Trying methods to improve archaeological inference

Advances in archaeological practice have led to the development of a plethora of inferential techniques to reconstruct the human past. However, their ability to achieve their goals can never be properly estimated, as we will never have an absolute knowledge of what really happened in the past.

Archaeo-riddle is a collaborative project developed to explore the current limits of such methods. We simulated a virtual world and its archaeological record, and ask you, the participants, to answer a series of questions about what happened. Just like in the real world, some questions will have a clear answer, while others may not. But in contrast to the real world, we actually know what happened, and can measure how accurate and robust our inferential techniques are. All participants will have access to the same virtual dataset and contextual background (modulo a few user-driven choices). They will have the opportunity to answer a series of research questions with the method of their choice. The process generating the simulated dataset, i.e. what happened, will be revealed only at the very end. We believe this to be an opportunity to reflect and discuss archaeological theory and method.

Which techniques work better than others? Why, and how much accuracy can they achieve? Do they converge to the same answer? Why does a particular statistical tool work?

We are aware that the range of archaeological inquiries is too vast to be incorporated in a virtual-world setting. Thus, we have decided to narrow our focus on a particular phenomenon (transition to farming), and a particular kind of dataset (site locations with radiocarbon dates and a background environmental map). We believe this provides an opportunity to apply a wide range of techniques, from dispersal models to radiocarbon based demographic inference, from site location analyses to eco-cultural niche models and more.

Our goal is not to make claims about the superiority of any method over another, but to create an opportunity for a collaborative reflection, in an entertaining and pressure-free environment (anonymous participation is also welcome). In doing this, we want to help the discipline in continuing the path already started towards a more robust theoretical and methodological framework.

If you are interested in participating, keep reading the next tabs!!


This project is being carried out by the CDAL, from the University of Cambridge. The researchers involved are:

Alfredo Cortell-Nicolau (as Dr. Pants)

Simon Carrignon (as speaker)

Leah Brainerd (as hunter-gatherer (Rab))

Charles B. Simmons (as farmer (Pop))

Joseph Lewis (as Dr. Stones)

Enrico R. Crema (as random unexpected walker)

With the priceless collaboration of Jasmine Vieri and Chris Stevens

If you want to join us, drop us a line!! Alfredo and Simon's contact at the 'How to participate?' tab!

There are no specific requirements. All you need to do is to download the data, play around, and be willing to share your thoughts with the wider community!! However, we do provide some general guidance to streamline your experience:


1. Context: 1. This is a simulated archaeological context. You can find some contextual information and what our virtual archaeologists currently know in the tab 'The story: what do we know?'

2. Data: The tab 'Data' provides some additional information. This consists of:

2.1. Core shared data: Core shared data: This dataset is the same for everyone. It primarily consists of a raster map containing some information about the background environment, coordinates of the locations of the known sites, and radiocarbon dates.

2.2. Additional data Additional data: Each participant can carry out additional virtual site surveys to collect more information. You can select up to five grid squares from our map and obtain additional data from those locations. To keep the game fair, we ask participants (and research groups) not to make multiple requests and download datasets from more than five grid squares. While we will inspect each request, we ask participants to collaborate with us. This is not about winning, it is about learning!

3. Objectives: Once participants (and research groups) have their own data, they can use any kind of methodology or analysis that they want in order to answer the research questions proposed in the tab 'RQs'. Participants can answer any number of RQs that they want to (even all of them!). In fact, they can also propose their own questions and are indeed encouraged to do so.

4. Timing and contact: The ‘experiment’ started officially on the 6th September 2022. After that, we are allowing approximately one year for the participants to play and develop their methods. During this year, we can be in touch as much as you want. We post new information and news (and videos!) periodically on twitter or mastodon, so you can follow us there to stay tuned!


In any case, if you need any additional information, or just want to discuss different aspects of the project, you can contact




5. Workshop: We will organise a dedicated workshop in the next EAA Conference at Belfast (Aug-Sept 2023). In this workshop, we will show what actually happened and share and discuss all of the mehodological proposals made by the participants.

6. Publication: At the conclusion of the workshop, anyone willing to do so, will be invited to participate in a collaborative paper, where we will show (1) the actual process and the model used to create it, (2) the different contributions and approaches used by the participants and (3) a collaborative reflection of the overall experiment.

7. Involvement: Of course, the level of involvement that you wish to have is entirely up to you. From participating anonymously, not submitting your analyses and results, to even help us coordinate things. Just feel free to join and play in any way you like!

Archaeoriddle is able to help participants wanting to give it a go. More particularly, we will give £650 to the best proposals for the attendance at our workshop in the next EAA meeting in Belfast. These funds will be granted to the researcher presenting the proposal (in the case it is a research group, it will be this group deciding how to dispose of them). They will be given as a prize/single payment, and winners will decide how to use them (no accountability or reports for archaeoriddle are needed). It doesn't matter whether you have already signed up for the EAA or not. If you have not signed up, you can use these to pay fees/travel etc, but if you have, you can use them as a refund to you own expenses. The only thing we do ask is that winners will have to present their own methodological solution at the workshop held in Belfast.

Now, if you want to participate and opt in for the $$, please provide




With a document containing the following information:

- Name and affiliation of authors involved

- Short abstract describing the general idea of the approach undertaken, specifying how the particular methods employed can answer the chosen research question(s) (max 200 words)

- Specific details on the methods employed/planned. Including methods, packages, algorithmic/mathematical description of the methods and script if relevant/available (max 200 words, not counting the script, if provided)

- Short explanation of the strategies used to select the additional tiles from this webpage (see data)

The deadline to send us your proposals will be 16th June. This is the real, for sure, very final deadline. There will be no extensions (we mean it), so please, do make sure to send your proposal on time!!

Each participant will present their results as a communication in a workshop at the next European Association of Archaeologists (EAA) conference (Belfast, 30th August to 2nd September) and, if willing to do so, will take part on a common publication. (Please, note that accepting the award implies the commitment to present your results at the workshop in the EAA)

We have a simulated land (see data for the map) called Rabbithole. We know from archaeological evidence that two distinct populations live in Rabbithole. One of them, the rabbit skinners, had a hunter-gatherer economy, while the other one, the poppy chewers, relied primarily on farming for their survival.

What archaeologists know about Rabbithole is that poppy chewers dispersed in this region and eventually led to the demise of the rabbit skinners culture. However, what they do not know is how exactly did this happen. Did the rabbit skinners promote the expansion of the poppy chewers by adopting their culture and subsistence economy? How was their growth rate compared to that of the poppy chewers? Did perhaps a conflict between the two populations led to the extinction and disappearance of the rabbit skinners?

Unfortunately, the acidic soils of Rabbithole do not provide an optimal context for the preservation of human remains. To date, very few human remains of poppy chewers have been recovered,but their level of preservation does not allow us to extract their DNA. Interestingly, some of these remains show evidence of violence, but the lack of remains from rabbit skinners does not allow us to determine whether these are episodic signs of violence, or evidence of systematic warfare either within poppy chewers communities or between poppy chewers and rabbit skinners.

But, why don't you watch Pop and Rab show you a bit about their cultural transmission processes!?

Our main goal is to reconstruct the relationship between the rabbit skinners and the poppy chewers, and the process that led to the transition to a farming economy in Rabbithole. In order to achieve this primary objective, we propose the following research questions:

RQ1. What was the relationship between the two groups? Was it peaceful or hostile?

RQ2. What was the population trajectory of each group?

RQ3. What was the rate of dispersal of poppy chewers?

You can try to answer one, two or all of them, or you can even propose questions of your own! The key is assessing how our methods work within a known environment!

This is where you can actually get your data, and this is how it works:

The dots you can see in the map are the currently known sites; by clicking on ‘bibliographic data’ you can download all the data that we have at the moment and by clicking on 'Maps' you will download a zip which contains a map and the paleoenvironmental model developed by Dr. Stones and Dr. Pants (basically another map with fitness values). You may have to work out your data a little bit, but that's just real life, isn't it?? This information will be the same for every participant.

However, you can also do your own surveys to complement your dataset! By clicking on the map you will see below the number of the patch you are surveying. You can select up to five more patches/grid squares for your own survey. Once you have decided which five additional patches you want to survey on, you can click on the survey button, which will open a google form. Fill it up with the required information and we will privately send the additional data to you.

Once one user (e-mail) has required five patches, we won’t send another five, so think well which ones you want to ask for. We advise to play a little bit with the common data before asking any additional patches. If different users from the same university/research group require the additional data, we will require confirmation that you are not working on the project together.

And we are on!!! Start downloading your data! But think well which additional patches you want,since you only have one shot!

Bibliographic data Maps

Possible square(s) to survey:
(click on the map to get their ID)


You'll need to copy paste these ID in your submission when you will click on "Survey".

Additional information

Your data consists of:

- One map with height values and another one with values with probabilities of settlement according to environmental fitness (these are common both for the rabbit skinners and the poppy chewers).

- For every site:

- Site name

- Coordinates

- Radiocarbon dates

- Cultural affiliation of the site (is it a rabbit skinner site or a poppy chewer site)

- It is assumed that there are no research biases in the collection of radiocarbon dates

- Great news!! We are happy to announce that we can provide some help for participants to attend to our last workshop in the next EAA Conference (Belfast). Therefore, send us your proposal and we will select the ten best proposals (see Workshop/grant sign up), which will be granted £650 to spend at their will at the next EAA conference!

- If you want to stay tuned, you can either drop us a line (you have the contacts in the 'How to participate?' tab), keep checking this website, or follow us on twitter.

The Archaeoriddle Workshop

Saturday 2nd September 2023, Lanyon Building, room G059

The Archaeoriddle Workshop. In this workshop the final results and the proposals from all participants will be presented. All the code and the methods used to generate the data will be unveiled. The workshop will be held during the EAA conference in Belfast on the Saturday Afternoon at Lanyon Building (G059). It will start at 14h00 and will finish early enough to leave time to people to join the annual EAA dinner party .


  • Simon Carrignon – 14:00
    Introduction to Archaeoriddle & a simple LLM solution.
  • Daniel Hromada – 14:20
    An ABM approach to the Archaeoriddle.
  • Alexes Mes – 14:40
    Quantile Regression and the Bayesian Framework to tackle regional variations in dispersal's tempo.
  • Peter Michael Yaworsky – 15:00
    A multimodal approach, from Logistic Regressions to Spatio-temporal Species Distribution Models.
  • Xuan Zhang – 15:20
    Using Point Process Modelling to detect cooperation vs competition.
  • Deborah Priß & Raphael Kahlenberg – 15:40
    AI and dialectic.
  • Coffee Break – 16:00
    We all need a break, don't we?
  • Alfredo Cortell-Nicolau – 16:30
    The final revelation, code and conclusions
  • Round table, Future work & Publication – 16:55
    During the round table we invite people who haven't submitted proposal to discuss other approaches they think would have been interesting to explore. We will also gather ideas and opinions about the whole projects and propose to anyone interested to join us in writing a paper about modelling in archaeology and why and how we should keep such approaches as open and collaborative as possible.

Every participant of the workshop is invited to join us on the day before for an Ice Breaker dinner at the Holohans. If you plan on comming write us