Lobby Theory / Privacy notice
Privacy notice
This website and the intended app are different. This notice explains what this preview currently does, then describes the production design. The app is unreleased. Production processing providers, permissions, retention arrangements, and notices are not finalized.
1. Operator and scope
Lobby Theory is developed and operated by Johnny Antwiler. This notice applies to this product-preview website and communications sent to the listed contact addresses. The application-design sections explain intentions; they do not activate collection or substitute for a production privacy notice.
2. What the website does today
The site serves HTML, styles, scripts, fonts, images, and an illustrative 13-second video. Those assets are hosted with the site. The video is requested when playback starts; desktop autoplay may initiate that request. Reduced-motion, mobile/tablet, and Save-Data visitors initially receive a still and may choose to play.
The three-view demonstration uses fixed illustrative shop offers, fictional role-based guide examples, a synthetic completed-match review, and editable experimental builds based on the hero’s patch-labeled roster. The site does not connect to a Riot account, inspect the running game, read the real shop, collect gameplay, upload game files, run an AI model, accept payment, or enroll anyone in research, testing, or a waitlist.
Switching views, choosing a guide, editing a draft, saving a demo build, and trying consent settings update page memory. A demo save is not durable storage: reloading or leaving the page clears it. The deletion button displays explanatory text; it does not send a request or delete evidence. No consent is recorded by these controls.
Requests, cookies, and storage
The implemented site code requests its page assets from the same site and does not add analytics, advertising scripts, tracking pixels, cookies, localStorage, sessionStorage, IndexedDB, or a service worker. It makes no gameplay or demo-data API requests. Font and artwork files are local site assets, not runtime requests to Google Fonts or Riot’s CDN.
The hosting provider may process IP addresses and browser/request details to deliver and secure the website. If access protection is enabled for a particular preview, its access provider may also process authentication information. Provider notices describe those services; this audit of the site’s code does not establish provider-side cookies or retention.
Links to other websites request those destinations when you open them. Email links open your configured email application; this site does not itself send your message. There is no contact form or file-upload control.
3. Email inquiries
If you email Lobby Theory, the operator and email delivery services process your address, message, and any attachments for responding to the inquiry and managing the correspondence. Share only what is needed. Do not send passwords, authentication/session material, API keys, or unnecessary personal information. A fixed retention schedule for inquiries has not been adopted; ask about access or deletion using the privacy contact below.
4. Application design: separate purposes and choices
These sections describe the app in development. This website performs none of this processing. Buying coaching, using the overlay, and contributing to research are separate decisions.
| Purpose | Choice and boundary |
|---|---|
| Free overlay and planning | Show static reference information and the guide the player selects. Local processing supports guide-category shop matching and factual readouts of your own gold, board value, and gold to the next level. The live readouts do not tell you when to roll or level or compare you with the lobby. Using these features does not automatically create a persistent match recording or cloud upload. Players enter planning inputs and manually switch guides. |
| Recording for personal review | A separate affirmative choice before gameplay enables persistent recording within the authorized field list for the player’s own completed-match review. It is distinct from purchasing premium, requesting cloud processing, or joining research. |
| Requested coaching | Process only evidence necessary for the review requested, after the entire match ends. Any cloud processing would send allowlisted derived evidence to Lobby Theory and disclosed providers. Opening an already saved review does not itself require another evidence upload. |
| Saved history and builds | Support revisiting custom builds and completed-match reviews. Production account requirements, local versus cloud storage, and retention remain to be decided. This preview saves only a demo build in page memory. |
| Research contribution | A separate opt-in, off by default, for free and paid users to record and contribute authorized evidence for research and evaluation. Neither a coaching purchase nor a coaching request is required. Once enabled in production, it may cover later games until disabled, as the eventual notice explains. |
| Training Lobby Theory’s own models | An additional opt-in, also off by default, with its own explanation. Research consent alone does not authorize training, including training models that may support paid coaching. In the demonstration, training requires research and a separate manual choice; enabling research never enables training. |
5. Evidence, data sources, and local processing
The planned distribution channel is Overwolf. Its Game Events Provider (GEP) is the planned structured source, alongside static game data, player-entered plans, and authorized official completed-match information. Overwolf’s TFT documentation lists player gold, level and XP, board pieces, and shop events. Technical availability does not establish permission for every use or guarantee a complete decision history.
When separately enabled, recording would concern the player’s board and positions, units and equipped items, bench, gold and economy, health, level and XP, rounds, shop offers, and available combat statistics. These are candidate fields, not a claim that each source can provide each field. Augment collection and analysis remain outside the initial scope pending clarification. Any opposing-board evidence would require specific permission for post-game collection and use; it must not enable live opponent scouting.
Raw source artifacts stay on the player’s PC in the production design. Any uploaded derived evidence must be specifically allowlisted and redacted. Raw artifacts can contain account identifiers, authentication-like material, machine context, and other information that should not leave the device. A field’s presence in a local artifact is not permission to upload it.
Redacted does not mean anonymous. Game identifiers, timestamps, account associations, other players’ information, and technical metadata can make evidence linkable even after names are removed. The app must disclose the actual uploaded fields and purposes before processing. Player consent does not replace Riot or platform permission.
Theorycrafting uses planning inputs entered by the player. The generator is not automatically fed the live board to produce in-match prescriptions. Any server processing of those planning inputs would require disclosure before that feature launches.
6. Coaching, research, and model processing
Coaching concerns a requested personal review. Research may evaluate supported relationships across contributed evidence, such as decisions, item interactions, combat outcomes, and positioning. Training Lobby Theory’s own models is a distinct use requiring the additional choice. A review must distinguish observations from inference and preserve uncertainty where evidence is missing.
No production AI vendor has been selected. How analysis will be divided among the PC, Lobby Theory’s service, and any provider remains open. The provider requirements do not authorize outside providers to train their own models on player evidence; the actual contracts, settings, subprocessors, and disclosures must be settled before processing begins. No existing vendor protection or production research approval is claimed.
7. Stopping contribution, withdrawing choices, and deletion
The designed controls allow users to stop future research contributions and withdraw permission for future training use. In this demonstration, turning research off also turns training off; turning research on again leaves training off. Personal recording and requested-coaching choices remain separate.
Stopping future contribution does not itself delete earlier evidence. A separate process will handle requests concerning prior research contributions, personal history, and other information held. The demo does not implement that process.
Before production, the handling of existing datasets, completed analyses, trained models, provider copies, and backups must be defined and disclosed. Deletion may have a different scope from stopping future collection or use. No model-unlearning guarantee is made. For a real inquiry about information already held, email the privacy address; we may need information reasonably necessary to identify the relevant correspondence and verify the requester.
8. Advertising, accounts, and operational information
Advertising arrangements remain undecided. No advertising SDK is integrated in this preview. Any future advertising and purchase prompts would need to comply with applicable platform rules; the design does not place ads in Riot’s in-game, loading-screen, or client surfaces.
Production account requirements are not finalized. If approved account linking is used, the design calls for Riot-hosted sign-in; this website requests no Riot credentials. Subscription, security, billing, review-usage, diagnostic, and other necessary operational records would need separate disclosure. No production behavior-analytics system is implemented here.
9. Decisions required before production
No production retention schedule has been selected. Separate rules are needed for raw local artifacts, transient review inputs, saved history and outputs, research records, training uses, operational records, provider-held copies, and backups. Storage locations, international processing, provider selection, security controls, access procedures, and deletion/withdrawal arrangements also remain unresolved.
Research availability depends on the necessary permissions for collection, upload, central storage, research and any model uses, together with appropriate disclosures and working controls. If those conditions are not ready, research remains inactive. A production privacy notice must accurately describe adopted and implemented practices before the relevant processing starts.
10. Changes and external services
This notice may be updated as the website changes or production practices are adopted. The date above identifies this version. External services and linked sites have their own terms and privacy practices. Material production processing changes must be explained before they are activated; the preview’s example switches are not advance consent to future practices.
11. Contact
Operator: Johnny Antwiler.
Privacy questions or requests: privacy@lobbytheory.com.
Product and support inquiries: support@lobbytheory.com.
Return to the data-choice demonstration · Read website terms