Might be interesting to predict the probability of multiple occurrences of wining numbers tho I think I’ve only ever heard of the exact same numbers being selected on two occasions. I also wonder if you go by worldwide numbers what the probability might be of winning numbers being duplicated (I think many of them use machines that pick the numbers publicly made by the same companies too).
I bet there are people obsessed with numbers already on it
I’m not sure about over there, but (IIRC) they use machines here with physical balls here, and perhaps variations in things like weight or smoothness/texture might play a part even if slight. Either way I would much prefer to let a computer try to guess the numbers than simply picking one of their ‘lucky dips’ - even if just out of curiosity and a little fun exercise
I might expand that lottery app I made one day, perhaps could be a good chance to play with Elixir’s Nx
It would appear that they do still use the balls in a tumbler. You can watch the draws online, here is a link for the draw in question
In terms of the physical properties of the balls, it would only really make a difference if the same set was used each time. If they use a new set each time (or change them with reasonable frequency) then you would not be able to find patterns caused by uneven wear or the likes. I guess you might find something with manufacturing quality or perhaps the weight differences by variable amounts of printing on the surface. I’d imagine even if you were able to glean some kind of advantage from this kind of analysis it would be extremely slight and quite an expensive endeavour to try get anything out of it.
If the universe is a mathematical object, then perhaps it’s only a matter of time? In fact I think we are so primitive in what we know, that (if as a species we survive) in just one thousand years people will look back at us now in a similar way as we look at early humans of 200,000 years ago or even our closest relatives of a couple million - the pace of advancement is going to speed up exponentially imo.
That is starting to head into some deeply philosophical areas in terms of the point though, it is a reasonable one however I do think it would take more information than we are talking about here in order to model anything meaningful. There are thousands of variables at play which you won’t be able to determine by what is a fairly small set of previous numbers. Things like the temperature and humidity are changing the way those balls behave and changing how those imperfections affect their behaviour. Other things like the power source not be absolutely perfectly clean are having similar impacts on how they are being shuffled. If you can capture and model all the relevant information then it probably would be conceptually possible to figure it out, maybe one day… the main thing I’m getting at is I don’t think looking for patterns in previous results is going to get such a model, it is more like looking for patterns in previous states of the universe