Challenge Rules

# Global Chess Challenge 2025 # Official Rules **NO PURCHASE NECESSARY TO PARTICIPATE. VOID WHERE PROHIBITED.** By participating in this Challenge, you agree to abide by these Official Rules and all decisions of the Organizers. ### Definitions **Sponsor:** Amazon Web Services, Inc. (“AWS”), providing financial and brand sponsorship. AWS is not responsible for the organization or execution of the Challenge. **Organizer:** AGI House, responsible for administering, managing, and operating the Challenge. **Platform Provider:** AIcrowd SA, responsible for providing the online platform and leaderboard infrastructure. AIcrowd is not an Organizer or Sponsor. ### Overview The Global Chess Challenge 2025 ("Challenge") is a global hybrid competition organized by AGI House and sponsored by Amazon Web Services (AWS). Infrastructure and leaderboard services are provided by AIcrowd. The competition focuses on leveraging **AWS Trainium chips** to train and evaluate a family of **Small Language Models (SLMs)** specialized for **reasoning in text-based chess environments**. Participants are tasked with building **text-only models** that, given a chess board position, can: * Produce **legal chess moves** * Generate **succinct textual explanations** justifying those moves * Operate efficiently within the computational constraints of AWS Trainium hardware All submitted models are executed independently by the Organizers on controlled infrastructure. Submissions must operate as **standalone language models**, producing moves solely via token prediction. External tool use, heuristic search procedures, or runtime access to chess engines or auxiliary decision systems is not permitted. The challenge emphasizes **reasoning efficiency, strategic quality, and hardware-aware optimization**, rather than raw model scale. **Stockfish** serves as the objective verifier for move legality and strategic soundness, while large public chess corpora may be used to provide scale for training and evaluation. ### Sponsor Amazon Web Services, Inc., 410 Terry Ave North, Seattle, WA 98109. AWS is the sponsor of the Challenge and is not responsible for the administration or organization of the event. ### Organizers & Admins AGI House is the sole Organizer of the AWS Trainium Challenge 2025. “Organizer Admins” are companies or entities authorized by the Organizer to support execution of the Challenge. For this event, this includes AIcrowd SA as a platform and evaluation infrastructure provider. Organizer Admins are not responsible for the overall organization or legal administration of the Challenge. ### Timeline, Website & Registration * **Launch & Registration Opens: AWS re:Invent, December 4, 2025** * **Round 1 Submissions Close: December 31, 2025 (23:55 UTC)** * **Round 2 Submissions Close: January 31, 2026 (23:55 UTC)** * **Final Tournament: Feb 1st, 2026 - Feb 7, 2026** * **Winners Announced: Feb 15th, 2026** ### Challenge Evaluation **Overview** Evaluation will proceed in multiple stages designed to assess **move quality, reasoning consistency, and competitive strength** across both baseline play and peer-to-peer competition. **Round 1 and Round 2: Baseline Evaluation** Across **Round 1** and **Round 2**, all submitted models will be evaluated against fixed Stockfish opponents to establish a stable, comparable performance baseline. For each round: * Each submission will play **50 games against Stockfish Skill Level 0 (Depth 1)** and **50 games against Stockfish Skill Level 0 (Depth 5)**. * Game positions, time controls, and computational resources will be standardized across all submissions. Scoring for Round 1 and Round 2 will be defined as follows: * **Primary Score – Average Centipawn Loss (ACPL):** ACPL will be computed by analyzing each played game using **Stockfish Level 20 (Depth 20)** as the reference evaluator. Lower ACPL indicates stronger move quality and positional reasoning. * **Secondary Score – Win Rate:** Win Rate will be calculated across all games played against both Stockfish configurations. **Important Clarification on Reasoning Text:** While submissions are required to output a textual explanation of their move, **the reasoning text is not scored directly**. Evaluation is based **exclusively on the UCI move provided within the `` tags**. Any text outside these tags is ignored for scoring purposes. If a submission does not provide a **valid UCI move** within the `` tags (including missing tags, malformed output, or illegal moves), the evaluation system will **retry the model up to three (3) times** for the same position. If the submission still fails to produce a valid UCI move after these retries, the model will be treated as having **resigned the game**, and the game will be recorded as a **loss** for that submission. ACPL will serve as the primary ranking signal for leaderboard ordering during Round 1 and Round 2. Win Rate will be used as a secondary metric for tie-breaking and additional analysis. **Eligibility for Final Tournament** At the conclusion of Round 2, submissions that achieve an **ACPL lower than the official baseline model defined by the Organizers** will be deemed _eligible submissions_ and will advance to the Final Tournament stage. **Final Tournament: Swiss-Style Competition** Eligible submissions will compete in a **Swiss-style tournament** to determine final placements and prize winners. Key properties of the Final Tournament include: * All eligible models will participate in a fixed number of rounds determined by the Organizers based on the number of qualifying submissions. * Pairings will be generated using a Swiss-system format, ensuring that models with similar performance records compete against one another in successive rounds. * Each match will consist of a predefined number of games with alternating colors and standardized time controls. * Game outcomes will be scored **solely using win/loss/draw results**. **ACPL will not be used** in the Swiss-style Final Tournament. **Tournament Scoring and Final Rankings** The Swiss-style Final Tournament will follow a **standard Swiss-system chess tournament format**, consistent with widely used rules such as those published by FIDE. * **Tournament performance** is defined strictly as the **total number of points earned from game outcomes** during the Swiss tournament rounds. * Each game will be scored using the conventional chess scoring system: * **Win:** 1 point * **Draw:** 0.5 points * **Loss:** 0 points * Final rankings will be determined by **total tournament points** accumulated across all Swiss rounds. **Tie-Breaking Rules** If two or more submissions finish with equal total tournament points, tie-breaks will be applied in the following order, unless otherwise announced prior to tournament start: 1. **Head-to-head result(s)** between tied submissions, if applicable 2. **Buchholz score** (sum of opponents’ final tournament points), or an equivalent strength-of-opposition metric 3. **Sonneborn–Berger score**, where applicable 4. Any additional tie-breaking procedure announced by the Organizers prior to the start of the Final Tournament No engine-quality metrics (including ACPL) will be used during the Swiss-style Final Tournament. Only game outcomes will be considered for scoring and ranking. All models will be run in a controlled environment with identical computational resources and execution constraints throughout all stages of evaluation. The Organizers reserve the right to refine tournament parameters, scoring aggregation, or tie-breaking rules prior to the Final Tournament to ensure fairness, scalability, and technical integrity. ### Am I Eligible to Enter the Challenge? You are eligible if you (and each team member): * Are an individual * Are 18 years or older (and of majority age in your jurisdiction) * Have Internet access, an email account, and a personal computer **Residents of the following are not eligible for any prizes, including cash prizes or compute credits:** Crimea region of Ukraine, Cuba, Iran, North Korea, Sudan, Syria, Quebec (Canada), Brazil, Italy, Russian Federation. _Note: Residents of these regions may still participate and retain their leaderboard rank. Any cash prizes will be passed to the next eligible team._ Employees, contractors, or household members of Amazon, AGI House, and AI Crowd are not eligible to receive prizes. You are responsible for reviewing and understanding your employer's and country's policies regarding participation. If you participate in violation of these policies, you and your Entry may be disqualified. Organizers disclaim liability for disputes arising between employers and employees or countries and residents. If your employer has rights to intellectual property created during employment, you must confirm you have received authorization to participate, or that your employer has waived such rights for this Challenge. Organizers may request written confirmation. ### Is the Entry an Eligible Entry? The Entry MUST: * be compatible with the official submission format * be in English * be the Team's own original work * not have been submitted previously in any promotion of any kind * not contain inappropriate, offensive, defamatory, or illegal content, or content that promotes discrimination, violence, or infringes on others' rights And The Team members MUST: * Obtain all necessary consents, approvals, and licenses for submission * Obtain agreement from all team members to these Rules * Not engage in false, fraudulent, or deceptive acts * Not tamper with or abuse any aspect of the Challenge * Ensure transparency in the use of external data sources with clear documentation * Ensure external datasets are permissible for non-commercial or academic research and comply with licensing terms * Provide rationale for any declared datasets **Violation of these terms will result in immediate disqualification.** ### Submission Timeouts and Infrastructure Requirements Submissions will be evaluated on a standard hardware configuration disclosed prior to the deadline. Each submission will be subject to strict runtime limits per task, including per-turn timeouts in interactive tasks. Submissions that exceed time limits or fail to execute on the designated infrastructure may be marked as failed. Participants are responsible for ensuring compatibility with evaluation hardware and accounting for resource constraints such as memory limits and disabled external network access. Submission Limits and Documentation * Teams may submit up to **20 entries per day**. * Winning solutions must be validated for compliance with allowed data and thoroughly documented. * Minimum performance thresholds must be met to qualify for prizes; specifically, submissions must achieve an Average Centipawn Loss (ACPL) lower than that of the official baseline model defined by the Organizers. * Winners must submit a comprehensive solution report for publication on a platform designated by organizers. **Eligible Models and Execution Backends** Only a **subset of models and execution backends** are eligible for evaluation in this Challenge, reflecting the **currently supported and performance-optimized model configurations for AWS Trainium** and the specific execution environment used for the competition. Participants must use **only those model architectures, model families, and inference or training backends explicitly supported by the official Challenge infrastructure**, as documented in the following reference: https://github.com/AIcrowd/global-chess-challenge-2025-starter-kit/blob/master/docs/neuron-and-vllm-tuning.md#supported-model-types-backends **Model Execution Constraints** Submissions are evaluated by executing the submitted model **as-is** on Organizer-controlled infrastructure. Models do not have access to external tools, function calling, heuristic search procedures, or runtime interaction with chess engines or auxiliary decision systems. All decisions must be produced solely through **token-level language model inference** based on the provided textual input. Any form of external computation, search, retrieval, tool invocation, or handcrafted rule-based logic executed at inference time is prohibited. Submissions that attempt to bypass these constraints (whether through embedded engines, indirect tool use, or hidden heuristics) may be disqualified at the discretion of the Organizers. **Model Size Limitation** Only models with a total parameter count of **strictly fewer than eight point five billion (8,500,000,000) parameters** are eligible for participation, leaderboard ranking, Final Tournament qualification, and prizes in this Challenge. The parameter count shall be determined based on the total number of trainable parameters in the submitted model at inference time, excluding optimizer states but including all model weights. Any submission found to exceed this parameter limit may be disqualified or deemed ineligible for prizes at the discretion of the Organizers. ### Disqualification If you, any Team member, or the Entry is found to be ineligible for any reason (including but not limited to conflicts within Teams, noncompliance with these Rules, or tampering with submissions), the Organizer reserves the right to disqualify the Entry and/or you and/or your Team members from this Challenge. A participant may not create or use multiple accounts to participate. Violating this rule will result in immediate disqualification and potential exclusion from future events. The Organizer also reserves the right to disqualify participants for fraud, deception, interference with the leaderboard, or any other activity that undermines the integrity of the Challenge. ### Intellectual Property and Use of Entries Participants retain ownership of their Entries, including any models or code developed during the Challenge. By submitting an Entry, participants grant AGI House, AWS, AIcrowd SA, and their respective partners, sponsors, and affiliates a non-exclusive, worldwide, royalty-free, irrevocable license to: * Review, evaluate, and display (or not display) the Entry in connection with the Challenge * Use the Entry for research, evaluation, publication, and promotional purposes related to the Challenge * Use the Entry in any media for non-commercial or commercial purposes in connection with marketing, sale, or promotion of Organizers, sponsors, partners, and their products and services * Publish results, model performance data, outputs, evaluation results, and descriptions of winning approaches The Base Model and any datasets provided by organizers remain their exclusive property and may only be used within the scope of the Challenge. Trained or derivative models that build upon the Base Model must remain open-sourced under a permissive license (e.g., MIT, Apache 2.0). Participants using their own data or third-party components are responsible for ensuring proper rights and license compliance. Participants acknowledge that: * They will not be compensated and may not be credited (at Organizers' discretion) for use of the Entry * Organizers, sponsors, and partners may have developed similar materials and waive any claims from similarities * Entries may be posted publicly and shared with influencers, media partners, and other third parties for promotional purposes; Organizers are not responsible for unauthorized use by visitors or third parties * Organizers are not obligated to use the Entry in any way, even if selected as a winning Entry ### How May the Entry Potentially Be Used? The Entry may be used in a few different ways. Organizers do not claim to own your Team's Entry; however, by submitting the Entry you and each member of your Team: * hereby grants to Organizers Admins a non-exclusive, irrevocable, royalty-free, world-wide right and license to review and analyze the Entry in relation to this Challenge * hereby grants to Organizers and Organizers Admins a non-exclusive, irrevocable, royalty-free, world-wide right and license to use the Entry or parts of your Entry in any media for any non-commercial or commercial purpose in connection with the marketing, sale, or promotion of Organizers, Organizers Admins and their respective products and services * agrees that each member will execute any necessary paperwork for Organizers and Organizers Admins to use the rights and licenses granted hereunder * acknowledges and agrees that the Team will not be compensated and may not be credited (at Organizers's sole discretion) for the use of the Entry as described in these Rules * acknowledges that the Organizers or Organizers Admins may have developed or commissioned materials similar to the Entry and waive any claims resulting from any similarities to the Entry * understand that the Entry may be posted on a public website or social media channel and that Organizers is not responsible for any unauthorized use of the Entry by visitors to such site * understand and acknowledge that, subject to provision of Prizes, Organizers are not obligated to use the Entry in any way, even if the Entry is selected as a winning Entry * Personal data you submit in relation to this Challenge will be used by Organizers and Organizer Admins in accordance to Section 15 of these Rules. Organizers may publish outputs, evaluation results, and analytical findings derived from submissions. ### How Will Winners Be Selected and Notified? Entries will be ranked on the AIcrowd leaderboard based on algorithmic scoring during Round 1 and Round 2 (**Primary: ACPL; Secondary: Win Rate**). At the conclusion of Round 2, only submissions that meet the minimum performance threshold (**ACPL lower than the official baseline model defined by the Organizers**) will be eligible to proceed. Final winners will be determined by the results of the **Swiss-style Final Tournament** among eligible submissions, using **game outcomes only** (**Win = 1, Draw = 0.5, Loss = 0**). Tie-breaks, if required, will follow the tie-breaking rules stated in the **Challenge Evaluation** section. Potential winners will be contacted via the email address associated with the **Team Leader’s** AIcrowd account. For the purposes of these Rules, the **Team Leader** is the individual designated on the AIcrowd platform as the primary point of contact and authorized representative of the Team. In the case of a **single-participant Entry** (i.e., an Entry submitted by an individual not formally registered as part of a Team), that individual shall be **deemed the Team Leader** for all purposes under these Rules. If a potential winner cannot be contacted, does not respond as directed, refuses the prize, or is found ineligible, the prize may be forfeited and awarded to the **next eligible Team on the leaderboard**, as determined by the Organizers. To the extent that there is any dispute as to the identity of the potential winner, the official account holder of the email address associated with the **Team Leader’s** AIcrowd account (as designated on the AIcrowd platform) will be deemed the official potential winner by the Organizers. In the case of a single-participant Entry not formally registered as part of a Team, the individual participant will be deemed the Team Leader for purposes of this provision. Winners are required to document their methodology in detail. This includes the submission of a comprehensive solution report, which must be prepared for publication on a platform designated by the organisers. As a **condition of prize eligibility and acceptance**, Winners are required to document their methodology in detail. This includes the submission of a **comprehensive solution report** prepared for publication on a platform designated by the Organizers. The report must accurately and substantively describe the technical approach used in the winning Entry; submissions that are materially incomplete, misleading, or consist primarily of generic or automatically generated content without meaningful technical detail may, at the Organizers’ discretion, result in disqualification or forfeiture of prizes. ### Prizes and Conditions Prize amounts and award categories will be announced on the official Challenge webpage. * Prizes are non-transferable and may not be substituted except at the Organizer's discretion * Winners are solely responsible for all applicable taxes, withholdings, and costs associated with prizes * Winners must complete required documentation (e.g., W-9 or W-8BEN tax forms) within 14 days of notification * Failure to provide complete documentation may result in forfeiture * The Organizer reserves the right not to award a prize if a winning Entry cannot be reproduced, verified, or validated * Prizes will be awarded within 6 months from the end of the Challenge * Prize distribution may be facilitated through external partners; winner information will be shared as necessary for prize fulfillment * Disputes among team members regarding prize sharing are not the responsibility of Organizers A list of winners will be posted on AIcrowd and may be announced on official AWS or AGI House channels. # Governing Law and Dispute Resolution All disputes will be administered by AGI House as Organizer. AWS, as Sponsor, will not serve as a party to dispute resolution. This Challenge and any dispute arising under or related to it will be governed by the laws of the State of California, without regard to conflict of law principles. All claims and disputes must be resolved exclusively in the state or federal courts located in San Mateo County, California, and each participant consents to personal jurisdiction in such courts. If such a venue is unenforceable, an alternative venue within the participant's jurisdiction may apply. Participants waive any right to participate in a class action or class arbitration related to the Challenge. # Personal Data and Privacy Organizers may use cookies and/or collect IP addresses for implementing rights and obligations under these Rules, identifying your location, or any other lawful purpose in accordance with the Privacy Policy. Organizers may use personal data you provide to: * Contact you regarding the Challenge * Confirm details of your Entry * Administer and execute the Challenge, including sharing with Organizer Admins and external partners for prize distribution * Credit you and/or your team for the Entry or identify you as a Winner * Meet obligations under these Rules or applicable law Only name and email address are required to participate. Please read the AIcrowd Site terms and conditions carefully to understand how your data may be used by AIcrowd SA. **Force Majeure and Event Disruption** Organizers reserve the right to cancel, suspend, or modify the Challenge if circumstances beyond their reasonable control compromise the fairness, integrity, or feasibility of the competition, including but not limited to natural disasters, war, public health emergencies, system failures, or regulatory actions. Organizers may determine how prizes will be awarded, if at all, based on eligible entries received prior to disruption. The Organizer reserves the right to make reasonable updates to the structure, schedule, evaluation criteria, and technical framework of the Challenge to enhance clarity, improve fairness, or respond to unforeseen circumstances. Updates will be communicated via official Challenge channels. Participants will not be required to formally re-accept these Rules unless updates materially alter eligibility, prize structure, data use rights, or other legally significant terms. ## Limitation of Liability and Additional Terms To the fullest extent permitted by law, you agree that the Organizer and their directors, officers, employees, agents and assigns will not be liable for personal injuries, death, damages, expenses, costs, or losses of any kind resulting from participation or inability to participate in this Challenge or acceptance, use, or inability to use a prize, whether under contract, tort (including negligence), warranty, or other theory. If Organizers determine that any portion of this Challenge is compromised by virus, bugs, unauthorized intervention, or other causes beyond their control that corrupt or impair the administration, security, fairness, or proper participation, Organizers may: (a) cancel the Challenge; (b) pause the Challenge until issues are resolved; or (c) consider only Entries submitted prior to compromise. Your use of any products or services required by these Rules is subject to the terms and conditions of those products or services, including the AIcrowd site and services. If any provision of these Rules is deemed unenforceable, the remaining provisions remain valid and in effect.