# Reinforcement Learning on CartPole v1 (test)

Source: [Reinforcement Learning on CartPole v1 (test)](<https://www.sota2.com/research/sota/reinforcement-learning-on-cartpole-v1-test>)

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

Data updated: 2026-02-26T06:07:28.047Z

Evaluation of Reinforcement Learning agents on the CartPole v1 environment (test split).

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 | Reinforcement Learning |
| Dataset (including any stated split) | CartPole v1 (test) |
| Metric | Total Reward |
| Unit | SCALAR |
| Direction | Higher is better |
| Results loaded | 25 of 25 |
| 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

- [Total Reward](<https://www.sota2.com/research/sota/reinforcement-learning-on-cartpole-v1-test.md?metric=Total%20Reward>): Higher is better; unit: SCALAR.

- [Metric M](<https://www.sota2.com/research/sota/reinforcement-learning-on-cartpole-v1-test.md?metric=Metric%20M>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Total Reward | Result date | Report paper |
| --- | --- | --- | --- |
| Qualitatively measured policy discrepancy w/ β | 500 | Not recorded | 68fa6e581e209131cea3f30b |
| Qualitatively measured policy discrepancy w/ β | 500 | Not recorded | 68fa6e581e209131cea3f30b |
| Qualitatively measured policy discrepancy w/ η | 500 | Not recorded | 68fa6e581e209131cea3f30b |
| Differentiable Decision Trees | 500 | Not recorded | 68fa6e581e209131cea3f30b |
| Differentiable Decision Trees | 500 | Not recorded | 68fa6e581e209131cea3f30b |
| two-level optimization scheme | 500 | Not recorded | 68fa6e581e209131cea3f30b |
| two-level optimization scheme | 500 | Not recorded | 68fa6e581e209131cea3f30b |
| Qualitatively measured policy discrepancy w/ β | 499.9 | Not recorded | 68fa6e581e209131cea3f30b |
| General Q(λ) | 499.9 | Not recorded | 68fa6e581e209131cea3f30b |
| Qualitatively measured policy discrepancy w/ η | 499.4 | Not recorded | 68fa6e581e209131cea3f30b |
| Importance-Sampling | 498.7 | Not recorded | 68fa6e581e209131cea3f30b |
| Peng & Williams's Q(λ) | 496.7 | Not recorded | 68fa6e581e209131cea3f30b |
| Qualitatively measured policy discrepancy w/ β | 494.9 | Not recorded | 68fa6e581e209131cea3f30b |
| Tree-Backup(λ) | 494.7 | Not recorded | 68fa6e581e209131cea3f30b |
| Qualitatively measured policy discrepancy w/ η | 493.3 | Not recorded | 68fa6e581e209131cea3f30b |
| Qualitatively measured policy discrepancy w/ η | 493.2 | Not recorded | 68fa6e581e209131cea3f30b |
| Q(λ) | 489.9 | Not recorded | 68fa6e581e209131cea3f30b |
| Watkins's Q(λ) | 484.3 | Not recorded | 68fa6e581e209131cea3f30b |
| Retrace(λ) | 461.1 | Not recorded | 68fa6e581e209131cea3f30b |
| Differentiable Decision Trees | 388.76 | Not recorded | 68fa6e581e209131cea3f30b |
| Deep Q Network | 327.3 | Not recorded | 68fa6e581e209131cea3f30b |
| Kronecker-Factored Approximate Curvature | 321 | Not recorded | 68fa6e581e209131cea3f30b |
| Bayesian Deep Reinforcement Learning weighted | 136.75 | Not recorded | 68fa6e581e209131cea3f30b |
| Bayesian Deep Reinforcement Learning | 113.52 | Not recorded | 68fa6e581e209131cea3f30b |
| Deep Q Network | 98.33 | Not recorded | 68fa6e581e209131cea3f30b |

### Qualitatively measured policy discrepancy w/ β

Total Reward: 500. Result ID: 69900e983a6c314849fdce23.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e581e209131cea3f30b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66b11e209131cea11d5a.md>)

Recorded evaluation configuration: {"source\_paper":"\[23\]","instance":3}. This record does not assert independent reproduction.

### Qualitatively measured policy discrepancy w/ β

Total Reward: 500. Result ID: 69900e983a6c314849fdce26.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e581e209131cea3f30b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66b11e209131cea11d5a.md>)

Recorded evaluation configuration: {"source\_paper":"\[23\]","instance":4}. This record does not assert independent reproduction.

### Qualitatively measured policy discrepancy w/ η

Total Reward: 500. Result ID: 69900e983a6c314849fdce27.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e581e209131cea3f30b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66b11e209131cea11d5a.md>)

Recorded evaluation configuration: {"source\_paper":"\[23\]","instance":4}. This record does not assert independent reproduction.

### Differentiable Decision Trees

Total Reward: 500. Result ID: 69900e983a6c314849fdce2d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e581e209131cea3f30b.md>)

Recorded evaluation configuration: {"source\_paper":"\[3\]","note":"Result confirmed by personal communication"}. This record does not assert independent reproduction.

### Differentiable Decision Trees

Total Reward: 500. Result ID: 69900e983a6c314849fdce2e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e581e209131cea3f30b.md>)

Recorded evaluation configuration: {"source\_paper":"\[3\]","note":"Tree simplified using the technique used in this work"}. This record does not assert independent reproduction.

### two-level optimization scheme

Total Reward: 500. Result ID: 69900e983a6c314849fdce2f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e581e209131cea3f30b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e581e209131cea3f30b.md>)

[Implementation](<https://gitlab.com/leocus/ge_dts>)

Recorded evaluation configuration: {"Split Type":"Orthogonal"}. This record does not assert independent reproduction.

### two-level optimization scheme

Total Reward: 500. Result ID: 69900e983a6c314849fdce30.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e581e209131cea3f30b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e581e209131cea3f30b.md>)

[Implementation](<https://gitlab.com/leocus/ge_dts>)

Recorded evaluation configuration: {"Split Type":"Oblique"}. This record does not assert independent reproduction.

### Qualitatively measured policy discrepancy w/ β

Total Reward: 499.9. Result ID: 69900e983a6c314849fdce1d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e581e209131cea3f30b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66b11e209131cea11d5a.md>)

Recorded evaluation configuration: {"source\_paper":"\[23\]","instance":1}. This record does not assert independent reproduction.

### General Q(λ)

Total Reward: 499.9. Result ID: 69900e983a6c314849fdce25.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e581e209131cea3f30b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66b11e209131cea11d5a.md>)

Recorded evaluation configuration: {"source\_paper":"\[23\]"}. This record does not assert independent reproduction.

### Qualitatively measured policy discrepancy w/ η

Total Reward: 499.4. Result ID: 69900e983a6c314849fdce24.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e581e209131cea3f30b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66b11e209131cea11d5a.md>)

Recorded evaluation configuration: {"source\_paper":"\[23\]","instance":3}. This record does not assert independent reproduction.

### Importance-Sampling

Total Reward: 498.7. Result ID: 69900e983a6c314849fdce1a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e581e209131cea3f30b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66b11e209131cea11d5a.md>)

Recorded evaluation configuration: {"source\_paper":"\[23\]"}. This record does not assert independent reproduction.

### Peng & Williams's Q(λ)

Total Reward: 496.7. Result ID: 69900e983a6c314849fdce22.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e581e209131cea3f30b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66b11e209131cea11d5a.md>)

Recorded evaluation configuration: {"source\_paper":"\[23\]"}. This record does not assert independent reproduction.

### Qualitatively measured policy discrepancy w/ β

Total Reward: 494.9. Result ID: 69900e983a6c314849fdce20.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e581e209131cea3f30b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66b11e209131cea11d5a.md>)

Recorded evaluation configuration: {"source\_paper":"\[23\]","instance":2}. This record does not assert independent reproduction.

### Tree-Backup(λ)

Total Reward: 494.7. Result ID: 69900e983a6c314849fdce19.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e581e209131cea3f30b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66b11e209131cea11d5a.md>)

Recorded evaluation configuration: {"source\_paper":"\[23\]"}. This record does not assert independent reproduction.

### Qualitatively measured policy discrepancy w/ η

Total Reward: 493.3. Result ID: 69900e983a6c314849fdce21.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e581e209131cea3f30b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66b11e209131cea11d5a.md>)

Recorded evaluation configuration: {"source\_paper":"\[23\]","instance":2}. This record does not assert independent reproduction.

### Qualitatively measured policy discrepancy w/ η

Total Reward: 493.2. Result ID: 69900e983a6c314849fdce1e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e581e209131cea3f30b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66b11e209131cea11d5a.md>)

Recorded evaluation configuration: {"source\_paper":"\[23\]","instance":1}. This record does not assert independent reproduction.

### Q(λ)

Total Reward: 489.9. Result ID: 69900e983a6c314849fdce1b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e581e209131cea3f30b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66b11e209131cea11d5a.md>)

Recorded evaluation configuration: {"source\_paper":"\[23\]"}. This record does not assert independent reproduction.

### Watkins's Q(λ)

Total Reward: 484.3. Result ID: 69900e983a6c314849fdce1f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e581e209131cea3f30b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66b11e209131cea11d5a.md>)

Recorded evaluation configuration: {"source\_paper":"\[23\]"}. This record does not assert independent reproduction.

### Retrace(λ)

Total Reward: 461.1. Result ID: 69900e983a6c314849fdce1c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e581e209131cea3f30b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66b11e209131cea11d5a.md>)

Recorded evaluation configuration: {"source\_paper":"\[23\]"}. This record does not assert independent reproduction.

### Differentiable Decision Trees

Total Reward: 388.76. Result ID: 69900e983a6c314849fdce2c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e581e209131cea3f30b.md>)

Recorded evaluation configuration: {"source\_paper":"\[3\]"}. This record does not assert independent reproduction.

### Deep Q Network

Total Reward: 327.3. Result ID: 69900e983a6c314849fdce18.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e581e209131cea3f30b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66b11e209131cea11d5a.md>)

Recorded evaluation configuration: {"source\_paper":"\[23\]"}. This record does not assert independent reproduction.

### Kronecker-Factored Approximate Curvature

Total Reward: 321. Result ID: 69900e983a6c314849fdce2b.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e581e209131cea3f30b.md>)

Recorded evaluation configuration: {"source\_paper":"\[25\]"}. This record does not assert independent reproduction.

### Bayesian Deep Reinforcement Learning weighted

Total Reward: 136.75. Result ID: 69900e983a6c314849fdce2a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e581e209131cea3f30b.md>)

Recorded evaluation configuration: {"source\_paper":"\[24\]"}. This record does not assert independent reproduction.

### Bayesian Deep Reinforcement Learning

Total Reward: 113.52. Result ID: 69900e983a6c314849fdce29.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e581e209131cea3f30b.md>)

Recorded evaluation configuration: {"source\_paper":"\[24\]"}. This record does not assert independent reproduction.

### Deep Q Network

Total Reward: 98.33. Result ID: 69900e983a6c314849fdce28.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e581e209131cea3f30b.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa66b11e209131cea11d5a.md>)

Recorded evaluation configuration: {"source\_paper":"\[24\]"}. This record does not assert independent reproduction.

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