What shipped, what was removed, and what was corrected. Newest first, dated the day it actually landed.
Removals and corrections are listed here beside the new things, because a changelog carrying only good news is not a record of anything. So this page is where you will read that a tool was deleted for not being good enough, that a claim on the pricing card was wrong and has been fixed, and that a panel on the home page was animating numbers it had invented until it was torn out.
Nothing here is forward-looking. There are no roadmap items, no coming-soon entries and no dates in the future — an entry appears on this page after the thing exists, which is the only version of a changelog worth reading.
Last updated August 16, 2026
Removed
Paper Trading has been removed
The practice account is gone — deleted rather than hidden, along with its page, its documentation and every mention of it across the site. It was not a good tool, and keeping a weak one on the list to make the list longer is how a product surface starts lying about itself.
If you had bookmarked it, both its old addresses now send you to the tool directory instead of a dead end. A few strategy write-ups used to tell you to rehearse a play there; those passages now say the more useful thing, which is to size the first position small enough that it is a tuition fee. A simulator teaches the mechanics of a market but not the part that actually costs people money, which is how differently you behave when the stake is real.
Everything else is unchanged, and the membership price did not move.
The Forward Test marks published signals forward from the day they printed — the only honest answer to what following them would have done. It needs snapshots taken before the outcomes are known, and the job that takes those snapshots existed but was never actually scheduled to run. So the page had nothing to show since the day it launched.
It runs daily now. Cohorts become eligible once they are a day old, so the first readable cohort appears roughly two days after collection begins, and the record grows from there. Until then the page says collection has just started, which is the truth.
Deliberately not fixed the fast way: replaying the leaderboard backwards would have filled the page immediately and measured nothing, because those wallets are on the board precisely because they just won. That is a sort order, not an edge.
A locked tool now explains itself instead of sending you home
Clicking a members-only tool while signed out used to drop you on the landing page — the site answering a question nobody asked, at the exact moment you had asked a very specific one.
Now that click lands on that tool's own documentation page: what it computes, which venue feeds it reads, how to read its output, and the section on what it cannot tell you. You get the answer you clicked for, and the decision about membership comes after it rather than instead of it.
Members are unaffected — the tool opens as it always did.
The sidebar and dashboard now lead with six tools — Live Sharp Feed, Sharp Leaderboard, Sharp Consensus, Insider Radar, Market Browser and the Arbitrage Scanner — in larger rows, each with its own line of context. The remaining tools sit under a More tools disclosure that opens itself whenever you are on one of them, so you are never on a page the rail cannot show you.
Nothing was removed and nothing became an add-on: every tool still ships, still works, still has its own public page, and is still included at the same price. A twenty-item menu is simply harder to read than a six-item one, and reading a menu is not what you came for.
The dashboard mirrors the same split: six full cards, then a compact block for the rest, under a line stating plainly that nothing there is an upsell.
New
Real figures on the pages where you decide
The pricing page and the tool directory now carry a small strip of live numbers from the same public feed the landing page uses: how many scored wallets are being tracked, their combined recent performance, and how many traders and markets are in the current tape.
The one claim an analytics product cannot make credibly is trust us. So the strip does not argue that the product is good — it states what it is watching right now, and every figure on it can be checked against the venues themselves.
It hides itself when it cannot be honest: each number is independently optional, and if fewer than two real values are available the strip renders nothing at all. There is no placeholder path, so there is no way for it to ever show an invented number.
Three new public pages explain the ground the product stands on: what a Polymarket wallet really publishes and what it does not, what Kalshi makes public and where inference has to stop, and an even-handed comparison of the two venues.
The Kalshi page leads with the uncomfortable part, because it is the honest wedge: Kalshi publishes no per-user data, so a Kalshi whale wallet is a product category that cannot exist. What Kalshi does publish — the trade tape, block prints, taker side, open interest — is real information about flow, and that is what we read.
One correction shipped with them. The methodology page claimed capital efficiency was the strongest weight in the Sharp Score. It is not; settled profit carries the larger coefficient. The page now states the real bounds instead of ranking them wrongly.
A documentation page for every tool, readable without a membership. Each one is written from the real implementation and carries a mandatory section on what that tool cannot tell you — the part that separates documentation from a sales page.
Alongside them: a brand page saying plainly who operates WhaleTracks and what it is not, and a forty-term glossary defined the way these two venues actually work rather than the way a dictionary would put it — Kalshi's fee curve, the difference between Polymarket's negative-risk flag and Kalshi's mutually-exclusive one, why every priced opportunity here quotes the executable ask and never the last trade.
Also a plumbing fix with a visible consequence: the site used to answer on two hostnames while telling search engines only one of them counted, splitting the brand across two addresses. The www address now redirects permanently to whaletracks.com.
Icons for every tool, and every claim re-checked against the code
Every tool now has its own drawn icon rather than a two-letter tile, and the written surface caught up with the restored product: the tool guides went from thirteen articles to twenty-three, the strategy write-ups from four to eight, and the comparison tables were rewritten line by line.
The comparisons are worth calling out. At the current price several rivals undercut us, so those pages now say so and argue on breadth instead of pretending otherwise. A comparison table you cannot lose a row on is not a comparison table.
What the pass refused to claim is the point of it: no mobile or background delivery for alerts, because there is none; no market-watching in the Watchlist, because it holds wallets; and no edge claimed for the Fade Board, because its inverse test came out negative.
A trim two weeks earlier had cut the product to six tools. This put the other fifteen back — the same real implementations, recovered and re-fixed against the current codebase, not stubs written to make a list look longer. Every restored page was loaded and every restored feed checked for live data before it went back on the list.
The tools are now organised in four groups: Track for following the wallets themselves, Signals for what the flow is saying, Trade for price and structure, and Plan for sizing, schedule and the write-up.
Two honesty corrections landed with the restore. The Fade Board no longer implies an edge — its inverse-tail test was negative and the page says so on the page rather than in a comment — and an unsourceable statistic about how many traders lose money was deleted rather than re-cited.
Membership settled on a single monthly plan at $49.99/mo with full access. The weekly plan was retired everywhere, including in the code that could still have charged for it, and there is still no free trial and no free tier.
A claim on the launch-offer card was also corrected, because it was false: it said membership would rise to the list price once the offer closed. It will not — when the offer closes the price falls back to the standard monthly rate, and the list price is a number nobody would ever have been charged. The card now claims only what is true, which is that the rate you buy at is the rate you keep while the membership stays active.
Signing in now takes you to the live trade feed rather than a dashboard. The feed is the product's pulse: it is the screen worth having open, so it is the one you land on.
The sidebar was redrawn at the same time — larger type, and real icons in place of the two-letter tiles.
Fixed
A proper 404, a faster dashboard, and a hole in the gate closed
A mistyped address used to render a bare white panel inside the dark site. There is a branded 404 now, and it routes you somewhere useful depending on whether you are a member.
A genuine paywall bug was closed at the same time: appending a dot-suffix to certain addresses served members-only pages to anyone. The rule that let that happen was narrowed so it can only ever apply to real static files.
On the performance side, the dashboard stopped being able to trigger the heavy market crawls it only ever needed cached results for, and the scheduled refresh took ownership of the crawl cadence. The visible effect is a dashboard that loads in bounded time instead of occasionally waiting on hundreds of upstream requests.
New
A dashboard and a permanent tool rail
The app got a proper shell: a dashboard as its front room, and a sidebar carrying the full tool list on every page, so moving between tools stopped being a trip through the header.
The Market Analyzer arrived in the same release — one market, every signal held on it, on a single screen.
NewFixed
You can cancel online, and the ticker stopped inventing numbers
The site said cancel anytime in six places while offering no way to cancel. That is now real: Manage subscription opens the payment provider's own customer portal, where you can cancel, change your card and read your invoices. It is one click from the header, the app sidebar and the pricing page.
Auto-renewal is now disclosed next to the buy button rather than only in the Terms, and checkout itself repeats that the membership renews until cancelled.
The worse finding was on the landing page: a panel labelled LIVE was animating upward gains that were fabricated in the browser. It was deleted. A number on this site is real or it is not shown, and that panel had been breaking the rule in the most visible place we had.
A round of performance work with a measurable effect on the screens you keep open: the membership check is cached briefly instead of hitting the database on every request, the live feed reuses its assembled payload between polls rather than rebuilding it every few seconds, and background pollers now stop when a tab is hidden.
The landing page also stopped waiting on a multi-megabyte upstream download to render two numbers, and the text above the fold now paints without waiting for JavaScript to load first.
Separately, the site's map for search engines was rewritten to list only pages that are genuinely public. It had been advertising hundreds of members-only pages, which is a good way to teach a crawler that none of your pages are worth trusting.
Fixed
Links stopped dead-ending at the paywall
Turning on the membership gate broke every link that pointed through it, and the worst of them were on the pages people arrive on: each guide and strategy article ended by linking to the tool it had just described, and a signed-out reader's click was spent on a bounce back to the home page.
Every one of those links now resolves by membership — members get the tool, everyone else gets the page that explains what it costs. Links to genuinely public pages still go straight there for everyone.
The site also split its navigation in two: a marketing header for the public pages and the full tool navigation inside the app, so the public site stopped advertising thirty links a visitor could not open.
Changed
Every free-tier claim removed
The product had stopped offering a free tier and a free trial, but the site was still advertising both in several places. All of it was removed — the pricing table's free column, the trial language, and the claims scattered through the comparison pages.
The Terms were rewritten in the same pass to describe the membership model as it actually works, so the legal page and the buy button finally agree.
The tools became members-only, and sign-in comes before checkout
The analytics tools moved behind a membership. The editorial side of the site — the guides, the strategy write-ups, the methodology and the legal pages — deliberately stayed open to everyone, and still is.
Sign-in became a hard requirement before checkout. Without it a payment carries nothing that links it to an account, which in testing produced exactly what it sounds like: completed payments with no way to grant the access they bought. Signing in first makes that impossible.
Changed
Sign in with Google, on our own domain
The login path was rebuilt so the code exchange happens on whaletracks.com rather than at a third-party authentication vendor. The practical difference is what Google's consent screen says when you sign in: continue to whaletracks.com, the site you are actually signing into.
Sessions are a signed, http-only cookie lasting ninety days, and your Google account is the identity a membership attaches to. There is still no password to set or lose.
The product was renamed and re-themed: a new name, a diving-orca mark carried through the whole icon set, and the near-black interface the site uses today.
The legal footing was set at the same time. The operating entity is WhaleTracks LLC, a California limited liability company; the Terms of Service and Privacy Policy were rewritten from scratch for this product, and one support address was published for everything.
The first working version: a live feed of large trades from scored wallets, a leaderboard ranking traders by how much of their record looks like repeatable skill rather than size, the Insider Radar for fresh wallets taking concentrated positions, a portfolio reader for any Polymarket address, and cross-venue market browsing.
Everything it showed then came from the same place it comes from now — the public APIs and public blockchain data the two exchanges already publish.
Found something that does not match?
If a page on this site claims something this history does not support — or a figure disagrees with the exchange it came from — that is a defect, and we would rather hear it than not. The help page has the address, along with how membership, cancellation and refunds actually work.
WhaleTracks is informational analytics, not financial or investment advice. It never places trades and never holds funds. Not affiliated with Polymarket, Kalshi, or Manifold. Past performance does not guarantee future results.