Website Visual Archive / electionbettingodds
Election Betting Odds takes prediction-market prices, bid and ask ranges, stake totals, and state-level data, then makes them read as a single chance. This report preserves the page as a dated static study. The mirror contains no live odds, forecast service, or betting function.
Selected as a WVA v2 seed specimen because the page exposes its calculation claim in the interface itself. This is a 22-route public capture window from July 19, 2026: 21 source pages returned HTTP 200, and one linked chart-notes destination returned HTTP 404. The exact static mirror lives under /preview/electionbettingodds/.
The archive notice stays visible in the mirror because the source describes frequent updates. Candidate tooltips and state-map hover survive locally. The source chart visuals are frozen images, range buttons are disabled, and only captured chart destinations remain active.
The route set begins with the homepage navigation and time-range controls, then includes chart destinations and the one directly linked 2025 nominations page found during local link QA. Twenty-one captured source pages render locally. The source chart-notes link returned 404, so its mirror destination explains that failure instead of sending a reader back to the live site.
Source images and the four map libraries are local. The build removes the five-minute reload, analytics, social widgets, StatCounter markup, and the newsletter iframe. Google Charts is replaced by locally stored rendered images. Uncaptured entries in the source chart selector are disabled, which keeps a static archive from implying that it holds the full historical catalog.
The home page is one vertical dashboard built from centered tables and fixed 700px regions. A boxed navigation block sits above three simultaneous readings: Senate control on the left, a state map in the middle, and presidential candidates below the first screen. The page gives each comparison a literal place rather than collapsing it into a card system. On a narrow viewport, that structure retains a horizontal overflow instead of becoming a new mobile layout.
Times carries almost every visible decision. The title and percentages run at sizes that would be excessive in a modern dashboard, yet the contrast makes the source easy to scan: names and market labels stay small, percentages become the reading anchors, and italic section titles mark each new question.
The page uses a white browser ground, black text, blue and red party colors, gray map states, and green or red directional arrows. Color indicates party, state status, and recent movement. The material character comes from default links, borders, tables, and raster portraits rather than a branded interface layer.
The page alternates between large percentage displays, small market-detail tables, and one wide map or chart. That coarse rhythm gives a dense data source a readable sequence: question, chance, evidence, then a broader geographic or historical view. Negative margins and inline table rules tighten individual clusters, which gives the page a compressed printed-news rhythm.
The source has no decorative motion. Candidate portraits reveal market ranges and stakes through CSS hover, while the Raphael map changes the selected-state and market panels on hover. The source also reloads the full page every five minutes. The mirror retains the local hover states and removes the reload loop.
The source has no viewport meta tag or responsive breakpoint. At 390px, its fixed 700px regions produce 340px of horizontal overflow. The mirror adds a viewport declaration so the archive notice remains legible, then preserves the fixed source surface and its horizontal overflow rather than redesigning it for a narrow screen.
The source uses an HTML 3.2 doctype, inline styles, tables, font tags, and a flat root-level route structure. The mirror keeps the captured markup and routes while adding noindex metadata, a dated archive notice, and local path rewriting.
jQuery, Raphael, jquery.usmap, and the source color helper are retained locally because the map and its hover detail are central evidence objects. The mirror has no source-origin map request.
Sixteen Google-chart surfaces were captured as images during the public-source window. The remote Google runtime is absent. The frozen chart image keeps the visible historical line and range slider; the source range buttons are disabled, and only captured chart routes remain selectable.
The build removes analytics, social widgets, StatCounter, the newsletter iframe, and all external scripts. The newsletter location becomes a small archive note. Editorial and market links remain ordinary outbound anchors.
The chart About control points to a same-origin local page that names the source-side 404. This preserves the observed failure without routing readers to an unverified live destination.
The page asks a reader to treat a large percentage as a useful synthesis of several prediction markets. "Election Betting Odds" alone leaves the method opaque: a reader still needs to know whose prices enter the number, whether the number represents a bid, an ask, a midpoint, a wager total, or a calculation across markets, and how a national number relates to individual state races. The page answers those questions through nested evidence. The large percentage creates a first read, while the candidate-tooltip table exposes each named market, its bid-ask range, and the amount wagered. The state map turns a statewide collection into a geographic index, and the chart turns a current-looking percentage into a time series. The candidate-tooltip table carries the most explanatory weight because it connects the headline number to a visible set of source-market inputs. For WVA, the transferable move is structural: place the conclusion first, then make the proof object one hover or one scroll away. The method carries without borrowing the source's Times typography, party palette, or old-web surface.
The page uses blunt visual means to make the aggregation feel direct. A white field, oversized Times percentages, underlined labels, and default-blue links make the interface resemble a public worksheet more than a trading terminal. That register reduces the gap between a market statistic and a reader who wants a plain answer. Red and blue locate party information, green and red arrows express recent movement, and the black tooltip creates a compact evidence pocket over the large percentage. The map does a different job: it gives the state-level claim a physical shape, even when the reader does not inspect every number. Time appears as literal update text, a range selector, and a historical line rather than through decorative animation. The page's restraint comes from exposing raw mechanisms in full view. A WVA report can borrow that causal sequence: state the volatile condition, show the current artifact, and keep the time boundary visible throughout the read.
Illustrated: the source depicts candidate chances, market range tables, state colors, and historical chart lines. Observed: this capture directly shows a fixed-width HTML 3.2 interface, candidate hover tables, local map-hover behavior, a source reload loop, and source text that names several markets. Proven: the dated source responses and rendered captures prove that those elements and values were presented during this capture window; they do not independently verify the underlying market data. Gap: the page asks the reader to accept that its aggregation, market coverage, liquidity, and resulting percentage are reliable enough to serve as a forecast. The mirror labels the result as a static archive, not a live probability or betting recommendation.
Lead with the number or recommendation, then make the inputs available through a nearby expandable detail when the reader needs to inspect the claim.
Use a map when geographic comparison changes the meaning of the data. Pair the color field with a text panel so the state result has a readable explanation.
State the update window and show a historical frame when a number changes. Time labels are evidence, not decoration.
Freeze a volatile page only after naming its capture window, stopping live refresh, and marking the interactions that remain active, frozen, or unavailable.