usually we have a range of numbers and we assume that any number in this range is equally likely to 00:14.850 --> 00:15.420 be chosen. 00:16.140 --> 00:21.990 So for example, if you have a pack of cards - playing cards - and you choose one at random, it is equally 00:21.990 --> 00:24.570 likely to be any one of the 52 possibilities. 00:25.770 --> 00:28.890 And that is what mathematicians call a "uniform distribution". 00:30.270 --> 00:35.160 We also assume that the next number we choose is not affected by the last number we chose. 00:35.700 --> 00:37.980 So the choices are "independent". 00:38.670 --> 00:40.230 So that is like tossing a coin. 00:40.230 --> 00:45.210 If you toss a coin and get heads, then you still have a 50-50 chance of getting heads or tails 00:45.210 --> 00:45.810 the next time. 00:49.530 --> 00:52.020 Random numbers are actually very useful in computing. 00:52.410 --> 00:58.710 The obvious example is in games. For example, if you have a card game, then you can deal the player 00:58.710 --> 01:00.120 a different hand every time. 01:01.050 --> 01:05.580 If you have an action game, then you can make the opponent move around unpredictably, which makes 01:05.580 --> 01:06.420 him harder to catch. 01:07.080 --> 01:11.280 If you have a game with a story, then you can choose different branches through the story. 01:11.640 --> 01:15.780 So random numbers will make games more interesting and unpredictable. 01:17.640 --> 01:23.280 Random numbers are also used in numerical computation. For example, in "Monte Carlo" simulations. 01:23.790 --> 01:27.630 If you do not have complete data, you can put in random numbers with sensible values. 01:28.050 --> 01:32.850 Then run the simulation lots of times, and take an average. And that should give you some idea of what is 01:32.850 --> 01:33.420 going to happen. 01:34.140 --> 01:36.060 So that is basically how weather forecasting is done, 01:36.060 --> 01:43.470 for example. In cryptography, we can use random numbers to create codes and ciphers, which are unbreakable. 01:43.470 --> 01:45.240 Or very hard to break, anyway! 01:46.020 --> 01:50.100 And they are also used in authentication for network security and crypto. 01:53.190 --> 01:58.530 How do we get random numbers? If we are using software, we have a bit of a problem, because software is 01:58.530 --> 02:01.380 deterministic, so it cannot do things by chance. 02:02.460 --> 02:05.820 However, there are things called "pseudo-random number generators". 02:06.540 --> 02:12.360 These are algorithms which will do some calculations, and that will generate a sequence of numbers which 02:12.360 --> 02:14.460 look fairly close to being random. 02:16.140 --> 02:18.900 Obviously, it will produce the same results every time. 02:20.160 --> 02:25.410 So to mix things up, you can provide a number, which is known as a "seed". That is used to initialize 02:25.410 --> 02:29.250 the generator, and then that will cause it to generate different sequences. 02:30.870 --> 02:35.760 Obviously, that is not terribly helpful if you are testing your code. If you make a change and the program 02:35.760 --> 02:36.630 behaves differently, 02:36.930 --> 02:40.560 is it because of your change, or is it because you got a different random number? 02:41.220 --> 02:44.070 So when you are testing, it is a good idea to use the same seed every time. 02:45.060 --> 02:47.760 And then you can change it to use a different seed when you go into production. 02:50.150 --> 02:54.650 It is also possible to get true random numbers, but this involves interacting with the physical world. 02:55.970 --> 02:59.420 The oldest ones are things like rolling dice or tossing coins. 03:00.380 --> 03:06.470 In the 20th century, we discovered that physical processes produce random numbers: things like thermal 03:06.590 --> 03:12.260 noise, from electrons rattling around in conductors. Radioactive decay. 03:12.320 --> 03:13.790 How many particles per second? 03:16.480 --> 03:21.550 Moving into the modern world, computers have quite a lot of internal stuff, which is random. Things 03:21.550 --> 03:26.260 like process id, activity on the mouse or keyboard, the temperature of the CPU. 03:26.530 --> 03:30.100 And if you mix them all up, you get something which is pretty close to being random. 03:31.750 --> 03:36.310 And you can also get specialized hardware devices, which you can connect up to your computer. 03:36.970 --> 03:42.850 And these will use things like air pressure fluctuations to generate random numbers, in real time. 03:43.780 --> 03:45.430 Okay, so that is it for this video. 03:45.850 --> 03:48.700 I will see you next time, but until then, keep coding!