Noodling Stats - Australian 5c, 10c And 20c Coins

Thanks for sharing your stats Hurcules.
One thing about stats like this - you need an awful lot before you can say that the numbers are statistically significant.
For example:
1. I have found a 1972 5c in $10 worth of 5c.
Or
2. I have gone a much longer time than you would expect between finds of 1972 5c in the past. I think I searched through more than $350 without finding any.
Even with heaps of data, the distribution of coins is not always even around the country, so stats for 2000 $1/10c mules or 2000 incuse flag 50c in Melbourne or Perth may be different to Sydney.
Stats are great (and I like doing this sort of analysis), just don't lose sight of other factors too.

I came to the exact same conclusion yesterday evening. I was going to keep all the 20c coins that I noodled with the Arnold Machin portrait on them, but then when I realised that I could barely carry them at once, I decided to take them all back to the bank.

Based on official mintage figures, the 1971 20c ought to be just as rare as the 1972 5c, but in reality, it's much more common. It really makes me wonder just how accurate the mint's official figures are.




I'm fully aware of this problem, which is why I'm trying to create as big a sample size as possible. I'm hoping that it will eventually reach the point where it becomes statistically significant. If not, then it will still be useful as a guide.

How big a sample size do you think I'll need in order for the figures to become statistically significant?




I've had similar experiences in the past. Some of the coins I've found were so rare that my eyes almost popped out when I saw them, yet others that were relatively common have taken me decades to find. There really is a lot of luck in this, and people will notice this once I start talking about errors and varieties.




I read your thread several months ago, and my stats support the higher mintage figure, but I'm afraid it's not the only error. Several years ago, the mint published the wrong mintage for the 2010 circulating standard 20c coin on their website, and popular coin guides like Maccas and Renniks copied it. Maccas has only just realised the mistake and corrected it, whereas Renniks is still printing the incorrect figure. I can think of several other examples, but this is the most serious.

I've often wondered about this, but because I'm getting my coins in rolls instead of bags, I think the risk is minimal. Do any of the major banks possess the ability to roll their own coins? Does having the name of a security company printed on the packaging ensure that the coins came from a centralised location, and thereby got mixed up with many others?

I thought it was interesting how the first few batches of 10c coins I noodled had one 1991 and 1997 on average, but then when I returned those coins to the bank, the well of 1991s and 1997s suddenly dried up and I started getting a 2011 in each batch. Now, the well of 2011s has also dried up, and I've started getting 1985s. Is this normal behaviour for a random sample?