Victorian YH shilling die axis variation

[Nick]

My examination of the die axis variation for Victorian young-head shillings is complete. I utilised a square piece of thick card with a shilling-sized hole that allows the reverse to be aligned with a mark on the card and then flipped over and the obverse photographed, whilst maintaining the reverse in the correct alignment. I arbitrarily chose as a zero point an obverse that looked to have the portrait in the most normally upright position, then measured each of the angular differences from this zero point.

The results from 65 YH shillings show a spread of 33 degrees, from 27 degrees clockwise to 6 degrees anti-clockwise. The results are normally distributed about the mean (8 degrees clockwise).

Apologies for posting a topic that may only be of interest to Rob and maybe he was just being polite.

[Rob]

Interesting, thanks Nick. A normal distribution curve suggests that a large deflection is not an unusual event which happens suddenly, rather a reading at either extreme as part of the normal spread. For the mean to be 8 degrees off to one side, this position can reasonably be assumed as the intended one for a 'perfect' strike working on the basis that most of the time, things would be done correctly. With a spread of 19 on one side and 14 the other, that isn't wildly skewed, and the sample size of 65 is large enough to be considered statistically significant. If there had been a clear lump in the distribtion at one extremity (which I was hoping for) then you could say this was a significant event, but a normal distribution rules that out. I don't think we can infer anything from this survey other than some engravers needed an eye test. Back to the drawing board.

Edited April 29, 2013 by Rob

[Nick]

Interesting, thanks Nick. A normal distribution curve suggests that a large deflection is not an unusual event which happens suddenly, rather a reading at either extreme as part of the normal spread. For the mean to be 8 degrees off to one side, this position can reasonably be assumed as the intended one for a 'perfect' strike working on the basis that most of the time, things would be done correctly. With a spread of 19 on one side and 14 the other, that isn't wildly skewed, and the sample size of 65 is large enough to be considered statistically significant. If there had been a clear lump in the distribtion at one extremity (which I was hoping for) then you could say this was a significant event, but a normal distribution rules that out. I don't think we can infer anything from this survey other than some engravers needed an eye test. Back to the drawing board.

Thanks Rob. Here is a histogram of the results.

[Coinery]

Interesting, thanks Nick. A normal distribution curve suggests that a large deflection is not an unusual event which happens suddenly, rather a reading at either extreme as part of the normal spread. For the mean to be 8 degrees off to one side, this position can reasonably be assumed as the intended one for a 'perfect' strike working on the basis that most of the time, things would be done correctly. With a spread of 19 on one side and 14 the other, that isn't wildly skewed, and the sample size of 65 is large enough to be considered statistically significant. If there had been a clear lump in the distribtion at one extremity (which I was hoping for) then you could say this was a significant event, but a normal distribution rules that out. I don't think we can infer anything from this survey other than some engravers needed an eye test. Back to the drawing board.

Thanks Rob. Here is a histogram of the results.

Neat work, Nick!

[Peter]

Good work chaps.Brilliant Nick

But don't send me down the path of Vickie YH shillings.