US ranked keywords went from 27 to 31. Five keywords entered, one left. Four of the five entries are IFVG queries, and one of them carries more search volume than every other keyword the site ranks for combined.
| New this week, US | Position | Volume | Ranking page |
|---|---|---|---|
inverse fair value gap | 21 | 1,900 | /indicators/inversion-fair-value-gap |
ifvg vs fvg | 46 | 390 | /indicators/inversion-fair-value-gap |
dodgy ifvg indicator | 12 | 30 | /indicators |
4dsky points value | 36 | 30 | /indicators |
precision swing point | 65 | 30 | /indicators |
what quadrant is degrees in | — | 320 | — |
The lost keyword was never relevant to the product. Losing it costs
nothing. inverse fair value gap at position 21 with 1,900 monthly searches is the
highest-volume term the site has ever ranked for.
This is new. Last week the page ranked in three. The same URL is now picking up the same query across every English market plus Germany and the Netherlands.
| Market | Query | Position | Volume |
|---|---|---|---|
| US | inverse fair value gap | 21 | 1,900 |
| GB | inverse fair value gap | 36 | 320 |
| CA | inverse fair value gap | 37 | 210 |
| DE | inverse fair value gap | 37 | 170 |
| AU | inverse fair value gap | 33 | 140 |
| DE | inversion fvg | 17 | 110 |
| NL | inverse fair value gap | 36 | 70 |
| IN | ready aim fire indicator | 54 | 260 |
India is the first ranking ever recorded in that market. India is also the second-largest organic market by traffic, at 57 sessions in 28 days. That gap between real traffic and rank-tool visibility is landmine T4, and this is the first time the two have met.
fractal model tradingview #6, dodgy indicator #6,
dodgysdd indicator #6, forever model trading #8,
dodgysdd ifvg indicator #9. Every one is a brand or a partner term. No generic
query has reached page 1.Three items have now sat unmoved for three weeks or more. When work does not happen for three weeks running, the usual cause is that it needs someone with access, not more time.
T13, the AI crawler block. GPTBot, ClaudeBot and Google-Extended are still blocked in Cloudflare. This is the single largest blocker on the citation-gap programme, and nothing in it can be measured until it clears.
T18, the two unindexed explainers. The IFVG explainer and the liquidity pools post are still unknown to Google. The internal-link fix has not shipped.
T11, the disavow file. Ready for six weeks, still not submitted.
Add one internal link from /indicators/inversion-fair-value-gap to the IFVG
explainer (G-002), and one from /indicators/liquidity-depth to the liquidity
pools post (G-003). Both indicator pages are indexed and crawled; the blog posts are not.
An internal link from an indexed page is the cheapest way to get a page discovered.
Done = both links in source, build passing.
Search Console, URL Inspection, Request Indexing on
/blog/inverse-fair-value-gap-explained-what-an-ifvg-is-and-how-to-trade-it
and /blog/liquidity-pools-indicator-for-tradingview-explained. Two clicks.
Done = both show "Indexing requested".
Cloudflare dashboard, AI Crawl Control, allow GPTBot, ClaudeBot and Google-Extended. Done = the three bots show Allowed and a fresh fetch from each succeeds.
clients/toodegrees/disavow-2026-08-23.txt, 20 domains, still matching this
week's pull. Upload through the Search Console disavow tool. Done = the tool shows the file
with today's date.
The explainer that targets inverse fair value gap already exists. Position
21 is not a content problem, it is an indexing problem. Items 1 and 2 are the whole path.
Adding a second page would split the cluster.
Gained this week: none. Lost this week: none. Compared against the 2026-09-11 snapshot.
| Referring domain | Links | Share | Dofollow | Spam | First seen |
|---|---|---|---|---|---|
| tradingview.com | 59 | 0 | 2 | 2024-03-10 | |
| whop.com | 15 | 15 | 0 | 2025-09-19 | |
| thelabtrading.com | 7 | 2 | 17 | 2025-09-26 | |
| rattibha.com | 6 | 3 | 0 | 2026-02-22 | |
| newlyregddomains.com | 6 | 0 | 10 | 2025-08-02 | |
| xstalk.com | 4 | 0 | 10 | 2025-07-08 | |
| screenshots.wiki | 2 | 0 | 50 | 2026-03-23 | |
| jake.eu | 2 | 0 | 50 | 2026-02-12 | |
| ates.academy | 2 | 2 | 0 | 2024-09-13 | |
| netkly.com | 1 | 1 | 0 | 2026-09-02 | |
| fashionclothingnews.com | 1 | 1 | 70 | 2026-09-01 | |
| uncledspizza.com | 1 | 1 | 60 | 2026-08-26 | |
| quotesblom.com | 1 | 1 | 60 | 2026-08-24 | |
| theforbestimes.com | 1 | 1 | 60 | 2026-08-21 | |
| betwinnermirror.com | 1 | 1 | 70 | 2026-08-17 |
Concentration. tradingview.com holds 44% of every link to the site, and all 59 are nofollow. The top three sources hold 60.4%. The dofollow equity that actually carries ranking signal comes from whop.com, rattibha.com, ates.academy and thelabtrading.com. That is four domains.
TTrades co-built the Fractal Model, ranks #1 for fractal model, and there is
no link in either direction. Ask: a credit link on their Fractal Model page, offering the
same back from /indicators/fractal-model.
The affiliate code TOO is live on /ecosystem and sends them
business. There is no link back. Ask: a partner listing linking to
toodegrees.trade. Warm relationship, exact topical match, likely dofollow.
Same position as Apex. Affiliate partner, no reciprocal link. Same ask, same email.
Solves: dofollow equity.All three are Tier 1 warm relationships and none needs new content. Reddit and Medium are not on this list on purpose: those links are nofollow and do nothing for the 35% dofollow share, whatever else they are good for.
Funnel counts from Rybbit are cumulative, never windowed (landmine W6). The weekly figure is this week's count minus last week's.
| Purchase Journey | All-time | Last week | This week | Conversion |
|---|---|---|---|---|
| / homepage | 9,301 | 9,102 | 199 | 100% |
| /pricing | 3,404 | 3,315 | 89 | 36.6% |
| /checkout | 639 | 618 | 21 | 6.87% |
| /welcome | 112 | 107 | 5 | 1.20% |
5 purchases this week, against 21 checkout starts. End-to-end conversion 1.20%, up from 1.18%. The Indicators funnel added 4 of those 5. Reached pricing across all organic traffic: 22.1%.
The Whop exit path is still untracked (T15), so any purchase completed on Whop rather than on site is missing from these counts. Treat 5 as a floor.