dorsal/arxiv
View SchemaUsing Artificial Market Models to Forecast Financial Time-Series
| Authors | Nachi Gupta, Raphael Hauser, Neil F. Johnson |
|---|---|
| Categories | |
| ArXiv ID | physics/0506134 |
| URL | https://arxiv.org/abs/physics/0506134 |
Abstract
We discuss the theoretical machinery involved in predicting financial market movements using an artificial market model which has been trained on real financial data. This approach to market prediction - in particular, forecasting financial time-series by training a third-party or 'black box' game on the financial data itself -- was discussed by Johnson et al. in cond-mat/0105303 and cond-mat/0105258 and was based on some encouraging preliminary investigations of the dollar-yen exchange rate, various individual stocks, and stock market indices. However, the initial attempts lacked a clear formal methodology. Here we present a detailed methodology, using optimization techniques to build an estimate of the strategy distribution across the multi-trader population. In contrast to earlier attempts, we are able to present a systematic method for identifying 'pockets of predictability' in real-world markets. We find that as each pocket closes up, the black-box system needs to be 'reset' - which is equivalent to saying that the current probability estimates of the strategy allocation across the multi-trader population are no longer accurate. Instead, new probability estimates need to be obtained by iterative updating, until a new 'pocket of predictability' emerges and reliable prediction can resume.
{
"annotation_id": "ea54d373-fe2f-4481-916a-14604e9cfba7",
"date_created": "2026-03-02T18:01:00.419000Z",
"date_modified": "2026-03-02T18:01:00.419000Z",
"file_hash": "41b0d8eb16624ed01fdcbf7ab0d37f5befa78905c9ef157a9b09bf8f16f6ed88",
"private": false,
"record": {
"abstract": "We discuss the theoretical machinery involved in predicting financial market\nmovements using an artificial market model which has been trained on real\nfinancial data. This approach to market prediction - in particular, forecasting\nfinancial time-series by training a third-party or \u0027black box\u0027 game on the\nfinancial data itself -- was discussed by Johnson et al. in cond-mat/0105303\nand cond-mat/0105258 and was based on some encouraging preliminary\ninvestigations of the dollar-yen exchange rate, various individual stocks, and\nstock market indices. However, the initial attempts lacked a clear formal\nmethodology. Here we present a detailed methodology, using optimization\ntechniques to build an estimate of the strategy distribution across the\nmulti-trader population. In contrast to earlier attempts, we are able to\npresent a systematic method for identifying \u0027pockets of predictability\u0027 in\nreal-world markets. We find that as each pocket closes up, the black-box system\nneeds to be \u0027reset\u0027 - which is equivalent to saying that the current\nprobability estimates of the strategy allocation across the multi-trader\npopulation are no longer accurate. Instead, new probability estimates need to\nbe obtained by iterative updating, until a new \u0027pocket of predictability\u0027\nemerges and reliable prediction can resume.",
"arxiv_id": "physics/0506134",
"authors": [
"Nachi Gupta",
"Raphael Hauser",
"Neil F. Johnson"
],
"categories": [
"physics.soc-ph",
"cond-mat.dis-nn"
],
"title": "Using Artificial Market Models to Forecast Financial Time-Series",
"url": "https://arxiv.org/abs/physics/0506134"
},
"schema_id": "dorsal/arxiv",
"source": {
"execution_id": "d710e395-a9c1-4115-b52a-b1411472c132",
"id": "arXiv Dataset IDs",
"type": "Model",
"variant": "snapshot-2026-03-01",
"version": "0.1.0"
},
"user_id": 1000002
}