Metin2 Okey strategy: points, card tracking and the 400-point goal

In the Gameforge Okey rules documented here, 300 points earns silver and 400 points earns gold. A same-colour 6–7–8 scores 100, but the best immediate combination can still leave you with a worse final score. Track the cards already used, then compare the points you can make from the cards that remain.

This guide uses the Gameforge EN scoring table, cross-checked against the German rules and Romanian rules and examples on 2 November 2026. It is a strategy guide for that ruleset, not an announcement that the event is currently running. A private server's Okey variant needs its own published rules before these values can be applied.

Generated illustration based on Metin2 game references. It is not a game screenshot; the exact card examples are in the text.

The deck has 24 different cards: red, blue and yellow, each with values 1–8. You hold up to five cards and score with three at a time. A run must have consecutive values without a gap. Two cards may share a colour in a mixed run; all three must share it to get the same-colour bonus.

There is no 9 in this deck, so 7–8–9 is not a combination to wait for. Equally, 5–6–8 has a gap and scores nothing. The colour bonus applies to a valid run; three unrelated cards of one colour do not score merely because their colours match.

For chest tiers, 0–299 is bronze, 300–399 silver and 400 or more gold in this documented model. That is the score of one round. Do not confuse the 24 collectable items needed for a card set with 24 separate rounds.

Open the active event through its event interface, start with the required card set and fee shown there, then draw your first five cards. Select a valid group of three to score. Keep the unused cards and draw replacements from the remaining deck. Right-clicking discards a card permanently; use the confirmation setting if you are prone to misclicks. The official rules allow ending the round with End; before doing that, count the points still available.

A practical record has three states for every coloured card: still held, already played/discarded, not yet drawn. After each action, update it before considering another move. A used red 7 will not return as a replacement red 7. Blue 7 and yellow 7 are different cards and must be recorded separately.

For example, if red 7 was discarded while red 6 and red 8 remain in your hand, the red 6–7–8 plan is dead. That does not make both held cards worthless: they might still fit other valid groups. Check those alternatives before discarding again.

Players in this Okey discussion ask why card tracking does not reliably produce the chest rates other people claim. Its replies suggest remembering cards and using helpers; they do not verify a 60% success rate. The Turkish forum explanation also describes combinations and reward tiers. A strategy still needs to account for the actual hand, not just repeat “keep high cards”.

Here is a constructed, fully specified endgame, not a recorded game session. Use these card labels: R=red, B=blue, Y=yellow.

You have 240 points and hold R5 R6 R7 B7 Y7. There is exactly one undrawn card, R4. This certainty requires an accurate record of every other card; it is not a prediction about an unknown deck. One possible earlier history is three same-colour 1–2–3 runs plus the triple 8:50+50+50+90=240, with the six unused blue/yellow 4–6 cards discarded. All 24 cards are accounted for.

The 90-point move consumes the red 7 that belongs to the three-colour triple. The 80-point move instead preserves red 5 and red 6 for the known final red 4. That is the decision the single-move score misses.

We checked every legal play and discard in this particular known-deck state with a small exhaustive model. 160 additional points is its best result. This does not prove an optimal policy for a whole round with unknown draws, and the example does not mean a player can force this hand to appear.

At 290 points, a mixed 1–2–3 is worth 10 and reaches 300. Ending immediately instead leaves the round below silver. At 390, the same 10 points reaches 400. So “only mixed runs remain” is not, by itself, a reason to end the round.

There is another useful arithmetic check: two disjoint mixed 6–7–8 runs can score 60+60=120. For instance, R6 B7 Y8 and B6 Y7 R8 share no card. A 100-point same-colour run therefore does not automatically outscore every pair of mixed combinations.

Earlier in a round, a small combination may consume cards that could later make something better. Near the end, it may be exactly what reaches the target. First ask how many more points you need, then which legal groups remain possible. Neither “always play the highest score” nor “always discard a weak run” answers both questions.

The following Python 3 example uses only the standard library. R/B/Y and values 1–8 identify unique cards; invalid input or a non-scoring group returns 0. It checks the two worked branches without connecting to Metin2, reading a client or making game moves.

from itertools import product

deck = {f"{c}{n}" for c, n in product("RBY", range(1, 9))}

def score(cards):
    if len(cards) != 3 or len(set(cards)) != 3:
        return 0
    if any(c not in deck for c in cards):
        return 0
    values = sorted(int(c[1]) for c in cards)
    if values[0] == values[2]:
        return 10 * values[0] + 10
    if values != list(range(values[0], values[0] + 3)):
        return 0
    bonus = 40 if len({c[0] for c in cards}) == 1 else 0
    return 10 * values[0] + bonus

first = score(("R5", "R6", "R7")) + score(("B7", "Y7", "R4"))
second = score(("R7", "B7", "Y7")) + score(("R4", "R5", "R6"))
assert (first, second) == (90, 160)
print(240 + first, 240 + second)

The output is 330 400. Separately, our validation checked all 2024 unique three-card selections from the deck: 8 same-value triples, 18 same-colour runs and 144 mixed runs score; the other 1854 do not. These are combinatorial checks against the stated rules, not measured chest odds or an in-game test.

A manual tracker can reduce memory mistakes, but a recommendation depends on the cards entered and the model used. An accidentally marked colour, a forgotten discard or a different server ruleset changes the result. A calculated probability is conditional on its assumptions; it does not reveal the next shuffled card.

Before following advice, check whether it respects values 1–8, preserves used-card history, scores two-colour runs correctly and distinguishes immediate points from the final target. The worked example above is a useful sanity check for the last distinction. It is not a certification of any third-party helper, a blanket permission to use tools on every server, or a guarantee of gold.

For a private server, read that operator's event rules and tool policy first. The Metin2 private-server directory is a starting point for finding the relevant project; a shared event name does not establish matching mechanics. For the official event's next dates and entry conditions, use the current in-game announcement or the regional Gameforge notice.

• Gameforge EN Okey rules and points, preserved revision 61963.
• Gameforge DE scoring and chest thresholds.
• Gameforge RO worked combinations and edition information.
• Player questions and replies on Reddit and Gamer.com.tr.

Sources were checked on 2 November 2026. We executed the score checks and the constructed endgame model. No game-client integration, live Okey session or real-world win-rate measurement was performed.