# The Courage to Stop: Overcoming Sunk Cost Fallacy in Deep Reinforcement Learning

arXiv: https://arxiv.org/abs/2506.13672 · HTML: https://mlsanity.com/paper/2506.13672/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- Placed in Deep RL algorithms and representations, one of 15 directions in Deep reinforcement learning methods.
- Builds on 48 of its 63 references that are in the map.
- First posted on arXiv 16 Jun 2025.

## Placed in

- Deep reinforcement learning methods / [Deep RL algorithms and representations](https://mlsanity.com/deeprl/deep-rl-algorithms-and-representations/)

## Builds on (most cited first, 30 of 48)

- [Soft Actor-Critic: Off-Policy Maximum Entropy Deep Reinforcement Learning with a Stochastic Actor](https://mlsanity.com/paper/1801.01290/) (2018)
- [Continuous control with deep reinforcement learning](https://mlsanity.com/paper/1509.02971/) (2015)
- [Addressing Function Approximation Error in Actor-Critic Methods](https://mlsanity.com/paper/1802.09477/) (2018)
- [Deep Reinforcement Learning with Double Q-learning](https://mlsanity.com/paper/1509.06461/) (2015)
- [OpenAI Gym](https://mlsanity.com/paper/1606.01540/) (2016)
- [Curiosity-driven Exploration by Self-supervised Prediction](https://mlsanity.com/paper/1705.05363/) (2017)
- [Prioritized Experience Replay](https://mlsanity.com/paper/1511.05952/) (2015)
- [DeepMind Control Suite](https://mlsanity.com/paper/1801.00690/) (2018)
- [Mastering Visual Continuous Control: Improved Data-Augmented Reinforcement Learning](https://mlsanity.com/paper/2107.09645/) (2021)
- [Randomized Ensembled Double Q-Learning: Learning Fast Without a Model](https://mlsanity.com/paper/2101.05982/) (2021)
- [The Primacy Bias in Deep Reinforcement Learning](https://mlsanity.com/paper/2205.07802/) (2022)
- [Scaling Robot Learning with Semantically Imagined Experience](https://mlsanity.com/paper/2302.11550/) (2023)
- [Revisiting Fundamentals of Experience Replay](https://mlsanity.com/paper/2007.06700/) (2020)
- [The Dormant Neuron Phenomenon in Deep Reinforcement Learning](https://mlsanity.com/paper/2302.12902/) (2023)
- [Bigger, Better, Faster: Human-level Atari with human-level efficiency](https://mlsanity.com/paper/2305.19452/) (2023)
- [Loss of Plasticity in Continual Deep Reinforcement Learning](https://mlsanity.com/paper/2303.07507/) (2023)
- [Dropout Q-Functions for Doubly Efficient Reinforcement Learning](https://mlsanity.com/paper/2110.02034/) (2021)
- [Optimism in Reinforcement Learning with Generalized Linear Function Approximation](https://mlsanity.com/paper/1912.04136/) (2019)
- [Bigger, Regularized, Optimistic: scaling for compute and sample-efficient continuous control](https://mlsanity.com/paper/2405.16158/) (2024)
- [Synthetic Experience Replay](https://mlsanity.com/paper/2303.06614/) (2023)
- [Time Limits in Reinforcement Learning](https://mlsanity.com/paper/1712.00378/) (2017)
- [Revisiting Rainbow: Promoting more Insightful and Inclusive Deep Reinforcement Learning Research](https://mlsanity.com/paper/2011.14826/) (2020)
- [Mixtures of Experts Unlock Parameter Scaling for Deep RL](https://mlsanity.com/paper/2402.08609/) (2024)
- [TACO: Temporal Latent Action-Driven Contrastive Loss for Visual Reinforcement Learning](https://mlsanity.com/paper/2306.13229/) (2023)
- [Tactical Optimism and Pessimism for Deep Reinforcement Learning](https://mlsanity.com/paper/2102.03765/) (2021)
- [DrM: Mastering Visual Reinforcement Learning through Dormant Ratio Minimization](https://mlsanity.com/paper/2310.19668/) (2023)
- [In value-based deep reinforcement learning, a pruned network is a good network](https://mlsanity.com/paper/2402.12479/) (2024)
- [Small batch deep reinforcement learning](https://mlsanity.com/paper/2310.03882/) (2023)
- [HyAR: Addressing Discrete-Continuous Action Reinforcement Learning via Hybrid Action Representation](https://mlsanity.com/paper/2109.05490/) (2021)
- [Softmax Deep Double Deterministic Policy Gradients](https://mlsanity.com/paper/2010.09177/) (2020)

## Built on by (most cited first, 6 of 6)

- [Stable Gradients for Stable Learning at Scale in Deep Reinforcement Learning](https://mlsanity.com/paper/2506.15544/) (2025)
- [Simplicial Embeddings Improve Sample Efficiency in Actor-Critic Agents](https://mlsanity.com/paper/2510.13704/) (2025)
- [Asymmetric Proximal Policy Optimization: mini-critics boost LLM reasoning](https://mlsanity.com/paper/2510.01656/) (2025)
- [Stable Deep Reinforcement Learning via Isotropic Gaussian Representations](https://mlsanity.com/paper/2602.19373/) (2026)
- [When are LLMs Sufficient Policy Optimizers for Sequential RL Tasks?](https://mlsanity.com/paper/2605.30719/) (2026)
- [EBRL: Asynchronous Embodied RL by Multi-Grained Resource Management](https://mlsanity.com/paper/2609.27547/) (2026)

## Abstract (arXiv)

Off-policy deep reinforcement learning (RL) typically leverages replay buffers for reusing past experiences during learning. This can help improve sample efficiency when the collected data is informative and aligned with the learning objectives; when that is not the case, it can have the effect of "polluting" the replay buffer with data which can exacerbate optimization challenges in addition to wasting environment interactions due to wasteful sampling. We argue that sampling these uninformative and wasteful transitions can be avoided by addressing the sunk cost fallacy, which, in the context of deep RL, is the tendency towards continuing an episode until termination. To address this, we propose learn to stop (LEAST), a lightweight mechanism that enables strategic early episode termination based on Q-value and gradient statistics, which helps agents recognize when to terminate unproductive episodes early. We demonstrate that our method improves learning efficiency on a variety of RL algorithms, evaluated on both the MuJoCo and DeepMind Control Suite benchmarks.