Fortune Gems Scratch by JILI

A scratch-card take on Fortune Gems

Fortune Gems Scratch works best as a simple reveal-and-read game: buy a card, uncover its nine squares and look for the crowned mask forming a complete line.

What I liked

✅ I like how outside prizes sit beside their lines, making a completed target line easy to price.

✅ Manual scratching and quick reveal offer two ways to uncover the same card

✅ Two-line wins are easy to follow; portrait keeps the card prominent

What I didn’t like

❌ Ordinary matching gems can mislead a new player who skips the mask rule; a matching-gem line can still pay nothing.

❌ The feature set is narrow compared with a reel game built around changing multipliers

❌ Landscape shrinks the card, and the maximum-prize sources disagree

This is JILI’s game 441, with a December 2024 release month on its official product page. It uses a 3×3 scratch card rather than the reel mechanics of other Fortune Gems titles. Our 100 settled demo cards cost 2,000 units and returned 640, a net loss of 1,360. Those results help show how this particular session unfolded; they cannot establish the game’s theoretical RTP.

Use the demo on this page to learn the target symbol and check the card price before starting. The review below explains the controls, the winning rule and three actual outcomes that make the differences clear.

Card 4: 12.4 seconds of real manual scratching. Still frames approximate the pauses between captures; this is a silent clip, not native screen recording. It ends before the card is fully revealed. The card’s later confirmed payout was 0 and it is already included in the 100-card sample.

Philippine search interest. Ahrefs’ Philippines database, checked on 5 October 2026, reported the following monthly search-volume estimates. These are search queries, not player counts; overlapping phrases should not be added together.

The returned estimates show a small named-game search footprint. A reported zero does not prove that nobody searches for the phrase; missing keyword data is a separate case.

The interface, artwork and controls

The warm gold ruins and jewel-like card cover fit the Fortune Gems name, but the important design choice is the tall central card. The picture keeps the nine squares, their outside prizes and the card cost together. Balance and WIN sit below, with three large action buttons underneath. I find that separation more useful than the oversized headline prize, which is easy to mistake for a statement about the current card.

Complete desktop interface before purchase: Balance 2,000, Bet 20 and WIN 0. The numbers identify the real areas below; they do not alter the game display.

1. 3×3 cover. Scratch these nine areas. The identical gold faces are the covering, not nine revealed winning symbols.

2. Eight outside prizes. Three above, three on the right and two diagonal corner positions correspond to the card’s possible straight lines.

3. WIN UP TO. The Bet 20 screen shows 200,000 here. It is a headline amount, not a credited payout.

4. Card price. The coin value is 20 in this screen. Confirm it before buying.

5. Balance. The demo wallet contains 2,000 before this purchase. Later screenshots show different points in the test.

6. WIN. This is the card’s displayed payout. Subtract its cost to work out the net result.

7. Bet Select. The blue button opens the available prices. Ten values were visible in this version.

8. Start. The green arrow buys a card at the current price. After purchase, this position becomes the red OPEN ALL button.

9. Auto Spin. Help says the yellow button starts automatic play and a second click stops it. Automatic runs were not tested.

10. Option. The left-hand gear opens the settings area, including access to rules and sound controls.

What the expanded menus show

The Bet menu offers 20, 30, 50, 80, 100, 150, 250, 500, 800 and 1,000 demo units. The screenshot below has 30 selected; the settled sample used 20 throughout. The options panel and Help were opened, while History’s detailed records, automatic play and audible sound were not assessed. Help calls the red lightning control “Scratch All”; on the live card it reads “OPEN ALL.” Both refer to revealing the purchased card quickly.

Actual Bet menu. A visible price option is not evidence that a card was bought at every listed amount.

Actual settings panel. Rule and sound controls are visible; their presence does not establish an audio or History test.

Portrait is the clearer layout

In a 390×844 desktop viewport, the full card, all eight prizes and the main controls fit comfortably. The Bet menu opened, and clicking outside dismissed it. Clicking the already-selected 20 left the menu open; no change to another price was demonstrated.

390×844 portrait simulation: the card and amounts remain prominent.

Same portrait viewport with the ten-price menu open.

844×390 landscape simulation. The card is only about 207 pixels wide, making the prizes and symbols smaller.

Landscape fits the complete layout, but it wastes much of the extra width on background. I would choose portrait for reading this card. The landscape Bet response was not verified: the session subsequently went blank and recovered after reopening the official demo. These were desktop viewport checks, not physical-phone, touch, mobile-network or iPhone playback tests.

How a card wins

Choose a price, press Start and reveal the squares by dragging over them or using OPEN ALL. Wait for the completed payout and the green Start button to return before counting the result. Repeated scratching gestures on one purchased card are still one card.

The winning target is the crowned, fanged mask shown in Help. Three of these masks on a complete line win the amount next to that line. “Mask” is a description of the artwork; the game does not supply a formal symbol name. There is no documented Wild substitution rule to apply to the letters or ordinary gems.

Official in-game rule illustration, not an outcome from the test. The dollar-marked example is illustrative and does not identify the currency of our demo balances.

Read the prize beside the completed line. These were the eight displayed prize positions at Bet 20. The layout count comes from the actual card; it is not a claim that the game uses eight slot paylines.

The revealed artwork observed in the test included the mask, A, J, Q, K, a six-sided gem, an eight-sided gem and a round faceted ornament. Help gives no separate paytable for the ordinary letters and gems. The gold covering is not an additional revealed symbol.

Eight actual revealed-symbol crops from the test. The gold face on an unscratched square is its covering, not the revealed target.

The reveal, the target line and two-line wins

A completed line can simply return the card cost

Card 9 is a useful first example. Buying it reduced the balance from 1,840 to 1,820. After revealing the left column’s three masks, the screen still briefly showed WIN 0 and the red OPEN ALL control. The confirmed result then highlighted the 20-unit prize above that column, displayed WIN 20 and restored the balance to 1,840.

Card 9, purchase: 20 units debited; the squares are still covered.

Card 9, reveal: the target line is visible, but WIN 0 is still a process reading.

Card 9, settlement: payout 20 minus cost 20 equals net 0. The green Start button has returned.

That is a break-even card. The award animation is welcome feedback, but the wallet has only recovered what this card cost.

Three ordinary matching gems do not satisfy the mask rule

Card 61 uncovered three identical eight-sided gems across the middle row. The row’s outside prize was 1,000, yet it did not highlight and the completed WIN was 0. This is the rule I would learn before playing: match the designated target, rather than treating any three matching pictures as a win.

Card 61: matching ordinary gems, no target-mask line, payout 0 and net −20.

Two winning lines add together

Card 78 showed masks across the top row and down the right column. The shared top-right mask helped complete both lines. The outside prizes were 20 and 100, and the final display independently confirmed WIN 120: 20 + 100 = 120. With a 20-unit card cost, that was a net gain of 100.

Card 78: top row 20 plus right column 100, final payout 120, Balance 1,400. It was the sample’s largest payout at 6× the card cost.

This is where the layout becomes satisfying: one mask can visibly connect two prize routes. There is no need to import a reel-game multiplier or bonus rule to explain it. The checked interface and complete Help did not document EX, a bonus-buy mode, free spins or Turbo. OPEN ALL is the quick-reveal control for the card already purchased.

Card costs, RTP and the maximum-prize problem

In the tested version, the visible range was 20–1,000 demo units. Each of our settled cards cost 20, with no extra charge recorded for OPEN ALL. A 20-unit payout is therefore break-even; a 40-unit payout produces a 20-unit profit. None of the balances had a verified currency code, so the amounts in this review are not Philippine pesos or cash winnings.

The maximum-prize information needs care. The JILI product page was recorded as showing 1,000×, while the current in-game Help stated a maximum multiplier of 10,567× and a maximum payout of 10,567,000. An older promotional-image note cited 5,000×, but that image was not independently rechecked in this pass. These are conflicting sources, not interchangeable numbers.

The captured Help limits for this version. They differ from the product page’s recorded maximum-win field.

At Bet 20, the eight outside prizes sum to 211,340; dividing that by 20 gives 10,567. That arithmetic agrees with the Help multiplier, but the card headline says 200,000. We did not uncover an all-mask card, and no matching build or regional configuration explained the disagreement. I would check the rules in the exact version being played instead of relying on the promotional headline.

The observed 32% return is calculated from this test’s completed cards only. It is not a published RTP, a forecast or a reason to expect a later recovery. Neither the method of scratching nor the length of a losing run was shown to improve the odds.

Our 100-card demo test

The test ran on 4–5 October 2026 UTC using the official English demo for JILI game 441. It recorded 100 fully settled ordinary cards at 20 units each: 99 used quick reveal, and card 4 supplied the manual-scratching clip. Every completed card is counted once, including that filmed card. The game’s build number and currency were not displayed.

Seventeen cards paid something, but only five made a net profit. Twelve returned exactly the card price and 83 paid nothing. The payout distribution was 83 × 0, 12 × 20, 2 × 40, 2 × 100 and 1 × 120. The longest zero-payout run was 14 cards, from card 10 through card 23. This describes one small sample, not the game’s long-run hit rate or volatility.

Two additional purchases each deducted 20 but never reached a verified final outcome in the saved record. They are excluded from the 100 settled cards, not recorded as losses. Across all 102 known purchases, the cost was 2,040 and the confirmed payout subtotal was 640; the final overall payout, net result and return remain unknown.

Three views of the same settled sample

The balance chart breaks at the session boundaries. Reopening the demo introduced non-wager adjustments of 180 and 200, followed by a zero adjustment; these are not winnings. After card 9 settled at 1,840, the later unresolved purchase deducted 20 to 1,820; reopening then added 180 to reach the new 2,000 baseline. A further 20-unit reset before the completed sample is also excluded. This is why simply subtracting the first wallet balance from the last would give the wrong sample profit.

Observed balance, with separate session baselines and breaks. All points come from the 100-card settled ledger; the two unresolved purchases are excluded.

Cumulative sample return means all confirmed payouts so far divided by all completed-card costs so far. It continues across the session boundaries rather than restarting after each reset. The last point is 640 ÷ 2,000 = 32%.

Cumulative sample return from card 1. The dashed 100% line marks sample break-even; this is not an RTP estimate.

The net chart subtracts the 20-unit cost from each card’s final payout. The tall positive bar is card 78’s +100; the twelve break-even cards sit at zero and the 83 zero-payout cards each show −20.

Net change per card = final payout − actual card cost. No reveal-stage WIN readings or demo resets are counted as payouts.

There were also practical interruptions. One earlier card stopped after a manual scratch attempt and an error, and another remained unsettled after a confirmed debit. The Help includes automatic-completion and malfunction clauses, but those words do not establish that either particular card was paid or refunded. For this review, unknown results stay unknown.

Verdict: readable target lines, limited depth

I would choose Fortune Gems Scratch when I wanted a straightforward card reveal and an immediately readable line award. Its best moment is seeing one mask finish two routes. I am less convinced by the large prize headline and the amount of empty background in landscape. The unresolved demo purchases also keep me from calling this test a trouble-free experience.

Editorial score: 7.0/10. The simple target-line design and clear portrait layout earn it a solid mark for accessibility. The narrow range of mechanics, conflicting maximum-prize information and incomplete reliability evidence stop it going higher. The score is a judgement about the design and the documented experience, not a rating derived from the sample loss.

These are format comparisons, not claims that one title pays better or has safer odds. Start with the demo on this page, read the mask rule and set a firm limit before any paid play. Chasing this sample’s losses would not change the next card’s prospects.

Sources: JILI’s Fortune Gems Scratch product page, the game’s captured English Help, and the 100-card official-demo record dated 4–5 October 2026 UTC. The official demonstration build identifies the tested source; the play invitation above refers to this page’s embedded demo.