Firstly it is strange that no one has bothered to provide or ask for evidence that the notes are indeed printed in a straight run from beginning to end.
That's because it is the null hypothesis, meaning it is the theory we accept when no other theory can be proven. That's how the scientific method works. If we choose to abandon the null hypothesis without having another theory, then we have absolutely nothing, meaning there's no proof that anyone is printing notes at all.
You want proof of sequential printing of serial numbers? OK. The one thing I remember off the top of my head is the Journey $50 note with the offset serial numbers. We had this discussion and you were part of it...
http://www.cdnpapermoney.com/forum/index.php?topic=13552.0Ink transfer from the younger higher-numbered note gets onto the older lower-numbered note. It only makes sense if they are printing successive reams higher instead of lower or in some other jumbled order.
I have seen no evidence of that. Perhaps I have missed something but am happy to be enlightened. I am sure most people will consider me crazy to even raise this as a possibility but any good research considers that possibility.
You're not crazy for proposing an alternate hypothesis. If you dismiss the null hypothesis outright before proving the alternate hypothesis, then you would be acting crazy. Having enthusiasm for research and new ideas is admirable, but the researcher needs to keep things in perspective.
So let me offer you the possibility that the existence of the "star" in the hockey stick on $5 notes is a result of something going on with the plates. It is not a random printing mishap or variation that happens.
Interesting theory, but I have looked at more than a few $5 notes and came away with the conclusion that they are purely the result of gradual changes in the printing process.
Why do I say that - Through my research I can now predict the specific runs within each series (almost all - still waiting to resolve some prefixes) that are likely to have stars or not have stars. I can now predict this with 85% probability. In any research based on samples that would be considered a good run rate.
85% is a good score in high school English class. In a statistical analysis, a confidence interval of 85% is not that great. Certainly, if you were doing political polls for Ipsos Reid or another major company, 85% confidence means that even if one party had a 14% lead over another, they would still be statistically tied, meaning there's still a significant possibility that your poll is inaccurate.
I'm glad the data in the SNDB is more than 85% accurate.
I would like to offer to you the possibility that HPA, HPV and HPZ were printed before any of the other prefixes. Until that time there were no issues with the plate and there were no stars in either of those prefixes or if there are stars they are limited to specific runs predicted by the print run data that is available on the wiki. I would like to further suggest to you that HPD and HPH were perhaps printed last. They are the ones with most frequency of stars.
Interesting ideas. There is certainly no need for the printers to use prefixes in alphabetical order, especially when they know ahead of time how many notes they need to produce. In other words, if the bank of Canada orders 100 million notes, that's 10 complete prefixes, so they could pick a run of 10 prefixes and print those prefixes in any order.
However, even with a strong demand for circulating tender that makes it feel like production will be endless, production is not a constant process. Machines start, machines stop. How often, we don't know. Even machines that run perfectly need to be stopped for routine maintenance now and then. Since I have experience as a laboratory technician, I know that machines and scientific instruments work best if they work on a consistent ongoing basis. In a strange way, it kind of proves Newton's Third Law of motion. It's easier to keep a machine working than it is to get one working after a long period of inactivity.
Before I get carried off on a tangent, here is something you need to consider about the hockey stick phenomenon. The designs of the $5 Journey notes are printed BEFORE the serial numbers. The serial numbers are added to notes in a separate printing process. Well, that's the null hypothesis for printing stage chronology. Maybe you would like to propose another theory, that the hockey sticks and serial numbers are printed at the same time. You are assuming that over the course of printing some 20 prefixes (200 million notes), that the change in hockey stick appearance is a linear process from beginning to end of production. Here are the problems with that theory:
- It assumes nearly constant production, or, the ability of the printers to keep exactly the same conditions throughout production that is non-constant.
- It assumes that the roughly 4.5 million sheets of paper used to make those 200 million notes were kept in exactly the same order through the two different stages of printing (first, the back design, then, the serial numbers).
We have seen evidence that this incredible order and synchronization does not happen. What evidence? Consider the double denomination errors that occurred on $10 notes of prefixes BTT and BTU. These error notes occur in small clusters scattered throughout a range of about 10 million notes, with no predictable pattern to the occurrences. If the printers were so fastidious as to keep everything in perfect order, a block of $20 security paper getting mixed into the $10 security paper would have remained one big block after the serial numbers were added. But we do not see this. What we do see appears to be quite chaotic.
I have no doubt that there is some pattern to how the stars on the hockey stick emerge. What we don't know is how often it emerges over the course of printing 4.5 million sheets. Trying to get the answer from correlating the phenomenon to serial numbers makes no sense to me. Again, I emphasize that serial numbers were added later. There are so many variables that are unknown between the two stages of printing that I'm surprised you could even get a predictability rate of 85%. That probably tells us there is very good synchronization going on at the printing presses. So the fact that you have engaged in this research is quite worthwhile. However, it is still critically flawed to suggest a completely revised chronology of prefixes from a printing process that has nothing to do with how serial numbers are applied to notes.
Finally, don't take feedback on your research too personally. I know emotion can fuel the desire to prove one thing over another, but it may not help one in getting closer to the truth, especially if the truth has been known all along.
