Added econometric estimation GSoC 2026 project

Author: tturocy

content/gsoc_2026.md

@@ -49,6 +49,8 @@ The following table summarises our ideas, and includes links to detailed descrip
 | [6](#idea-6-development-and-integration-of-lemkes-algorithm-package)    | Development and integration of Lemke's algorithm package          | Medium/Hard           | Mathematical optimization; interoperability                                | Low |
 | [7](#idea-7-new-algorithms-for-finding-nash-equilibria)                 | New algorithms for finding Nash equilibria                        | Medium/Hard           | Nash equilibrium, algorithms, mathematical optimization                    | High |
 | [8](#idea-8-extensive-form-game-efg-exporter-for-deepminds-openspiel)   | Extensive-form Game .efg exporter for DeepMind's OpenSpiel        | Easy                  | Extensive-form-game; normal-form game; .efg format; .nfg format; OpenSpiel | Medium |
+| [9](#idea-9-econometric-estimation-of-behaviour-in-games)               | Econometric estimation of behaviour in games                      | Medium/Hard           | Econometrics; quantal response; cognitive hierarchy; experience-weighted attraction | Low  |
+
 
 ### Idea 1: "Updating the Gambit GUI"
 
@@ -353,3 +355,54 @@ documentation and tests.
 #### Difficulty rating (easy/medium/hard)
 
 Easy.
+
+
+### Idea 9: "Econometric estimation of behaviour in games"
+
+#### Description
+
+An important use case of Gambit over the years has been to support computations involving statistical or econometric estimation
+in games.  This is especially relevant for research in human-participant experiments in game theory.
+In particular, Gambit's implementation of computation and estimation of
+[quantal response equilibrium](https://gambitproject.readthedocs.io/en/latest/tutorials/advanced_tutorials/quantal_response.html) is the
+de-facto reference implementation. 
+
+However, this implementation is not as easily accessible to users, or integrable into reproducible computational workflows,
+as it could be.
+
+In this project, a student would work on improving our support for econometric estimation.  Depending on interests and
+background, this could take a number of forms:
+- Extending our existing simple tutorials, and replicating the estimations of quantal response equilibria in a variety of published papers
+- Developing a test suite for QRE estimation
+- Developing higher-level functions for estimation of payoff/game parameters
+- Extending Gambit with other statistical and econometric methods used in experimental game theory, such as estimation of
+  cognitive hierarchy, level-k, and/or models such as Experience-Weighted Attraction.
+
+#### Expected outcomes
+
+Depending on the specific direction chosen, the expected outcome would be the integration of improved tutorials, documentations,
+and test/benchmarking suites into Gambit as appropriate.  This project could also be suitable for writing a technical or scientific
+paper suitable for submission to a peer-reviewed journal or conference.
+
+#### Skills required
+
+This project does require some basic familiarity with the relevant econometric and statistical methods and how they are
+applied in the analysis of data from experiments in the literature.  This will not be a project suitable for someone without
+at least an initial familiarity with this literature or directly-relevant coursework.
+
+#### Possible mentors
+
+[Ted Turocy](mailto:ted.turocy@gmail.com)
+
+#### Expected size of project (90, 175, or 350 hours)
+
+The size of the project would depend on the level of ambition.  Simply helping with tutorials and documentation might be a
+90-hour project, while more extensive work on testing and developing general estimation code could be 175 or 350
+depending on the agreed scope.
+
+#### Difficulty rating (easy/medium/hard)
+
+Medium/Hard.
+
+
+