# About SIRE
**The first autonomous sports intelligence network designed for prediction markets.** SIRE is an agentic, on-chain sports-betting hedge fund that transforms live sports data into deployable intelligence. It fuses Score Vision’s decentralised computer-vision engine with advanced quantitative models to identify market inefficiencies and make them accessible through on-chain tools. #### Core Products * **aVault** — Automated strategy pools where users can **add USDC to receive aVault units** and access model-driven execution. * **aLink** — A token-gated LLM terminal for SIRE holders, offering live model insights, market mispricing alerts, and quantitative reasoning in plain language. #### How It Works SIRE connects quantitative modelling, visual intelligence, machine learning, and DeFi infrastructure within a single adaptive framework. The system ingests real-time match footage and market data, analyses probabilities through proprietary models, and continuously refines its predictions based on live outcomes. #### The Goal To democratise institutional-grade sports intelligence: turning every match into an opportunity for data-driven participation, with transparent systems and permissionless access. SIRE makes advanced sports analytics and agentic execution accessible to everyone, transforming real-world sports activity into an on-chain, uncorrelated potential yield stream driven by data rather than market speculation. #### Tokenomics at a Glance SIRE’s token economy aligns network activity with long-term sustainability.\ Each action: adding to aVault, staking, or using aLink, feeds into transparent on-chain fee flows designed to support the protocol. **Core Utilities** * **Stake** – Participate in fee-funded potential uncorrelated returns from SIRE products adoption. * **Access** – Unlock aLink and its analytical tools using SIRE. * **Reduce Fees** – Lower aVault withdrawal and performance fees through staking. * **Buybacks** – A share of protocol activity funds' ongoing SIRE buybacks. *For full details on fees, distributions, and buyback mechanisms, see the* [Tokenomics](/tokenomics/overview) page # Why SIRE Exists #### Opening up a closed world of sports intelligence Sports betting is one of the world’s largest data markets, yet the intelligence that drives it has always been locked away. A handful of syndicates, bookmakers, and hedge funds control private models that interpret billions of data points. For everyone else, betting remains a black box: decisions made on incomplete information and outcomes shaped by hidden odds engines. #### The Problem Traditional betting systems are built around opacity and imbalance. * **Markets are reactive, not predictive.** Odds often reflect liquidity and sentiment, not the true probability of an event. * **Data is trapped.** Most high-quality sports data is proprietary, expensive, and unavailable to the public. * **Access is gated.** Even when intelligence exists, it lives behind closed APIs and private algorithms. This creates an ecosystem where only a select few can identify market inefficiencies, while the majority play blind. #### **The Opportunity** Sports markets handle hundreds of billions of dollars annually. By connecting this continual flow of real-world activity to DeFi, SIRE transforms it into programmable, on-chain intelligence. This connection turns sporting events into a new foundation for uncorrelated, data-driven performance that moves independently of crypto markets. #### The Breakthrough SIRE emerged from **Score Vision**, a decentralised computer-vision system initially designed to analyse football matches in real time. By transforming raw footage into structured data, the team built a foundation for transparent, verifiable intelligence that anyone could access. From there, the concept evolved: combine that data layer with quantitative models and deploy it through on-chain infrastructure. The result is an agentic framework that connects sports analytics directly to programmable execution. #### The Vision SIRE exists to make sports intelligence itself permissionless. Our mission is to decentralise predictive infrastructure, allowing anyone to access and interact with the same analytical power once reserved for professional quants. Every addition of data, model, and strategy strengthens the network, creating a transparent ecosystem where collective intelligence compounds over time. > *SIRE turns private edge into public infrastructure, opening the world’s most data-rich market to everyone.* # Mission & Purpose #### **Mission** SIRE’s mission is to make advanced sports intelligence openly accessible through decentralised infrastructure.\ It connects data science, machine learning, and on-chain automation to create a transparent intelligence layer anyone can use, democratising institutional-grade analytics for real-world, uncorrelated performance. #### **Core Principles** * **Transparency:** All activity, fees, and agentic execution are verifiable on-chain. * **Accessibility:** Open participation in SIRE’s intelligence network, with access to data-driven, uncorrelated performance. * **Alignment:** Users, holders, and the protocol advance together through shared incentives. * **Autonomy:** AI agents continually learn from live data, refine their models, and execute strategies directly on-chain. * **Fairness:** SIRE launched through a fair, open-market distribution — with no presale, private allocation, or insider advantage — ensuring the network’s ownership and opportunity remain in the hands of its community. #### **Who It’s For** * **Sports Bettors:** Use model-driven insights to make data-backed decisions and manage risk. * **DeFi Users:** Diversify exposure with uncorrelated performance opportunities by adding USDC to receive αVault units and staking SIRE for reduced fees. * **DeFi Protocols & Partners:** Integrate αVault and αLink to give communities direct access to SIRE’s intelligence stack and on-chain execution. * **Contributors & DAO Members:** As a community-owned protocol, anyone can contribute research, build integrations, propose improvements, or support ecosystem growth through governance and bounties. # Core Contributors Meet the Core Contributors behind the first agentic sports betting DAO ## Core Contributors SIRE is developed and maintained by a distributed network of contributors coordinated through the SIRE DAO. The protocol was initiated by the team behind **Score Technologies**, who built the decentralised computer-vision system **Score Vision**, operating on **Bittensor Subnet 44**.
Today, SIRE continues to evolve through open collaboration between engineers, data scientists, analysts, and community members from around the world. #### Founding Contributors The following individuals established the core infrastructure and research foundations of SIRE: * **Maxime Sebti** – Co-founder and CEO of Score Technologies, guiding overall protocol strategy, partnerships, and ecosystem development at SIRE. * **Tim Kalic** – Co-founder and CTO of Score Technologies, where he built the Score Vision computer-vision infrastructure. He also contributes to the technical development of SIRE. * **Nigel Grant** – Co-founder and Chief Revenue Officer at Score Technologies. He leads commercial strategy and partnerships, leveraging an extensive global network across sports and technology to support SIRE’s ecosystem growth. * **Dr. Peter Cotton** – Quantitative research lead at Score Technologies, specialising in predictive modelling and statistical learning. Formerly with Microsoft and JP Morgan. #### Community & DAO Governance SIRE is a community-owned **protocol** built entirely by its community. There was **no presale.** All tokens were acquired through the open market.\ Anyone can contribute by participating in governance, supporting development, or helping grow the ecosystem through community-driven initiatives. The DAO represents the next stage of SIRE’s evolution — a collective intelligence network owned and advanced by its community. > *SIRE is built by its community. From research and model design to ecosystem partnerships, every contribution strengthens the intelligence network that powers the protocol.* \ **IMPORTANT LINKS**\ \ [Score corporate website](https://www.wearescore.com)\ [Score vision subnet website](https://www.scorevision.io)\ [Score twitter account](https://www.x.com/webuildscore) # Overview Explore SIRE's revolutionary approach to sports betting intelligence SIRE operates as a multi-layered intelligence network that transforms real-world sports data into on-chain execution. It combines computer vision, quantitative modelling, and AI-agent automation to detect inefficiencies and act on them programmatically. #### The Technology SIRE runs on three interlocking layers that transform live sport into deployable intelligence. While elements of this architecture are already powering αVault’s strategies, several components — particularly the advanced vision and multi-source optimisation systems — are actively in development. #### **1. Unique Data Layer** **Score Vision (Bittensor Subnet 44)** is a primary contributor. It converts live video into tracking-grade coordinates and events, then packages **Vision-Language-Action (VLA)** and **Player Value Function (PVF)** insights as contributor “packets.” These packets sit alongside proprietary datasets, live match feeds, market odds, and analyst notes to form a unified, structured stream. #### **2. Quantitative Core** The core combines and reweights multi-source signals in real time. For each target, it: * Ingests contributor packets, including Score’s VLA and PVF outputs. * Fuses them with the SIRE LLM, using multiple passes to estimate fair value and potential edge. * Logs outcomes, scores contributor sources by live calibration and PnL tracking, and reweights on the fly. * Sizes positions using fractional Kelly with explicit edge thresholds.\ The LLM-first design searches large context windows to uncover independent signals for live, adaptive performance. #### **3. Autonomous Agent** Signals surface in the token-gated **aLink** terminal and execute through automated pools and **aVault**. Every decision is recorded from odds to settlement, which keeps performance verifiable and enables continuous improvement. #### Continuous data loop The engine ingests and recalculates every second, incorporating: * Real-time match events and visual streams * Player and team performance metrics * Historical patterns and contextual priors * Market prices, liquidity gaps, and risk parameters The result is a live stream of institutional-grade intelligence that bettors, partners, and automated vaults can use at scale. # Architecture Understand the technical infrastructure behind SIRE SIRE’s architecture runs as a sequence of interlocking processes, each one refining, weighting, and executing intelligence in real time. From intake to optimisation, every layer builds on the last, creating a continuous adaptive loop.
#### System Flow **1. \ Information Intake** Multi-source contributor packets enter the system, including Score Vision (Subnet 44) data, proprietary datasets, live match feeds, and market odds.\ Each packet contains unique predictive signals that seed the forecasting engine. **2. Model Processing** The SIRE LLM processes all packets, interpreting statistical and visual relationships, generating structured representations of match context, and aligning them with model priors. **3. Multiple Prediction Runs** The engine performs ensemble inference: multiple independent predictions across different packet combinations, to estimate fair value, edge, and volatility-adjusted confidence intervals. **4. Outcome Tracking** Every prediction is recorded against real-world outcomes. Market prices, settlement data, and live match results are logged to benchmark and recalibrate performance continuously. **5. Performance Scoring** Each contributor packet is evaluated using statistical metrics such as calibration, Sharpe ratio, and information value. Contributors are ranked based on predictive quality and long-term reliability. **6. Dynamic Weighting** High-performing contributors are up-weighted; underperforming or correlated ones are reduced or replaced. Sizing logic applies fractional Kelly scaling with adaptive exposure limits to maintain consistent risk profiles. **7. Iterative Optimisation \** The engine retrains on new data, updates weights, and redeploys refined models automatically. Sharper, verified signals flow to **aLink** for visibility and **aVault** for autonomous execution. #### Why It Works * Unified pipeline connecting data ingestion, modelling, and execution. * Continuous performance feedback ensures adaptive learning. * Modular structure allows new models, datasets, or signals to integrate without redesign. * Verifiable on-chain transparency at every step. Read more in [The Multi-Source Prediction Engine](/research-and-development/the-multi-source-prediction-engine) section. # Roadmap Key milestones and DeFAI integration plans SIRE’s roadmap follows a phased approach focused on building intelligent infrastructure, expanding the ecosystem, and integrating fully into DeFi. Each phase compounds on the last, moving the network closer to autonomous, data-driven market execution. #### **Phase 1 – Building the Infrastructure** **Objective:** Deploy core betting intelligence infrastructure combining computer-vision analytics with autonomous agent distribution. * **Agent Deployment:** Launch of the SIRE autonomous agent on social platforms, interacting directly with followers. * **Documentation & Community:** Release of full documentation outlining the vision, technology stack, and dual-wager system. * **aLink Launch:** Launch of the αLink Terminal for token holders to access predictions. * **Advanced Features:** Deployment of enhanced prediction models and risk-management systems. * **Mobile Web App:** Access SIRE anywhere through **sire.bot**, ensuring users never miss a time-critical edge. ✅ *Status: Completed.* #### Phase 2 – Expanding the ecosystem **Objective:** Scale across sports and deepen DeFi connectivity. * **aVault launch:** First aVault goes live. Users can add USDC to receive aVault units & earn potential uncorrelated returns, with automated on-chain strategy execution. * **New sports coverage:** aVault is now live and executing across multiple sports, with additional pipelines coming online. * **Benchmarking dashboard:** Interactive comparisons vs LLMs, elite tipsters, and major sportsbooks. * **Hedge-fund partnership:** $300M AUM deployment using SIRE’s intelligence engine; a share of performance fees supports SIRE buybacks. * **On-chain Exchanges & Prediction markets integrated into aLink:** aLink connects to integrated venues, showing live prices beside SIRE’s “true odds” and surfacing positive-EV or arbitrage opportunities users can act on instantly via a self-custodial wallet. * **Market-making DeltaVault:** Dedicated market-making vault on Kalshi, designed as a lower-volatility alternative that targets **potential uncorrelated rewards** through order-book liquidity strategies. 🔄 *Status: In Progress.* #### **Phase 3 – Full DeFi Integration & Autonomy** **Objective:** Advance SIRE into a self-optimising intelligence and execution network. * **Multi-Source Prediction Engine:**\ Extend the SIRE LLM with an optimisation layer that dynamically combines and weights multiple intelligence sources, including Score (Subnet 44) and its player evaluation models. This system continuously learns from outcomes to refine accuracy. [\[Read more → Multi-Source Prediction Engine\]](mailto:undefined) * **Cross-Venue Routing:**\ Aggregate liquidity across Kalshi, Polymarket, SX Bet, and future venues with smart order routing for best-price execution. * **Composable DeFi Integration:**\ Make aVaults ERC-4626-compatible and interoperable with Base, Aerodrome, and other DeFi protocols. 🕒 *Status: Upcoming.* ## Phase 4: \[REDACTED] ## Phase 5: \[REDACTED] # Overview #### **The Economic Engine Behind SIRE** SIRE’s token economy is built around transparency, fairness, alignment, and sustainability.\ Every action: staking, adding to aVault, or using aLink will flow back into the network through automated, on-chain mechanisms that reward participation and strengthen protocol resilience. #### **aVault System Flow**
Every protocol action feeds into staking rewards, treasury funding, and ongoing buybacks, creating a transparent, self-sustaining loop. #### **How It Works** **1. Participation → Protocol Activity** Users add USDC to aVault, stake SIRE to reduce fees. These activities generate on-chain performance and platform fees. **2. Fee Structure** * **Performance Fee:** 20% base, scaling down to 10% depending on SIRE staked. * **Withdrawal Fee:** 2% base, scaling down to 1% with higher staking tiers. * **Management Fee:** 2% annualised, split between treasury funding and SIRE burns. **3. On-Chain Distribution**\ All fees are distributed automatically: * 50% → Staking pool (potential rewards for stakers) * 30% → DAO treasury (ecosystem development) * 10% → dTAO purchase (infrastructure support) * 10% → SIRE buybacks (market sustainability) #### **Token Utility** * **Stake:** Participate in potential on-chain rewards from αVault performance. * **Access:** Unlock αLink tools and analytics. * **Reduce Fees:** Lower performance and withdrawal fees by staking SIRE. * **Buybacks:** A share of protocol fees funds ongoing buybacks to support network stability. #### **Transparency & Alignment** Every fee, burn, and buyback is verifiable on-chain.\ The more users participate, the stronger the feedback loop between SIRE’s utility, treasury growth, and long-term sustainability. # Token Utility & Distribution #### **Overview** SIRE’s tokenomics are designed for function first: every utility is built to work directly with the protocol’s core mechanics. Staking reduces fees and unlocks access, revenue share through potential staking rewards, and buybacks recycle protocol activity into sustainability. Each component scales naturally with adoption: as aVault activity, aLink usage, and on-chain volume grow, so does the token’s role within the system. This structure is paired with a commitment to transparency and fairness: **SIRE launched with no presale or insider allocation**; all circulating tokens were acquired openly on the market. Every transaction, fee, and buyback is recorded on-chain, ensuring that network growth and value distribution remain visible, verifiable, and shared by all participants. #### **Token Utility** * **Stake:** Access potential rewards funded by protocol activity and lower aVault fees through alignment with the protocol. * **Access:** Unlock aLink and its analytics suite for live model outputs, market insights, and performance data. * **Reduce Fees:** Lower aVault performance and withdrawal fees based on SIRE staked. * **Buybacks:** A share of protocol fees is continuously used for on-chain SIRE buybacks, reinforcing sustainability and alignment. #### **Token Distribution** The SIRE token embodies a fair, transparent distribution model designed for community ownership and long-term alignment. * **Community Power (85.48%)**\ Eighty-five percent of supply entered circulation on day one. Access and potential upside are fully in the hands of the SIRE community. * **Core Contributors & project development (14.52%)**\ The core team purchased its allocation on the open market. A portion of these tokens is allocated to ongoing project development, ecosystem growth, and contributor rewards. * **Strategic Treasury**\ There is no pre-minted war chest.\ The treasury accrues protocol performance fees and allocates them toward:\ • Product research and development\ • Partnerships and liquidity initiatives\ • Ecosystem expansion All funds are secured through a multi-signature wallet for transparent and responsible management. [\[Link to treasury SAFE multisig wallet\]](mailto:undefined) # Dual Wager Mechanism DKING features two distinct approaches to sports betting intelligence ## **Agentic Wagers** The foundation of our ecosystem, where participants stake $DKING and deposit USDC into pools managed entirely by our autonomous agent. Leveraging Score's computer vision and ML models, these pools execute quantitative strategies automatically, requiring no human intervention or sports knowledge. ## **Human/Partner Wagers** A complementary system allowing verified professionals, betting syndicates, and other AI agents to establish specialized pools. These operators must undergo a rigorous validation process and stake 100,000 $DKING tokens. This category creates opportunities for: * Professional bettors to scale their proven strategies * Betting syndicates to leverage Score's vision technology * Other AI agents to integrate their prediction models Together, these systems create a comprehensive ecosystem where sophisticated quantitative strategies meet professional expertise, all powered by advanced sports intelligence. # Human Wager How the human wager works The integration of professional bettors and syndicates into DKING follows a rigorous five-step validation process to maintain ecosystem integrity: 1. **Performance Verification** * Submit comprehensive betting history * Demonstrate consistent profitability * Provide detailed strategy documentation * Show risk management approach 2. **Model Validation** * Undergo thorough backtesting * Validate strategy robustness * Test compatibility with Score's vision data * Assess scalability potential 3. **Community Vote** * Present strategy to DAO members * Share performance metrics * Outline pool parameters * Obtain majority approval 4. **Token Lock** * Stake 100,000 $DKING tokens * Long-term alignment with protocol * Skin in the game requirement * Protocol security measure 5. **Fee Structure** * Propose performance fee structure * Maintain 2% management fee * Obtain community approval * Set pool parameters This process ensures that only qualified operators with proven track records can launch pools, maintaining DKING's high standards while offering token holders access to diverse, validated betting strategies. Through this selection mechanism, DKING creates a sustainable bridge between professional betting operations and decentralized finance. # αVault How SIRE's AlphaVault works #### **Overview** αVault automates SIRE’s intelligence layer on-chain. It transforms model-driven strategies into fully autonomous execution, allowing users to add USDC, receive αVault units, and participate in our directional strategies. When you enter αVault, you receive **vtSIRE units**, representing your vault share. These units remain in your wallet and are **non-transferable** for fairness and security. αVault is governed by strict safety rules, so you hold your vtSIRE Units and let the system trade and compound on your behalf. #### **Fee & Flow Overview**
This visual outlines the complete fee and reward structure: * Management and performance fees are distributed to the DAO treasury, staking pool, and buybacks. * Staking reduces both performance and withdrawal fees. * Every on-chain flow supports the long-term sustainability of the ecosystem. [*Refer to the Tokenomics section for the full network-wide breakdown.*](mailto:undefined) #### **αVault Units & High-Water Mark** αVault uses a **High-Water Mark (HWM)** system to ensure fairness: * Performance fees apply **only to new gains** above your personal HWM. * Your HWM starts at your entry unit price and updates each time a new high is reached (after fees). * If the unit price falls below your HWM, no performance fee is charged again until it exceeds that previous high. **Example:**\ You receive units at an entry price of **1.00 USDC** → HWM = 1.00.\ The price rises to **1.10 USDC**. A 20% performance fee applies to the $0.10 gain ( $0.02 ).\ Your net unit value becomes **1.08 USDC**, and your HWM updates to 1.08.\ If performance later falls to 1.05, no fee is charged again until the price climbs above 1.08. This structure prevents double-charging and ensures complete fee transparency. #### **Fee Structure** αVault fees are designed to align incentives between users and the protocol. * **Management Fee:** 2% annually (1% DAO Treasury / 1% SIRE burn) * **Performance Fee:** reduces from 20% → 10%, depending on SIRE staked * **Withdrawal Fee:** 2% → 1%, also reduced by staking level All fee flows occur automatically on-chain: funding staking rewards, SIRE buybacks, and ongoing product development. #### **Pool Mechanics** Anyone can add USDC to αVault and participate in automated, model-driven strategies as long as there is capacity in the vault. αVault is not token-gated. The vault follows strict quantitative risk rules to ensure disciplined execution: * **Max 1% exposure per trade** * **Max 12% of total bankroll deployed at once** These constraints prevent overexposure and promote long-term capital safety.\ \ Profits from successful trades are **streamed back into the vault over 24 hours**, creating a smoother return curve and preventing short-term arbitrage opportunities. \ \ On the [vault page](https://vault.sire.bot/buy-units), the **“pending”** amount represents profits from winning bets that are in the process of being streamed into the vault over the next 24 hours . This ensures transparent, time-distributed performance updates for all participants.
> *Note: αVault participation is currently capped for liquidity management and strategy calibration purposes. Caps may expand progressively as total value locked (TVL) and on-chain execution capacity scale.* #### Withdrawals To exit αVault, initiate a withdrawal request.\ Withdrawals are finalised after a **7-day period**, aligning with the vault’s profit-streaming cycle. After this window, you’ll receive your USDC equivalent withdrawal, **minus a small 2% withdrawal fee** (reduced if you have SIRE staked). This structure ensures transparent settlement, fairness across participants, and consistent vault balance updates during the streaming period. #### **Performance Results:** αVault operates with absolute transparency. Users can monitor the vault's performance via a complete feed of all settled bets. On the [Positions page](https://vaults.sire.bot/performance) of the vault, you can track the Total P\&L and view the specific details of every wager: * Game & Position * Stake & Payout * Individual P\&L & ROI #### **Community dashboard** For advanced analytics on αVault performance, on-chain trends & more, the[ SIRE Community Dashboard created](https://www.icm-analytics.com/sire-dashboard/) by [ICM Analytics ](https://x.com/icmanalytics) provides a comprehensive overview of aggregate performance and historical data. ### Why It Matters αVault is the bridge between sports intelligence and Decentralised Finance (DeFi). * Access: It converts institutional-grade data into autonomous strategies anyone can access. * Alignment: Transparent fees and buybacks ensure the protocol grows alongside user profits. * Uncorrelated Yield: By relying on sports outcomes rather than crypto market sentiment, αVault produces performance streams that help diversify on-chain portfolios. # αLink High-level overview of the SIRE AlphaLink
#### **Overview** αLink is SIRE’s intelligence interface: a token-gated terminal that lets holders interact directly with the network’s live models, uncover inefficiencies across markets, and understand the quantitative logic behind every decision. It combines the reasoning power of SIRE’s AI with real-time market data, turning natural-language conversation into actionable insight. αLink is being developed into **the sports-intelligence terminal for prediction markets**, integrating on-chain execution so users can identify and act on opportunities from a single interface. #### **Key Features** **Currently live:** * **Token-Gated Terminal:** Chat with SIRE to access model outputs, probability forecasts, and live analysis for upcoming events. * **+EV Edge Scanner:** Continuously sweeps major sportsbooks to flag mispriced markets and display potential positive-EV edges. * **Strategy Insights:** Learn the reasoning behind each model-driven recommendation in clear, interpretable language. * **Risk & Bankroll Tools:** Input your bankroll parameters and explore suggested bet sizing based on statistical confidence and risk appetite. **In Development:** * **Live Signals:** In-game predictions and tactical alerts as matches unfold. * **Educator Agent:** Interactive onboarding for new users, covering odds literacy, bankroll basics, and decision frameworks. * **Strategy Guides & Prompt Library:** One-click prompts for key betting concepts such as Value, Kelly, Arbitrage, and Middles. * **Prediction-Market Integrations:** Connect directly to Kalshi and SX Bet to analyse odds, detect arbitrage, and route orders from one interface via a self-custodial wallet. #### **Why It Matters** αLink transforms complex analytics into accessible, explainable intelligence. It is the connective tissue between AI-powered prediction and on-chain execution, enabling users to explore, learn, and act transparently within the same system. By integrating prediction-market data and execution rails, αLink is evolving into the **definitive sports-intelligence terminal for decentralised sports markets**. # Automated Betting Through our Agentic Wager Pool, participants can access these institutional-grade strategies Most sports bettors rely on intuition, but professional syndicates achieve 60%+ win rates by treating matches as pure mathematical equations. Our automated betting system brings this quantitative approach to everyone, combining computer vision with ML models to turn sports intelligence into profitable opportunities. Transform your USDC into quantitative strategies powered by advanced sports intelligence. ## **How It Works** 1. **Entry** * Buy and stake DKING tokens * $1 worth of DKING token unlocks $10 USDC deposit capacity * Deposit USDC into the Agentic Wager Pool 2. **Strategy Execution** * Automated betting strategies process 98% of deposits * Computer vision analysis informs position sizing * Risk management systems protect capital 3. **Fee Structure** * 2% Management Fee * 1% to DAO Treasury * 1% to DKING Token Burn * 20% Performance Fee * 80% to Pool Participants * 10% to TAO/qTAO * 10% to DAO Treasury 4. **Returns** * 80% of profits distributed to participants * Automated withdrawals available * Un-staking DKING removes pool access Through this automated system, participants can access sophisticated sports betting strategies without requiring technical knowledge or manual execution. # Staking Learn how staking works and its benefits #### **Overview** Staking connects holders directly to SIRE’s network activity. By staking SIRE, users help stabilise the ecosystem and participate in a fee-based incentive structure that rewards long-term alignment. Stakers receive protocol benefits including reduced αVault fees, access to the token-gated αLink LLM, and eligibility for ecosystem-based rewards as new integrations come online. #### **Mechanics** When you stake SIRE, your tokens are locked in a non-custodial contract that tracks your tier level and unlocks protocol benefits in real time. **Staking Benefits:** * **Reduced αVault Fees:** Lower both performance and withdrawal fees, scaled by the amount of SIRE staked. * **Staking Rewards:** A portion of protocol fees from αVault and future network products is distributed to stakers. * **αLink access:** Stakers unlock the token-gated terminal with live models and market analysis tools. * **Ecosystem Rewards:** Stakers are eligible to receive additional benefits tied to broader ecosystem participation, including integrations with prediction-market platforms and on-chain initiatives aligned with SIRE’s mission. **Staking Tiers:** * 1K SIRE → 18% performance fee | 1.8% withdrawal fee * 5K SIRE → 16% | 1.6% * 10K SIRE → 14% | 1.4% * 50K SIRE → 12% | 1.2% * 100K SIRE → 10% | 1% All staking and fee adjustments occur automatically on-chain. Stakers maintain full self-custody and can unstake at any time, subject to cooldown periods when applicable. #### **Why It Matters** Staking aligns participants with the long-term success of the protocol. It strengthens the feedback loop between αVault, αLink, and the wider SIRE ecosystem, ensuring contributors benefit from transparency and shared growth. As new integrations launch, staking becomes the access layer linking holders to the expanding network of tools, incentives, and market applications built around SIRE. #### **How It All Connects** SIRE operates as a continuous loop between intelligence, execution, and alignment. αLink surfaces live analytical insight from sports data and prediction markets. αVault turns those insights into autonomous, on-chain strategies that generate measurable performance. Staking completes the cycle: rewarding participation, reducing fees, and feeding value back into the network. Together, these elements create a self-learning & compounding ecosystem where every match, signal, and transaction strengthens SIRE’s intelligence engine and supports its long-term sustainability. # SIRE Intelligence: Data Science Report How In-House Modelling, LLMs and a community of Machine Learning Engineers Are Powering a New Era of Sustainable +EV Betting ### Date Science Report **1. Summary:** Over the past season, Score’s Data Science team was given a clear mandate: develop a framework we could expand into a sustainable, scalable betting strategy. From day one, the end goal was a Multi-Source Prediction Engine, a self-optimising system capable of learning from multiple contributors and adapting in real time, which would power our LLM approach. To measure the impact of such an approach, we developed robust baseline models at first because they would: * Quantify the predictive edge of traditional statistical models. * Serve as benchmarks for both our in-house Large Language Models (LLM) and models developed by the Machine Learning (ML) community. * Provide feature inputs the LLM could draw on if it determined they were valuable.
Building these market models taught us that historical simulations have severe limitations. They can overfit to past data, fail to capture changing market dynamics, and create a false sense of stability. In today’s sports markets, where odds are efficiently priced and heavily influenced by bookmakers, traditional models often end up being little more than noise. The real edge lies elsewhere. Our new approach leverages LLM to uncover orthogonal signals, hidden within massive context windows of data. This isn’t just an incremental improvement, it is a fundamentally new way of extracting value, designed for live, adaptive performance. As Score Co-Founder and “[MicroPrediction - Building an Open AI Network](https://mitpress.mit.edu/9780262047326/microprediction/)” author, Peter Cotton puts it: “The LLMs we have now, we didn’t have six months ago. Statistics can be a form of reasoning, but it’s not the only one at our disposal.” With those lessons learned, we moved from static baselines to real-time LLM predictions powered by the Multi-Source Prediction Engine. ### 2. The Modelling Foundations There are four different in-house models developed to predict football match outcomes. (1X2) 1. #### Elo The Elo model was initially developed to calculate the skill levels of players in Chess. Each player’s elo will either increase or decrease based on the outcome of the match. In our implementation, we estimate the probability of each outcome based on the rating difference between home and away teams. 2. #### Davidson The Davidson model extends the Bradley-Terry framework by incorporating the possibility of a draw in pairwise comparisons. Each team is assigned a strength, and match outcomes are modelled probabilistically based on the relative strengths of the teams. These parameters are fitted using maximum likelihood: * Team strength * Home advantage * Draw parameter
3. #### Dixon-Coles The Dixon-Coles model is designed to model the number of goals scored by each team. It extends the Poisson regression framework by introducing a correction factor for low-scoring games and accounting for the dependency between teams' goals. In our implementation, we estimate the attacking and defensive strengths of each team to derive outcome probabilities from predicted goal distributions. 4. #### Sarmanov The Sarmanov Negative Binomial model is designed to model the number of goals scored by each team using a bivariate negative binomial distribution. It extends the standard framework by allowing for overdispersion in goal counts and introducing a dependency structure between the two teams’ scores through the Sarmanov distribution. In our implementation, we estimate attacking and defensive strengths along with a correlation parameter to generate outcome probabilities from the joint goal distribution. ### 3. Building Meta Models With the four individual models performing well, we moved toward ensemble modelling, combining multiple models to improve predictive accuracy. 1. #### Meta Pairwise The first meta model developed is meta pairwise, which combines the outputs from Elo model, Davidson model, and implied probability from bookmaker odds.
To measure the performance of the Meta Model, we tested the model on bankroll growth using match data from the top 5 leagues (Premier League, La Liga, Serie A, Ligue 1, and Bundesliga) across 2 seasons (23/24 and 24/25). The line chart below illustrates the bankroll growth for five different models and strategies.
1. ChatGPT (Flat Lines) 2. Sportsmonks (Flat Lines) 3. Agentic Model (Flat Lines) 4. Sportsmonks (+ Kelly Criterion) 5. The Meta Model (Agentic Execution)
As illustrated in the chart, both the ChatGPT model and the Sportsmonks model (based on a premium paid football data subscription) experienced early bankroll depletion, while the Agentic Model went bust near the end of the 2023/24 season. However, when bankroll management was applied using a fractional Kelly Criterion combined with an edge threshold, the Sportsmonks model lowered its loss to approximately 10% over two seasons. In contrast, the Meta Model achieved a solid return of +32.96% over the same period. This highlights the critical role of bankroll management in long-term betting sustainability. Additionally, the Meta Model consistently outperformed the Sportsmonks approach, further demonstrating the effectiveness and added value of our model. 2. #### Meta Logistic Meta Logistic combines the outputs from Elo, Davidson, Dixon Coles, and Sarmanov models, alongside implied probability from bookmaker odds and historical team statistics to predict the outcome of a football match.

After backtesting our model from 19/20 to 24/25 season, considering different Kelly Fraction and edge thresholds, we have the ROI illustrated in this grid search plot.
The best parameters returned 370.3% on betting draw outcomes. This improvement is significant compared to the first meta model, as we introduce more models and features to the ensemble model. Although a 370.3% return on paper seems impressive, plotting the bankroll growth reveals that the model's performance is unstable, likely due to the aggressive nature of Kelly betting, despite using a 30% Kelly fraction.
Key takeaways: Although one strategy in our Meta Logistic backtest yielded +370%, the bankroll growth was unstable, indicating that the ROI is not sustainable when deployed live without careful adjustment. ### 4. Expanding to LLMs In today’s fast-paced sports prediction industry, winning means adapting in real time. That’s why we’ve embraced a more ambitious, next-generation approach with Large Language Models (LLMs), moving beyond the constraints of traditional approaches to harness live performance data, accelerate learning, and enable autonomous improvement. This is more than a technical upgrade; it’s a strategic transformation designed to put SIRE LLM at the forefront of AI-driven sports forecasting. The following pillars outline the foundation of this upgrade: **Moving Beyond Backtesting** Traditional backtesting often suffers from overfitting and fails to capture real-world adaptability. Past performance alone doesn’t guarantee future results, so instead of relying on historical simulations, we are shifting our focus toward live performance evaluation, measuring SIRE LLM’s predictions in real-time, where it truly matters.
**Proven Real-World Results** [During the Club World Cup](https://x.com/thedkingdao/status/1945185360381608062), our LLM-powered SIRE terminal delivered outstanding performance, achieving over 3x bankroll growth. These results underscore the model’s capacity to adapt, learn, and thrive under real competitive conditions. **A Vision for Autonomous Model Evolution** We are entering an era where neural networks are trained, evaluated, analyzed, and iterated without direct human intervention. This is the path toward Artificial Superintelligence (ASI). Our goal for SIRE LLM is to learn from its past predictions, autonomously refine itself, and deliver increasingly precise forecasts.
**Strategic Role of the Data Science Team** Our in-house Data Science team will focus exclusively on advancing SIRE LLM’s architecture and performance. Meanwhile, our community of sports machine learning (ML) experts will focus on modelling and backtesting to uncover +EV (positive expected value) features and models for integration into the LLM.
**Collaborative Advantage** With improved modelling from our community of sports ML experts and a streamlined focus on the SIRE LLM, we will generate more accurate predictions. These predictions will feed into a meta-model responsible for optimal bankroll management and bet sizing, ensuring that every insight from the LLM is translated into maximum strategic advantage.
**The Future Is Now** By combining cutting-edge live evaluation with autonomous improvement, SIRE LLM isn’t just keeping pace with the future of AI in sports prediction — it’s leading the charge.
\
# The Multi-Source Prediction Engine With the SIRE LLM now delivering live, adaptive predictions, our next phase is to extend its capabilities with the Multi-Source Prediction Engine, a planned optimisation layer designed to combine and dynamically weight insights from multiple contributors in real time. Score (Subnet 44) will be one of the main contributors, with our Vision-language-action (VLA) and Player Value Function (PVF) models. These specialised models will provide high-resolution visual insights and player-level evaluations, adding unique dimensions to the prediction process.\ \ Our new approach leverages LLM to uncover orthogonal signals, hidden within massive context windows of data. This isn’t just an incremental improvement, it is a fundamentally new way of extracting value, designed for live, adaptive performance. As Score Co-Founder and “[MicroPrediction - Building an Open AI Network](https://mitpress.mit.edu/9780262047326/microprediction/)” author, Peter Cotton puts it: “The LLMs we have now, we didn’t have six months ago. Statistics can be a form of reasoning, but it’s not the only one at our disposal.” #### Planned Design 1. Information Intake: For each prediction target, the engine will open several slots.\ Each slot will be filled by an information packet from a contributor, containing signals relevant to the future outcome. Contributors can be models (including VLA and PVS), human analysts, or other data-driven systems.
2. Model Processing: The SIRE LLM will process all packets, interpreting both the individual signals and the relationships between them to produce forecasts.
3. Multiple Prediction Runs: Instead of a single forecast, the system will generate several independent predictions per target, with each run using a randomly selected sample of slots.
4. Outcome Tracking: Once the real-world result is known, the system will measure how much each packet contributed to predictive success.
5. Performance Scoring: Packets will be ranked based on historical profit and loss (PnL) providing a domain-agnostic measure of their value.
6. Dynamic Weighting: Packets with strong PnL histories will be up-weighted in future predictions. Weaker packets will be down-weighted or replaced over time.
7. Iterative Optimisation: This process will repeat for every prediction target, continuously refining the mix of contributors and improving overall forecast accuracy. #### Why It Matters * Combines diverse intelligence sources in a single, adaptive framework. * Continuously self-improves based on live market outcomes. * Maximises long-term +EV by prioritising the most profitable contributors. * Scalable and future-proof, able to integrate new models, data feeds, or human experts without changing the core system. The Multi-Source Prediction Engine will be the next leap forward for SIRE, moving from a single adaptive LLM to a networked, self-optimising intelligence system designed to stay ahead of increasingly efficient betting markets.