Quantile ranges

[Charles Lindsay]

In Bayliss & Marshall, 2022, Radiocarbon Dating and Chronological Modelling, reference is made to  "quantile ranges" (pp.9-10) as a simple summary of the overall distribution of probability of a calibrated date. As illustrated, the quantile range is somewhat different than the outer limits of the HPD range(s) for a calibrated date.

I could not find any reference to a quantile range in the manual for OxCal, nor as an option under the Edit button on the output of a calibrated date. Does Oxcal offer this calculation?

[Christopher Ramsey]

You can do this if you set the "Whole (floruits)" option on in [Options > Ranges] - in general I would not recommend this though as this range includes areas with lower probability and so is less informative.

Best wishes

Christopher

> On 2 Apr 2025, at 18:36, Charles Lindsay <clind...@gmail.com> wrote:
>
> In Bayliss & Marshall, 2022, Radiocarbon Dating and Chronological Modelling, reference is made to "quantile ranges" (pp.9-10) as a simple summary of the overall distribution of probability of a calibrated date. As illustrated, the quantile range is somewhat different than the outer limits of the HPD range(s) for a calibrated date.
> I could not find any reference to a quantile range in the manual for OxCal, nor as an option under the Edit button on the output of a calibrated date. Does Oxcal offer this calculation?
>

> --
> You received this message because you are subscribed to the Google Groups "OxCal" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to oxcal+un...@googlegroups.com.
> To view this discussion visit https://groups.google.com/d/msgid/oxcal/4585599b-e5ea-4ad1-a3cb-33757e86339fn%40googlegroups.com.

[Bayliss, Alex]

Hi Charles,

I don't agree with Chris. The advantage of the quantile range is that it always gives a single range. This can be convenient in cases where you have one or two 'range-finder' dates, which simply tell you a deposit is (e.g.) "Iron Age". The calibrated date using the probability method is more exact, but probably all a single sample is telling you is anyway "it's early 1st millennium BC". So why junk up your text with three split ranges that don't tell you more than this anyway? Calibration is anyway throw-away as next time someone looks at this measurement, the IntCal handle will have turned and it will be recalibrated (or, hopefully, contribute to a chronological model.

So, I think it has its place. But I don't use it very often.

Best wishes,

Alex

Ensuring our heritage lives on and is loved for longer.​​​​ historicengland.org.uk This e‑mail (and any attachments) is confidential and may contain personal views which are not the views of Historic England unless specifically stated. If you have received it in error, please delete it from your system and notify the sender immediately.  ​Do not use, copy or disclose the information in any way nor act in reliance on it. Any information sent to Historic England may become publicly available. For information about our use of your personal data please visit: historicengland.org.uk/terms/privacy

-----Original Message-----
From: 'Christopher Ramsey' via OxCal <ox...@googlegroups.com>
Sent: 02 April 2025 18:46
To: OxCal group <ox...@googlegroups.com>
Subject: Re: Quantile ranges

-- WARNING: This is an external message. Please use caution when replying, opening attachments or clicking on any links in this e-mail.--

To view this discussion visit https://groups.google.com/d/msgid/oxcal/8C5FFE5E-E29D-481A-9BC1-6DC896FFA865%40arch.ox.ac.uk.

[Christopher Ramsey]

Dear Alex and Charles

I did not mean that it was not useful or have its place - it is just not as useful as the HPD range in most circumstances which is why it is not the default option and not used that often.

Best wishes

Christopher

> On 2 Apr 2025, at 19:42, Bayliss, Alex <Alex.B...@HistoricEngland.org.uk> wrote:
>
> Hi Charles,
>
> I don't agree with Chris. The advantage of the quantile range is that it always gives a single range. This can be convenient in cases where you have one or two 'range-finder' dates, which simply tell you a deposit is (e.g.) "Iron Age". The calibrated date using the probability method is more exact, but probably all a single sample is telling you is anyway "it's early 1st millennium BC". So why junk up your text with three split ranges that don't tell you more than this anyway? Calibration is anyway throw-away as next time someone looks at this measurement, the IntCal handle will have turned and it will be recalibrated (or, hopefully, contribute to a chronological model.
>
> So, I think it has its place. But I don't use it very often.
>
> Best wishes,
>
> Alex
>
>
>

> <image297290.jpg>Ensuring our heritage lives on and is loved for longer.​​​​

> To view this discussion visit https://groups.google.com/d/msgid/oxcal/LO4P265MB3839EBD332821E019D08B633A5AF2%40LO4P265MB3839.GBRP265.PROD.OUTLOOK.COM.

[Charles Lindsay]

Good morning Alex and Christopher:

[Excuse me if this is a duplicate reply, I don't seem to be able to find what I was drafting a few minutes ago!]

Thank you both for your prompt responses to my query.

Actually, I think I agree with both of you. Yes, the quantile range omits some information, and yes the (often multiple) HPD ranges do clutter up a text with long strings of numbers to the point of making it almost unreadable. My usual solution is  to use the "Whole" or "Quantile" ranges for in-text reference to individual dates and include the full HPD information for all dates in tabular form in an Appendix or SI.

Again, thank you both for your insights on this.

Cheers

Charles