[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"doc-detail-84918-en":3,"doc-seo-84918-105":29,"detail-sidebar-cat-0-en-105":91},{"code":4,"msg":5,"data":6},0,"success",{"doc_id":7,"user_id":8,"nickname":9,"user_avatar":10,"doc_module":4,"category_id":11,"category_name":12,"doc_title":13,"doc_description":14,"doc_content":15,"file_id":16,"file_url":17,"file_type":18,"file_size":19,"view_count":20,"is_deleted":4,"is_public":20,"is_downloadable":20,"audit_status":20,"page_count":21,"language":22,"language_code":23,"site_id":24,"html_lang":23,"table_of_contents":25,"faqs":26,"seo_title":13,"seo_description":14,"update_tm":27,"read_time":28},84918,1099514068035,"Ezra","https://ap-avatar.wpscdn.com/davatar_276721f389ce27ea32af1340a28f341c",8,"Research & Report","Pitwall Faithful Natural-Language Race-Strategy Briefings from a Calibrated Real-Time Monte Carlo Engine","Live sports commentary must generate deadline-bound statements about named athletes while the grounding state changes every few seconds and no reference text exists at generation time. Pitwall is a production system that produces natural-language Formula 1 strategy briefings in English, Spanish, and Portuguese. Faithfulness is enforced architecturally: each sentence is decomposed into typed factual claims and verified against a calibrated vectorized Monte Carlo race-state engine. A verifier filters fine-tuning data, retaining only fully state-supported targets and falling back to faithful templates. The approach was validated in live 2026 Grands Prix.","arXiv :2607 .06495v 1 [ cs .CL] 7 Jul 2026  \nPitwall: Faithful Natural-Language Race-Strategy  \nBriefings  \nfrom a Calibrated Real-Time Monte Carlo Engine  \nJuan S. Santillana  \nIndependent Researcher  \n[juan@jsantillana. com](juan@jsantillana. com)  \nAbstract  \nLive sports commentary is grounded generation under a deadline: statements concern real, named athletes, the grounding state changes every few seconds, and no reference text exists at generation time. We present Pitwall, a production system that generates natural-language Formula 1 strategy briefings in English, Spanish, and Portuguese, and that treats faithfulness asan architectural property rather than an aspiration: every published sentence is decomposed into typed factual claims (positions, gaps, tyres, pace, overtakes, race control) and each claim is verified against the probabilistic race state that prompted it. The same verifier gates the fine-tuning data—of 3,045 model-written targets, only the 81 .9% whose every claim is statesupported are retained; the rest fall back to a provably faithful template—so the generator is trained never to have seen an ungrounded target. What makes verification meaningful is the grounding substrate: a vectorized Monte Carlo engine (N =2 ,000 per-lap race continuations) whose probabilities are calibrated on 126 races (2018–2024) and validated on fully held-out 2025–2026 seasons (winner-in-top-3 90 .3% over 155 backtests; held-out Brier 0 .0745) . A recurring finding spans both halves of the system: virtues trade off and must be gated separately. In simulation, calibration-optimal is not decision-optimal; in generation, fine-tuning on richer targets buys vividness that collapses into hallucination precisely when the grounding state is sparse—a collapse that a four-base replication traces to the base model’s instruction adherence rather than to scale, and that base-model selection under a sparse-context audit removes in production. End-to-end operation—live timing to verified trilingual briefings—was confirmed at two consecutive live Grands Prix (Austria and Britain, 2026); at Silverstone a timestamped probability trace, committed to disk before the outcome was known, locked onto the eventual winner ten laps before the flag.  \n1 Introduction  \nA modern Formula 1 race is decided as much on the pit wall as on the track. A pit stop costs roughly 20–25s of race time, tyre compounds degrade non-linearly and at circuit-specific rates, overtaking difficulty varies by an order of magnitude between venues, and Safety-Car interruptions arrive stochastically and reprice every open strategic option within a single lap. The teams’ strategists must therefore answer questions of the form “if we box now, what is the probability we emerge ahead of the car we are racing—and how does that change if we wait two laps?” in well under a minute, continuously, for two hours. Every top team operates simulation tooling for this purpose, but that tooling is proprietary, its calibration has never been publicly audited, and the academic literature on race strategy simulation [Bekker and Lotz, 2009, Heilmeier et al., 2020a] stops short of the real-time, probability-calibrated, decision-focused setting in which the problem is actually solved.  \nThe answers, moreover, are consumed as language. Engineers, broadcasters, and audiences do not read probability vectors; they read sentences—“Norris pits and rejoins P4, 2.1s behind Sainzon mediums”—and a generated sentence about a real, named athlete that is confidently wrong is worse than no sentence at all. Live commentary is therefore a demanding instance of grounded generation [Wiseman et al., 2017, Ji et al. , 2023]: the grounding state mutates every few seconds, no reference text exists at generation time (ruling out reference-based evaluation entirely), and the sparse moments—early laps, partial state after an ingestion hiccup—are exactly the moments a fluent generator is most tempted to fill from its prior.  \nThis p","cbCain3V7sQNVkGd","https://ap.wps.com/l/cbCain3V7sQNVkGd","pdf",408956,1,21,"English","en",105,"# Introduction\n## Grounded live natural-language race briefings\n## Pitwall system overview and faithfulness verification","[{\"question\":\"What is the core challenge Pitwall addresses in live Formula 1 commentary?\",\"answer\":\"Generating natural-language strategy statements under a deadline when the grounding state updates every few seconds and no reference text is available at generation time.\"},{\"question\":\"How does Pitwall ensure faithfulness of its generated sentences?\",\"answer\":\"Each published sentence is decomposed into typed factual claims and verified against the probabilistic race state that prompted it, and the same verifier gates fine-tuning targets.\"},{\"question\":\"What grounding engine does Pitwall use, and how was it validated?\",\"answer\":\"A vectorized Monte Carlo engine with probabilities calibrated on 126 races from 2018–2024 and validated on held-out 2025–2026 seasons using backtests and Brier scoring.\"}]",1784199339,53,{"code":4,"msg":30,"data":31},"ok",{"site_id":24,"language":23,"slug":32,"title":13,"keywords":33,"description":14,"schema_data":34,"social_meta":86,"head_meta":88,"extra_data":90,"updated_unix":27},"pitwall-faithful-natural-language-race-strategy-briefings-from-a-calibrated-real-time-monte-carlo-engine","",{"@graph":35,"@context":85},[36,53,68],{"@type":37,"itemListElement":38},"BreadcrumbList",[39,43,47,50],{"item":40,"name":41,"@type":42,"position":20},"https://docshare.wps.com","Home","ListItem",{"item":44,"name":45,"@type":42,"position":46},"https://docshare.wps.com/document/","Document",2,{"item":48,"name":12,"@type":42,"position":49},"https://docshare.wps.com/document/research-report/",3,{"item":51,"name":13,"@type":42,"position":52},"https://docshare.wps.com/document/pitwall-faithful-natural-language-race-strategy-briefings-from-a-calibrated-real-time-monte-carlo-engine/84918/",4,{"url":51,"name":13,"@type":54,"author":55,"headline":13,"publisher":57,"fileFormat":60,"inLanguage":23,"description":14,"dateModified":61,"datePublished":62,"encodingFormat":60,"isAccessibleForFree":63,"interactionStatistic":64},"DigitalDocument",{"name":9,"@type":56},"Person",{"url":40,"name":58,"@type":59},"DocShare","Organization","application/pdf","2026-07-17","2026-07-16",true,{"@type":65,"interactionType":66,"userInteractionCount":20},"InteractionCounter",{"@type":67},"ViewAction",{"@type":69,"mainEntity":70},"FAQPage",[71,77,81],{"name":72,"@type":73,"acceptedAnswer":74},"What is the core challenge Pitwall addresses in live Formula 1 commentary?","Question",{"text":75,"@type":76},"Generating natural-language strategy statements under a deadline when the grounding state updates every few seconds and no reference text is available at generation time.","Answer",{"name":78,"@type":73,"acceptedAnswer":79},"How does Pitwall ensure faithfulness of its generated sentences?",{"text":80,"@type":76},"Each published sentence is decomposed into typed factual claims and verified against the probabilistic race state that prompted it, and the same verifier gates fine-tuning targets.",{"name":82,"@type":73,"acceptedAnswer":83},"What grounding engine does Pitwall use, and how was it validated?",{"text":84,"@type":76},"A vectorized Monte Carlo engine with probabilities calibrated on 126 races from 2018–2024 and validated on held-out 2025–2026 seasons using backtests and Brier scoring.","https://schema.org",{"og:url":51,"og:type":87,"og:title":13,"og:site_name":58,"og:description":14},"article",{"robots":89,"canonical":51},"index,follow",{"doc_id":7,"site_id":24},{"code":4,"msg":5,"data":92},[93,97,101,105,110,115,120,123,128,131,135],{"id":20,"doc_module":4,"doc_module_name":45,"category_name":94,"show_sort_weight":95,"slug":96},"Story & Novel",90,"story-novel",{"id":46,"doc_module":4,"doc_module_name":45,"category_name":98,"show_sort_weight":99,"slug":100},"Literature",80,"literature",{"id":52,"doc_module":4,"doc_module_name":45,"category_name":102,"show_sort_weight":103,"slug":104},"Exam",70,"exam",{"id":106,"doc_module":4,"doc_module_name":45,"category_name":107,"show_sort_weight":108,"slug":109},5,"Comic",60,"comic",{"id":111,"doc_module":4,"doc_module_name":45,"category_name":112,"show_sort_weight":113,"slug":114},6,"Technology",50,"technology",{"id":116,"doc_module":4,"doc_module_name":45,"category_name":117,"show_sort_weight":118,"slug":119},7,"Healthcare",40,"healthcare",{"id":11,"doc_module":4,"doc_module_name":45,"category_name":12,"show_sort_weight":121,"slug":122},30,"research-report",{"id":124,"doc_module":4,"doc_module_name":45,"category_name":125,"show_sort_weight":126,"slug":127},9,"Religion & Spirituality",20,"religion-spirituality",{"id":126,"doc_module":4,"doc_module_name":45,"category_name":129,"show_sort_weight":126,"slug":130},"World Cup","world-cup",{"id":132,"doc_module":4,"doc_module_name":45,"category_name":133,"show_sort_weight":132,"slug":134},10,"Lifestyle","lifestyle",{"id":136,"doc_module":4,"doc_module_name":45,"category_name":137,"show_sort_weight":106,"slug":138},19,"General","general"]