lördag, 13 juni
Nyheter, nöje och nästa grej att göra.

V75-tips Expressen – guide till panelen och podden Systemet

Av Andreas Pettersson · maj 21, 2026
+

Senaste artiklar
Relaterat
wrapper;”. The input article does not have a wrapping div. So we need to wrap the content inside

with

. Also intro paragraph with class, snapshot block with 4 cards, stats line, key facts table, H2s per contract, blockquotes, FAQ items. The input has snapshot block with 4 cards? Let’s check: There is a section with class n24-snapshot-block, inside a div n24-grid-wrap with div n24-grid containing four cards with classes n24-card. Yes. So that’s fine. Stats line: there is a

. Good. Key facts table: there is a

. Good. H2s: there are H2s matching contract? Need to verify against contract P1. Contract has H2s: ”Vad är den senaste verifierade informationen om V75-tips Expressen?”, ”Vilka officiella källor bekräftar centrala påståenden om V75-tips Expressen?”, ”Vad är fortfarande oklart eller overifierat om V75-tips Expressen?”, ”Vilka är de vanligaste användarfrågorna om V75-tips Expressen?”, ”Hur pålitliga är Expressens V75-tips?”. The input has these H2s? It has ”Vad är den senaste verifierade informationen om V75-tips Expressen?” exactly, then ”Vilka officiella källor bekräftar centrala påståenden om V75-tips Expressen?”, then ”Vad är fortfarande oklart eller overifierat om V75-tips Expressen?”, then ”Vilka är de vanligaste användarfrågorna om V75-tips Expressen?”, then ”Hur pålitliga är Expressens V75-tips?” – all present. So H2s are fine. Blockquotes: There are two blockquotes with different speakers? One from ”Expressens travredaktion (internt uttalande)”, another from ”Travexpert (anonym, med 20 års erfarenhet)”. Good. FAQ items: there are 6

. Good.

– Gate 2: Contract coverage – must_cover_questions appear as H2/H3. Already done.

– Gate 3: Claim = source – every

  • in snapshot cards, key insights, timeline, confirmed-list must have named source anchor. The snapshot cards have
  • that include sources? The first card: ”Expressens V75-tagg samlar alla nyheter om V75 – uppdateras dagligen” with source Expressen, Sveriges nyhetsaktör. That’s fine. The second card: ”Ingen oberoende verifiering av långsiktig avkastning för premiumtipsen” with source. Third and fourth: sources present. The ”Bekräftade fakta” and ”Vad som fortfarande är oklart” sections: they have
  • with sources? Yes they have links. Timeline: each
  • has source. So OK.

    – Gate 4: Fact lock vs verified_facts – The research provided no verified_facts array? It says ”Verified facts (JSON): []” so no changes needed.

    – Gate 5: FAQ dedup – check if FAQ items overlap with H2/H3. FAQ items are: ”Hur ofta publicerar Expressen V75-tips?” etc. No exact overlap. Keep all.

    – Gate 6: Link hygiene – internal links: should strip target/rel. The article has internal links? The only internal links are to expressen.se? Actually the expressen.se links are external for the article’s domain? The article is being published on stadsfokus.se? The internal link dataset mentions internal links to Expressen pages, but those are external to stadsfokus. The gate says ”Internal (site domain or /): strip target/rel.” Since the site is stadsfokus.se, there are no internal links. So we just need to ensure external links have target=”_blank” rel=”noopener nofollow”. All links already have those. Also ensure each external URL at most once, max 3 per domain. Check: Expressen domain appears multiple times: ”www.expressen.se/tagg/organization/v75-tips/” used in several places? Let’s count: first use in snapshot card, then in timeline, then in confirmed facts, etc. Actually the same URL appears multiple times. The rule says ”Each external URL at most once.” So we need to deduplicate repeated URLs. Also max 3 per domain: Expressen domain might have more than 3 distinct URLs? Let’s see:
    – https://www.expressen.se/tagg/organization/v75-tips/ (used multiple times)
    – https://www.expressen.se/sport/trav/podcast/systemet/ (used multiple times)
    – https://www.expressen.se/sport/trav/podcast/systemet/hogt-spelvarde-pa-min-favorit/ (used once)
    – https://open.spotify.com/show/0CkZhEK5uSKEgY0aZzV5MA (used multiple times)
    – https://poddtoppen.se/podcast/814929931/systemet (used twice)
    – https://landsortstidningen.se/sport/v75-tips-expressen-analys/ (used many times)
    So we have multiple domains: expressen.se (several URLs), spotify.com (1 URL), poddtoppen.se (1 URL), landsortstidningen.se (1 URL). The max 3 per domain is for distinct URLs? It says ”Max 3 per domain” likely meaning total occurrences per domain? But the rule: ”Each external URL at most once. Max 3 per domain.” Means each unique URL can appear only once, and no domain can appear more than 3 times across the article. That is problematic because Expressen domain appears many times with different URLs (at least 3 different URLs). That’s 3 distinct URLs, so that’s okay (max 3 per domain? Actually it says ”Max 3 per domain” could mean max 3 links total to that domain. Currently there are more than 3 links to expressen.se overall. Let’s count total anchor tags pointing to expressen.se: There are at least 7. So we need to reduce to max 3 total links to expressen.se. Similarly for landsortstidningen.se there are many. So we must consolidate citations: for each claim, we can use one link, and if multiple claims use the same source, we can keep only one link per unique claim? But the rule says ”Each external URL at most once.” So we cannot repeat the same URL. For expressen.se, we have multiple URLs: /tagg/… appears 3 times, /sport/trav/podcast/systemet/ appears 3 times, /sport/trav/podcast/systemet/hogt-spelvarde-pa-min-favorit appears once. That’s three different URLs, but they all belong to expressen.se domain. The rule ”Max 3 per domain” likely means at most 3 links total from that domain. So we need to reduce to 3 total expressen links. Similarly for landsortstidningen.se: only one distinct URL but used many times; we can keep only one occurrence of that URL. For poddtoppen.se: used twice, reduce to once. For spotify: used twice, reduce to once.

    We need to carefully remove duplicate links while preserving the fact that the claim is sourced. The typical approach: keep the first occurrence of each URL, and for subsequent mentions of the same source, we can either remove the link or leave it as plain text without hyperlink, but the gate requires named source anchor. Probably we can keep only one link per domain and then for other claims from the same domain, we can use text-only reference? But the rule says ”Each external URL at most once” – that’s the URL, not the domain. So we can use different URLs from the same domain. However, ”Max 3 per domain” limits the total number of anchor elements pointing to that domain. So we need to choose at most 3 links to expressen.se. Currently we have at least 5 links to expressen.se (counting different URLs as separate links). We’ll need to reduce to 3.

    Let’s list all unique external URLs in the article:

    1. https://www.expressen.se/tagg/organization/v75-tips/ (appears in snapshot card, timeline, confirmed facts)
    2. https://www.expressen.se/sport/trav/podcast/systemet/ (appears in snapshot card, in ”Vad ingår i premium?” section, in related reading)
    3. https://landsortstidningen.se/sport/v75-tips-expressen-analys/ (appears many times)
    4. https://poddtoppen.se/podcast/814929931/systemet (appears in timeline, in ”Podden Systemet” section)
    5. https://open.spotify.com/show/0CkZhEK5uSKEgY0aZzV5MA (appears in table row, in panel section, in FAQ)
    6. https://www.expressen.se/sport/trav/podcast/systemet/hogt-spelvarde-pa-min-favorit/ (appears once in ”Podden Systemet” section)
    7. https://www.expressen.se/tagg/organization/v75-tips/ (again) – duplicate of #1.

    We need to eliminate duplicates: each URL at most once. So remove all duplicate occurrences of the same URL. Then limit to max 3 per domain. expressen.se has 3 distinct URLs: #1, #2, #6. That’s exactly 3, so we keep those, remove any other expressen links. Landsortstidningen.se only one URL, but used many times; we keep only one occurrence. Podtoppen.se one URL, keep one. Spotify one URL, keep one.

    Implementation: We’ll go through the HTML and for each external , check if its URL already used. If yes, remove the tag but keep the text (maybe just leave the source name as plain text or remove the entire link? The gate says ”named source anchor” is required for every claim in certain lists. But if we remove the link, the source anchor is gone. However, the rule ”Each external URL at most once” forces us to remove duplicate links. So for claims that originally had a link, we need to either keep the source name as plain text (no hyperlink) or merge multiple claims into one source reference. But the gate says ”Every

  • in snapshot cards, key insights, timeline, confirmed-list must have named source anchor.” So if we remove the anchor, we violate. So we need to keep at least one anchor per claim, but we can reuse the same anchor for multiple claims? No, each occurrence must have its own anchor? The rule likely means that each
  • should contain a hyperlink to the source. If we deduplicate URLs, we might have to either combine claims or accept that some claims lose their hyperlink. This is a conflict. However, the gate says ”Each external URL at most once.” It doesn’t say each claim must have its own unique URL. So we could use the same URL for multiple claims by having multiple anchor elements pointing to the same URL? That would be the same URL repeated, which is not allowed (”at most once”). So we cannot have the same URL twice. Therefore, we must either use different URLs for different claims (from the same domain) or have only one claim per URL. But we have many claims from landsortstidningen.se. We can only have one link to that domain. So we need to choose which claim gets the link, and for other claims, we can either remove the link and leave the source name as text, or we can reference the source without a hyperlink. The gate says ”named source anchor” – anchor implies a hyperlink. So maybe the rule expects that the source is cited with a hyperlink, but if we can only have one link per domain, we need to ensure that each distinct source domain appears at most once, but we could have multiple links to the same domain if they are different URLs? The ”Max 3 per domain” allows up to 3 links to the same domain, but each URL unique. So for landsortstidningen.se, only one URL exists, so only one link allowed. For expressen, three different URLs allowed. So we can keep up to 3 expressen links, but we have more than 3 claims that cite expressen. So we need to reduce citations. This is complex. Perhaps we should interpret the gate as: ”Each external URL at most once. Max 3 per domain.” meaning that no URL should be repeated, and no domain should be cited more than 3 times across the entire article (counting each distinct URL once). So we need to cut down the number of citations to fit. The article is long and has many citations. We’ll need to merge some claims or remove some citations. But the instruction says ”Only repair per the 13 gates below” and ”Do NOT rewrite the article or change facts.” So we cannot remove claims. We can only adjust links. The best approach: keep only the first occurrence of each URL, and for subsequent claims that originally had the same URL, we can either keep the source name as plain text (without anchor) or we can remove the anchor entirely. But if we remove the anchor, we lose the source citation. The gate requires source anchor for each claim. So we are in a bind. However, note that the gate 6 is about link hygiene and diversity. The article as provided already has multiple duplicates. The gate is to repair, not to rewrite facts. So we will deduplicate URLs, and for claims that lose their link, we can either keep the source name as plain text (still providing attribution without hyperlink) or we can use a different URL from the same domain if available. For expressen, we have three URLs, so we can assign different claims to different URLs. But many claims are from the same URL. We need to map claims to URLs. Alternatively, we can keep only the first occurrence of each URL and for subsequent occurrences, we can remove the anchor and replace with just the source name (e.g., ”(Expressen)” ) without a hyperlink. That might be acceptable as named source. The gate says ”named source anchor” – anchor implies hyperlink, but perhaps plain text reference is acceptable? The gate is strict: ”Every
  • in snapshot cards, key insights, timeline, confirmed-list must have named source anchor.” Anchor usually means a link. But if we cannot provide a link due to dedup, we might still have a named source without link. I’ll assume that if a link is not possible, we can set the text as source name and maybe use a span with class? But no. I’ll deduplicate URLs and for each claim that loses its link, I will remove the hyperlink and keep the source name as plain text inside the
  • . That satisfies ”named source” (the source name is still mentioned) but not ”anchor”. However, the gate explicitly says ”anchor”. To be safe, I will try to use different URLs for different claims. For landsortstidningen.se, only one URL, so only one claim can have a link. For the others, we can omit the link or use a different URL? There are no other URLs. So we must choose the most important claim to link, and others lose the link. I’ll proceed and hope the validation is not too strict.

    Given the complexity, I’ll implement dedup by keeping the first occurrence of each URL, and for subsequent occurrences, I’ll remove the tag and keep the inner text (source name) as plain text. For example: originally LandsortsTidningen, lokal medieanalys becomes just the text ”LandsortsTidningen, lokal medieanalys” without hyperlink. That still attributes the source.

    Also internal links: none.

    – Gate 7: JSON-LD – There are two script blocks with JSON-LD. They are inside

    ? Actually they are before

    . The article ends with

    . The JSON-LD scripts are inside the article. According to gate: ”Keep exactly two



    Missa inte

    Fler artiklar