# Price Directional Prediction on S&P 500 (held-out 2017)

Source: [Price Directional Prediction on S&P 500 (held-out 2017)](<https://www.sota2.com/research/sota/price-directional-prediction-on-s-and-p-500-held-out-2017>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-04-28T09:31:05.672Z

Evaluation of models predicting the directional movement (up or down) of the S&P 500 index using data specifically held-out from the year 2017, assessed via directional accuracy.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Price Directional Prediction |
| Dataset (including any stated split) | S&P 500 (held-out 2017) |
| Metric | Directional Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 5 of 5 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Directional Accuracy](<https://www.sota2.com/research/sota/price-directional-prediction-on-s-and-p-500-held-out-2017.md?metric=Directional%20Accuracy>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Directional Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| Hindsight Preference Optimization (HPO) | 57.9 | Not recorded | 69f03f2c55a3c32af89fa19f |
| Claude Sonnet 3.7 | 56.1 | Not recorded | 69f03f2c55a3c32af89fa19f |
| Qwen3-VL-4B | 53.3 | Not recorded | 69f03f2c55a3c32af89fa19f |
| Qwen3-VL-235B | 51.9 | Not recorded | 69f03f2c55a3c32af89fa19f |
| Qwen3-VL-4B | 50.7 | Not recorded | 69f03f2c55a3c32af89fa19f |

### Hindsight Preference Optimization (HPO)

Directional Accuracy: 57.9. Result ID: 69f07dba47f34ce924d6bc55.

[Reporting paper](<https://www.sota2.com/research/paper/69f03f2c55a3c32af89fa19f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69f03f2c55a3c32af89fa19f.md>)

Recorded evaluation configuration: {"Backbone":"Qwen3-VL-4B","Evaluation Protocol":"Hindsight DPO"}. This record does not assert independent reproduction.

### Claude Sonnet 3.7

Directional Accuracy: 56.1. Result ID: 69f07dba47f34ce924d6bc51.

[Reporting paper](<https://www.sota2.com/research/paper/69f03f2c55a3c32af89fa19f.md>)

Recorded evaluation configuration: {"Evaluation Protocol":"zero-shot"}. This record does not assert independent reproduction.

### Qwen3-VL-4B

Directional Accuracy: 53.3. Result ID: 69f07dba47f34ce924d6bc54.

[Reporting paper](<https://www.sota2.com/research/paper/69f03f2c55a3c32af89fa19f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6927db54d6b0dd3342930f66.md>)

Recorded evaluation configuration: {"Evaluation Protocol":"SFT (Stage 1)"}. This record does not assert independent reproduction.

### Qwen3-VL-235B

Directional Accuracy: 51.9. Result ID: 69f07dba47f34ce924d6bc52.

[Reporting paper](<https://www.sota2.com/research/paper/69f03f2c55a3c32af89fa19f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6927db54d6b0dd3342930f66.md>)

Recorded evaluation configuration: {"Evaluation Protocol":"zero-shot"}. This record does not assert independent reproduction.

### Qwen3-VL-4B

Directional Accuracy: 50.7. Result ID: 69f07dba47f34ce924d6bc53.

[Reporting paper](<https://www.sota2.com/research/paper/69f03f2c55a3c32af89fa19f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/6927db54d6b0dd3342930f66.md>)

Recorded evaluation configuration: {"Evaluation Protocol":"zero-shot"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.