How We Rate Pack Opening Sites: Methodology

Score	Weight	What it measures
Reputation	30%	Star ratings on Trustpilot, the App Store and Google Play (plus PackRipper community reviews), weighted by how many ratings back them up.
Fairness	20%	Provably-fair implementation, our seed verification, and odds/RTP transparency.
Payouts	20%	Measured cash-out time, buyback credit, and physical shipping speed.
Value	15%	Buyback percentage, pack pricing versus contents, and fees.
Experience	10%	Site/app usability, reveal transparency, account limits and player tools.
Support	5%	Response time and quality on our scripted support tickets.

Principles

Source everything. Every fact and rating on a listing links to where it came from, with the date we checked it.
Claims aren't results. What a platform says is marked *. What users typically report is marked ~. Neither is presented as our own measurement. Figures we measured ourselves appear only in official reviews.
Aggregate, don't copy. We summarize recurring themes from user reviews and link to the originals. We don't reproduce reviews or record reviewers' identities.
Transparency over luck. A run of good or bad pulls says little about a platform's true odds. We score whether odds are published, verifiable and consistent.
One formula, in the open. Every listing shows its five scores and the arithmetic behind the total.

Directory listings & sources

Each listing is built from two kinds of public information, both linked in its Sources section:

What the platform publishes: homepage, pack pages, FAQ and help center, terms of service, and fairness or odds documentation. That covers prices, odds tables, RTP or EV claims, buyback terms, payout and shipping promises, KYC, restricted regions, and responsible-play tools.
What others report: review sites, app stores, the BBB, press coverage and published third-party tests. Affiliate review sites are labeled as such and given little weight.

Anything we couldn't confirm is left out or clearly attributed, for example "per third-party reviews". Unverified allegations are not published.

Aggregated reviews

Each listing collects star ratings from Trustpilot, the App Store and Google Play, plus the BBB grade where one exists, each with its review count and the date we checked. The headline user ratingis the plain average of the per-source star ratings, with every source weighted equally. A single source with hundreds of thousands of prompted in-app ratings shouldn't drown out the others. A source needs at least 20 ratings to count toward the average. Smaller ones are still shown on the listing, as are ratings for a whole company app (for example, a fantasy-sports app that added packs), which are labeled "whole app" and left out of the average.

We read the recent reviews on each source and summarize the praise and complaints that recur, noting how many reviews and which dates they cover. We also flag patterns that make ratings less reliable: solicited-review bursts, unclaimed profiles, very small samples, and large gaps between app-store and Trustpilot scores. Where the recent written reviews clearly contradict the star average, which is common with in-app rating prompts, the listing shows a ⚠ rating caution with the numbers.

Directory scores

Directory listings get six scores. Reputation is calculated automatically from the aggregated ratings above. The other five use the score rubrics below, applied to published data and user reports:

Reputation: the review-site star average, weighted by how many ratings back it up. See the Reputation rules.
Fairness: the points rubric, applied to the platform's published fairness documentation and odds tables. The 3 points for independent verification are awarded only when outcomes have actually been reproduced, by us in an official review or by a credible published third party. "No unannounced odds changes" gets a default 0.5 unless there's evidence either way. Anything confirmed only through third parties earns half credit.
Payouts: the payout and shipping bands, applied to typical user-reported times from the last 12 months, weighted by how many reports agree and how recent they are (marked ~). Where there are too few reports, we use the platform's own stated time (marked *) scored one band lower. With no data at all, the band defaults to 4. Crypto-only cash-out and recurring ID-check or fee complaints each cost up to 1 point.
Value: the advertised buyback percentage, published pack prices and odds tables, and fees in the terms. A buyback paid only in site credit that can't be withdrawn scores in the lowest band. Independent evidence of low realized value, such as on-chain data or a published third-party test, counts against the platform's own EV claims.
Experience: the public site and apps, app-store ratings, and terms on KYC, limits and responsible-play tools.
Support: recurring themes in user reviews about response times and resolutions, and how the company responds on review sites.

Each listing ends with a short note explaining how its scores were set.

Official reviews: test protocol

Official reviews are our own hands-on tests, published in a separate section and linked from the platform's listing. We open an account like any new user, fund it ourselves, and run the same seven tests in a single window, usually 2–4 weeks. The date the window closes is the Tested date on the review. Measured payout and shipping times from an official review replace user-reported figures in the directory.

01 Deposit test

We deposit using every payment method offered to a new US account: card, bank, crypto, and third-party wallets where available. For each method we record:

Minimum deposit and any deposit fees, spreads or conversion markups.
Time from payment to usable balance.
Whether identity verification (KYC) is requested, and at what point: before deposit, before rip, or only before withdrawal.
Whether deposit limits and other responsible-play tools can be set before the first deposit.
The cheapest rip available, recorded as minRipPrice.

Fees and minimums feed Value. KYC friction and limit tools feed Experience.

02 Odds & RTP claims audit

We capture the odds table for every pack we test, plus any advertised return-to-player (RTP) or expected-value figure. We check:

Completeness. Is there an odds table for every pack, covering every possible hit, with probabilities summing to 100%?
Specificity. Are card values stated with a source and date (for example, recent sold comps), or only as an unexplained "value"?
Consistency. We compute each pack's expected value from its own odds table and compare it with any RTP the platform advertises. The advertised figure is recorded as claimedRtp, or null if none is published.
Stability. We snapshot odds tables at the start and end of the test window and note any unannounced changes.

We also report the observed return on our own rips in the review, with the sample size. We do not score it: at typical test volumes, variance dominates.

03 Provably-fair verification

"Provably fair" means the platform commits to a random seed before you rip, so it can't change the outcome afterwards. We don't take the badge on trust. For at least 20 rips per platform:

Before ripping, we record the platform's hashed server seed (the commitment) and set our own client seed.
We rip and record the nonce, the pack, and the item pulled.
We rotate the seed pair, which makes the platform reveal the unhashed server seed.
We check that hashing the revealed seed reproduces the commitment exactly.
Using the platform's published algorithm (typically HMAC-SHA256 over the server seed, client seed and nonce), we recompute each roll independently, map it onto the published odds table, and confirm it produces the item we actually received.

A platform with a commitment scheme gets provablyFair: true. Only if every sampled rip reproduces correctly do we mark fairnessVerified: true. A missing or unclear algorithm, or seeds that can't be revealed, means not verified. Our guides show how to run the same check yourself.

04 Buyback test

For pulled items, we note the instant buyback (sell-back) offer and express it as a percentage of the platform's own stated value. That is buybackPercent. Where values are listed, we also compare the stated value against recent market sales for a sample of cards. A generous buyback percentage on an inflated value isn't generous. We accept at least one buyback and record how quickly the credit lands and whether it can be withdrawn as cash or only re-spent on site.

05 Payout speed test

We request a cash withdrawal of site balance using the fastest method offered, and time it from request submitted to funds received in our account. That elapsed time, in hours, is payoutHoursTested. We record any extra KYC, minimum withdrawal amounts, fees, and whether the platform's stated processing time held up.

06 Shipping test

We request physical delivery of at least one pulled item to a US address. We time it from ship request to delivery scan (shippingDaysTested), and inspect packaging, card condition, and whether the item matches the pull (including grade and cert number for slabs). We record shipping fees and minimums.

07 Support ticket test

We open three scripted tickets through the platform's primary support channel:

A pre-deposit question about odds or fees.
A status check on our pending withdrawal or shipment.
A mild dispute, such as questioning a stated card value.

We record the time to first human response and whether each answer was accurate, specific and resolved the issue.

Score rubrics

Reputation · 30%

Reputation is what users think, and it's the largest single part of the score. It's calculated, never set by hand:

Stars to 0–10. Take the equal-weight average of the qualifying sources (Trustpilot, App Store, Google Play, and PackRipper community reviews once a platform has 10 or more approved). 1★ maps to 0 and 5★ to 10.
Rating caution. If the listing carries a ⚠ rating caution (prompted or contradicted stars), subtract 1.5.
Volume. Ratings only count fully when there are enough of them. Confidence is log₁₀(total ratings) ÷ 4, so 100 ratings count 50%, 1,000 count 75% and 10,000 or more count 100%.
Unproven default. Whatever confidence is missing pulls the score toward 2.0. A platform with no qualifying ratings scores 2.0.

So a 4.5★ average from 50,000 ratings scores 8.8, while the same 4.5★ from 60 ratings scores about 5.0. Official lab reviews don't include Reputation; their weights are rescaled over the other five scores.

The other five scores are set from 0 to 10 using the rubrics below. Half-points are allowed. Official reviews apply them to our own measurements; directory listings apply them to published data and user reports, as described above.

Fairness · 20%

Points-based, out of 10:

Criterion	Points
Provably-fair commitment scheme in place before each rip	3
Our independent verification reproduces every sampled rip	3
Complete odds table published for every pack	2
RTP/EV disclosed and consistent with the odds table	1
No unannounced odds changes during the test window	1

Partial credit applies. For example, odds tables that exist but omit item values score 1 of 2.

Payouts · 20%

60% cash-out speed band and 40% shipping band, then adjusted ±1 for friction such as surprise KYC, high minimums or fees.

Cash-out time	Band	Shipping time	Band
≤ 24 hours	9–10	≤ 5 days	9–10
24–48 hours	7–8	6–8 days	7–8
48–72 hours	5–6	9–12 days	5–6
3–7 days	3–4	13–20 days	3–4
> 7 days	1–2	> 20 days	1–2
Not received	0	Not received / wrong item	0

Value · 15%

Starts from the buyback band, then adjusted up to ±2 for pack EV versus price (from the odds audit), stated-value accuracy, entry price and fees.

Buyback % of stated value	Band
≥ 90%	10
85–89%	9
80–84%	8
75–79%	7
70–74%	6
60–69%	4–5
< 60%	0–3

Experience · 10%

Scored across four equally weighted areas:

Onboarding: clear KYC requirements up front and no bait-and-switch at withdrawal.
Rip transparency: odds, item values and the seed pair are visible before you commit.
Performance: speed and usability on mobile and desktop, and how stable reveals are.
Player tools: deposit limits, cool-offs, self-exclusion and spend history. A platform with none of these is capped at 6.

Support · 5%

Based on the median first-response time across our three tickets, adjusted ±2 for accuracy and resolution.

Median first response	Band
< 1 hour	9–10
1–4 hours	7–8
4–12 hours	5–6
12–48 hours	3–4
> 48 hours or no reply	0–2

PackRipper Score & ranking

The PackRipper Score is the weighted sum of the six scores, rounded to one decimal. A worked example:

Reputation   8.0 × 0.30 = 2.400
Fairness     8.5 × 0.20 = 1.700
Payouts      9.0 × 0.20 = 1.800
Value        7.5 × 0.15 = 1.125
Experience   8.0 × 0.10 = 0.800
Support      7.0 × 0.05 = 0.350
────────────────────────────────
Overall      8.2

The directory sorts by PackRipper Score by default, with ties broken alphabetically. You can re-sort by any column, including user rating. Rarely, we pin a manual rank, and the listing always explains why.

Hard caps

Some failures outweigh everything else. When one happens, the cap overrides the formula and the listing or review says so.

Withdrawal not received within 14 days: overall capped at 3.0.
Provably-fair claimed but verification fails: Fairness capped at 2.0, with a warning on the review.
Shipped item missing, wrong, or significantly misgraded, and not resolved: Payouts capped at 3.0.
No odds table at all: Fairness capped at 4.0.
Fewer than 100 ratings across review sites: overall capped at 4.0, and the listing is tagged "Few reviews". Good published terms alone can't earn a top ranking.

Community score

Readers can leave their own reviews on any listing. Each one is read by a moderator before it's published, following our community guidelines. The community score is the plain average of approved reader star ratings (1 to 5). It appears once a platform has at least three approved reviews, alongside "% would recommend" and the most common cash-out and shipping answers.

Each listing shows three signals:

PackRipper Score (0 to 10): the formula above, including Reputation.
User rating (0 to 5): Trustpilot, App Store and Google Play ratings, aggregated and linked.
Community score (1 to 5): moderated reviews left by PackRipper readers.

Once a platform has 10 approved community reviews, their average counts as one more source in its Reputation score, updated each time the site is rebuilt. When several readers report the same concrete issue, such as withdrawals taking weeks, it also becomes a lead for our research and gets sourced like any other data point.

Keeping listings current

Every listing shows the date its facts and ratings were last checked, and each source shows its own access date. We recheck listings regularly, and straight away after major changes such as new ownership, a new fairness system, changed payout terms or a spike in complaints. Readers can flag anything out of date. Official reviews are retested after major changes, and every score change is dated.

Independence

Some outbound links may be affiliate links. Affiliate status is not an input to any score, and platforms cannot pay for placement, faster testing, or a preview of their listing or review. See About for our full policy. Spot an error? Email contact@packripper.net. Corrections are dated on the affected page.