14:02 ET · NHL OPEN
Home side on the money line, taken at 1.85
Opponent on the second night of a back-to-back with a backup expected. Number has shortened since posting, which is the only part of this we can measure at the time we post.
PICKS DESK · UPDATED AUG 9, 2026
Today's picks
Posted with the price we took, timestamped, before the number moves. The archive keeps every one of them, including the ones that lost.
The three cards below show the format of a posted position. Matchups are described generically and no page on this site claims the outcome of a real fixture.
11:40 ET · NHL OPEN
Road side +1.5 on the puck line at 1.95
Rested opponent, short puck line, thin edge on the number. Small position, one unit, and it stays one unit whatever happens to the price after.
09:15 ET · NHL SETTLED
Total over 6.5, taken at 1.93
Wrong on the number and wrong on the price: it drifted after posting, so we paid over the market for a position that did not land. Left up, because deleting losers is how a record turns into fiction.
THE RECORD, ALL 240 POSITIONS
+1.8%MEDIAN GAP ON THE MARKET
54.2%POSTED AHEAD OF THE MARKET
240POSTED · 0 DELETED
45.8% of positions went up at a worse number than the market median at that moment. Those are in the archive too, at the same size, with the same one-line reasoning attached.
How the archive works →The three product cards, today
Two of the cards below carry a button that pays Loon Picks a commission if you open an account. It buys nothing here: the app is picked on the teardown numbers and the offer on the arithmetic printed on the card.
THE APP 
Bet99 Casino 3 taps from the lobby to a submitted withdrawal, against a board median of
5. 147 MB, 0 crashes in 40 sessions.
Get the app The teardown →
Bet99 Casino 3 taps from the lobby to a submitted withdrawal, against a board median of
5. 147 MB, 0 crashes in 40 sessions.
Get the app The teardown → THE OFFER 
Betty Casino 100 wager-free spins at no playthrough.
The arithmetic turns a C$20 headline into C$20.
Claim this offer The full sum →
THE SKIP Betty Casino 100 wager-free spins at no playthrough.
The arithmetic turns a C$20 headline into C$20.
Claim this offer The full sum → 
Captain Cooks Casino 200x playthrough means C$99,000 of
staking, which is 3960 days at C$25 a day. The headline is doing all
the work here.
Why it prices badly → How a position is priced
Start with a number we think is wrong and say why in one line. Then check what the rest of the market is showing at that moment. The gap between the price we can get and the market median at the time of posting is the only measurable thing available before the event happens, and it is the figure printed on every card.
That figure is not a result and it does not predict one. It says whether the price taken was better or worse than what everyone else was offering when the bet went on, which is the part a desk can actually control.
Where a market is thin, we say so and the position gets a smaller stake. Where the number moves against us before posting, the position does not go up at all.
The stake plan, written down
One unit is one percent of the desk bankroll. Almost everything is one unit. Two units is the ceiling and it is rare enough to be worth noticing when it appears.
Nothing is staked up after a loss and nothing is staked up because a position feels good. A flat stake is the only arrangement that lets anybody tell the difference between a bad method and a bad month, which is the entire reason for keeping a record.
The same plan, in a version you can hold at midnight, is written up at a stake plan.
What this desk does not do
No result service. We do not publish claimed outcomes of specific fixtures and there is no page here telling you who won anything.
No guaranteed strike rate, no locks, no subscription. A genuine edge is worth more to its owner than a subscription is, which is worth remembering every time somebody offers to sell you one.
And no deleted positions. The archive keeps everything, and the aggregate at the top of it is computed from the whole set rather than from a selection.
The other three cards
Alongside the positions, three product cards drop each weekday: the app, the offer, the skip. Those are about builds and terms rather than about games, and they use the same numbers as the rest of the site.
The app card is a package that changed. The offer card is a promotion live that morning with the playthrough attached. The skip is one thing to avoid today, which on most days is an offer whose headline is doing all the work.
The three cards
Card one is The App. One operator, one reason, three numbers: store rating with its ratings count, package size in MB, and taps to withdraw. The reason is always product, never brand. A card might read as simply as a build dropping from 96MB to 61MB after a cleanup release.
Card two is The Offer. A promotion that is live that morning, with the wagering requirement, the qualifying deposit and the expiry pulled from the operator's published terms. If the terms page and the app's in-product copy disagree, the card says so and the offer does not get recommended.
Card three is The Skip. One thing to avoid today. Usually a broken release, a payout method that has quietly disappeared, or an offer whose playthrough makes the headline number decorative.
Every card is numbered and permanent. Pick 412 stays at pick 412 even after it ages out of the top of the feed.
What goes into a card
Sources are boring on purpose. Store listings supply size, version history and rating. The operator's published terms supply payout windows, limits, wagering and expiry. Release notes supply what changed, when the operator bothers to write them.
Behaviour that no store page shows, such as how often a login loop appears after an update, comes from patterns reported across player accounts and public review text. Those are described as reported, never as verified.
Nothing on a card is a first hand payout claim. If a withdrawal window is quoted, it is the window the operator publishes, labelled as the operator's own figure. We do not print a gap against it. Grading a company against its own advertising is a comparison this desk is not in a position to stand up, so it does not appear anywhere on the site.
A card needs at least two independent sources to name a specific failure. One angry review is weather. Forty reviews naming the same screen in the same week is a pattern.
Why The Skip exists
Most affiliate feeds only push. This one publishes a negative every day, and it is the most read card in the set.
A skip is not a permanent verdict. It is a snapshot: this build, this week, this offer. When an operator fixes the thing, the fix gets its own card and the old skip stays online with a dated update on top of it.
The skip card also catches quiet regressions. An app that goes from three taps to five, or that starts asking for a document upload on every withdrawal instead of the first one, has made the product worse without changing the marketing at all.
Reading it on a phone
The feed is built for a thumb. Cards are full width, the three numbers sit above the fold, and the whole day fits in two scrolls on a standard 6.1 inch screen.
No autoplay video, no interstitial, no popup asking for notification permission. Total page weight for a day of cards runs under 400KB, which loads on transit LTE without a spinner.
Links out are labelled. If a card links to an operator, the label says so, and the arrangement behind that link is explained at how we rate.
What a pick is not
A pick is a product observation, not a prediction and not advice. Nothing in the feed says a bet will win, and nothing in it estimates anybody's odds of walking away ahead.
Offers are described, not endorsed. A card listing a 30x wagering requirement on deposit plus bonus is telling you the arithmetic, and the arithmetic usually favours the house.
Picks are not personalised. They do not know your bankroll, your province or your history, and they should not be treated as a plan for either.
Gambling is 19+ in Ontario and most provinces, 18+ in Alberta, Manitoba and Quebec. ConnexOntario is free and confidential at 1-866-531-2600, 24 hours a day.
The archive
Every card stays up. Nothing is quietly deleted, because a feed that erases its misses is not a record of anything.
Corrections appear on the card itself with a date stamp, not in a separate notice nobody finds. If a payout window was quoted wrong, the wrong number stays visible with a strike through it.
The archive is also the fastest way to see how an operator behaves over time. Six months of cards showing repeated login failures after each release says more than any single teardown or any position on the ranked list.