Skip to main content
⌖ OF Intel

GUIDES

FFA Rating: The OFstats Ladder That Replaced the Old Ranked Queue

How OpenFront's OFstats FFA Rating works in v36: the 22-rank ladder from Iron III to Grandmaster, the 10-game calibration, the rating-with-uncertainty display, the 8-week UTC seasons, and how to read the win screen, the top-250 leaderboard tabs, and a season reset.

Guides Published Nov 14, 2026 Updated Nov 14, 2026 Reviewed by OpenFront Intel editors #FFA#free-for-all#ladder#rating#ofstats#season#strategy

Direct answer. FFA Rating is OpenFront’s free-for-all ladder, added in v0.36.0 by OFstats. It replaces the old ranked FFA queue: every public FFA game now counts toward a rating shown with an uncertainty, like 1613 ± 85. You finish ten calibration games before placement, then climb a 22-rank ladder — Iron III up to Grandmaster — across eight 8-week seasons that reset at UTC midnight. The win screen shows your rating change, the season dates, and the time left, and the leaderboard opens on your rank with its top 250.

What FFA Rating is and when you should use it

FFA Rating is the ladder for the free-for-all mode, and it is the single biggest change the v36 series made to how competitive FFA players measure themselves. Before v36 there was no public FFA ladder at all: you could play FFA and get a win, but there was no persistent number that tracked your FFA skill the way the 1v1 and 2v2 ranked ladders do. People filled that gap with screenshots of finish positions, with win-rate tallies they kept by hand, and with informal “who is the best public FFA player” threads that never agreed on an answer. FFA Rating closes that gap. The moment you play a public FFA game, that result feeds a rating that OFstats maintains for you, and the number you see after the match is the authoritative measure of your FFA standing, not a screenshot you have to defend in a comment section.

The first thing to internalize is the version boundary, because it changes what you should expect in your client. FFA Rating shipped in v0.36.0 and is present in the current stable build v0.36.2, so it is a released feature with a stable tag behind it, not an unreleased queue. That matters for two reasons. First, every rule in this guide — the 22 ranks, the ten calibration games, the eight 8-week seasons, the top-250 leaderboard tabs, the uncertainty display — is anchored to the v0.36.0 release body and the v0.36.2 tag source, so if a value in your client looks different, the version you are running is the first thing to check, and you should expect v0.36.0-and-newer clients to match this guide. Second, and more important for your day to day: the old ranked FFA queue that the community asked for for years is gone. It has been replaced by this OFstats system. So if you were following an older guide or a “what is the ranked FFA lobby contract” writeup that describes a four-to-ten-player ranked queue on a six-map pool, that is the wrong model now. The current system is not a fixed-size queue you matchmake into; it is a rating that every public FFA game contributes to, with a placement after ten calibration games and a leaderboard you can browse by rank tab. Read this guide as the current source of truth for how FFA competition works, and read the older ranked-FFA material as the retired design that motivated it. The practical question — “how do I climb FFA now?” — is answered by the sections below, and the short version is: play public FFA, finish your ten calibration games, and then treat your rating and the leaderboard as the thing you are climbing.

How the rating itself works, and what the number means

The rating is the number you see on the win screen and in your game history, and it is always shown with an uncertainty rather than as a single fixed value. A typical display is 1613 ± 85: the 1613 is the point estimate of your skill, and the ± 85 is how uncertain OFstats currently is about that estimate. The ± is not decoration; it is doing real work. Early in your FFA history the system has played you against a limited sample, so the margin around your number is wide — your true skill could plausibly be meaningfully higher or lower than the point estimate. As you play more public FFA games, the system accumulates evidence about how you actually perform, and the margin narrows. A player who has finished hundreds of public FFA games will show a small ±, while a player in their first week will show a large one. So the number you should actually watch is the pair: the point estimate tells you where you sit, and the ± tells you how much to trust that position this week. A rating of 1613 ± 85 is a different statement than 1613 ± 12, even though the point estimate is identical, because the second one means the system is far more confident about where you are.

Three related displays complete the picture, and all of them appear on the win screen. The first is your rating change for the game you just finished — the difference between your rating before the match and your rating after it — shown as a signed number so you can see at a glance whether the game moved you up or down. The second is the expected value, which the interface renders as a percentage: it is the share of the outcome the model expected you to take before the match started, so it tells you whether the result was a good or a bad one relative to what your rating predicted. The third is the season context — the season dates and the remaining time in the current season — so you always know how much of the current ladder window is left. The way these three work together is the point of the system. A win against a higher-rated field should move your number up by more than a win against a lower-rated field, a loss against a much weaker field should hurt less than a loss against a much stronger field, and the ± should shrink as the sample grows. That is the feedback loop you are climbing on: every public FFA game is a data point that nudges your estimate and tightens your uncertainty, and the win screen is where you read the nudge. The practical habit is to look at the change and the ± after every match, not just the win/loss, because the rating change is the actual score of the game in this system.

The 22-rank ladder: every tier from Iron III to Grandmaster

The ladder has eight tier bases in fixed order from lowest to highest: Iron, Bronze, Silver, Gold, Platinum, Diamond, Master, and Grandmaster. The first seven bases are each split into three divisions labelled I, II, and III, and Grandmaster is the single top tier with no divisions. That gives 7 × 3 = 21 divided ranks plus Grandmaster for a total of 22 distinct ranks, and the source assigns them integer rank numbers from 0 to 21, with Iron III at 0 and Grandmaster at 21. Within each divided base the divisions are ordered III, then II, then I — so Iron III is the lowest rank of all, Iron II is above it, Iron I is above that, and then Bronze III begins the next base. You move up one division at a time, and when you clear the I division of a base you step into the III division of the next base up. So the full path from the bottom is Iron III, Iron II, Iron I, Bronze III, Bronze II, Bronze I, Silver III, Silver II, Silver I, Gold III, Gold II, Gold I, Platinum III, Platinum II, Platinum I, Diamond III, Diamond II, Diamond I, Master III, Master II, Master I, and finally Grandmaster.

The table below is the complete ladder with the rank number the source assigns to each tier, which is the same ordering the leaderboard tabs use:

RankTierRankTier
0Iron III11Gold I
1Iron II12Platinum III
2Iron I13Platinum II
3Bronze III14Platinum I
4Bronze II15Diamond III
5Bronze I16Diamond II
6Silver III17Diamond I
7Silver II18Master III
8Silver I19Master II
9Gold III20Master I
10Gold II21Grandmaster

The two design choices worth understanding are why the divisions exist and why Grandmaster is separate. The divisions exist so that the ladder has enough distinct steps to separate players without needing a long list of tier names: three sub-steps inside each base means you can show progress within a tier before you cross into the next metal. Grandmaster is separated because it is the cap of the ladder — there is no “Grandmaster I” or “Grandmaster II,” so the top of the climb is a single, unambiguous rank. The practical meaning is that your tier label is not a free-form string the interface guesses at; it is parsed into a base and a division, and the rank number is computed from that parse. That is also why the leaderboard can open directly on your rank: it knows your exact base and division, so it can jump to the right tab and the right row. If you see a tier you do not recognize in the list, the system does not fabricate a rank for it — an unparseable tier yields no rank — so the eight bases and their divisions are the complete set of labels you will ever see on a v0.36 ladder.

The 10-game calibration and how placement works

When you are new to FFA Rating — or when a new season starts and the system needs to re-establish your position — it does not assign you a tier immediately. You must finish ten calibration games before OFstats gives you a tier, and that ten-game number is a fixed contract in the source, not a suggestion. The purpose of the calibration window is to let the system collect enough evidence about your FFA skill to place you at the rank that best matches your demonstrated performance, rather than starting you at an arbitrary floor and letting you grind up. During those ten games your rating is still moving and the uncertainty around it is still large, so the number you see mid-calibration is a moving estimate, not your final placement. Your placement is the tier you are shown after the tenth calibration game finishes, and that is the rank your ladder climb begins from.

There are a few practical things to expect from the calibration window. First, the ten games are public FFA games, so you are playing real FFA matches against real players, not a separate practice queue — the calibration is just the system watching your real play. Second, the uncertainty is widest during calibration, so the ± around your rating is at its largest in the first several games and narrows as the ten complete and as you keep playing. Third, because your placement is determined by what the system observes across the ten games, your goal during calibration is not to win every match — it is to play cleanly so the system measures your actual skill. A fluke win streak does not place you higher than your real level, and a bad start does not lock you into a low rank, because the estimate is updated game by game rather than set once. The honest read is that calibration is the system’s way of finding your starting rank efficiently, and the ten games are the minimum sample it needs to do that. You do not have a separate calibration queue to find and you do not pay a special entry fee for it — the ten games are simply your next ten public FFA matches, which is what makes the window cheap to survive without changing how you already play. The practical habit is to treat the first ten games as a data-gathering phase: play them with your normal FFA routine, watch how the rating and the ± move, and expect your placement tier to appear once the tenth game is done. From there you are on the ladder, and the climb is governed by the rating change and the uncertainty you see on every subsequent win screen.

Season resets: the 8-week cycle and the UTC midnight boundary

FFA Rating runs on seasons, and each season lasts eight weeks. The interface shows the season dates and the remaining time in the current season, so you can see at a glance when the next reset is due. The reset itself happens at UTC midnight, which is the fixed boundary the system uses to end one season and start the next. That boundary is in UTC, not in your local timezone, so you should work out what UTC midnight is in your own timezone before you plan anything around the reset. For a player in a UTC+3 timezone, UTC midnight is 3 a.m. local; for a player in a UTC-5 timezone, it is 7 p.m. local the previous day. The practical point is that the season does not end at “midnight” on your clock — it ends at midnight in UTC, and the two can differ by hours, so checking the season countdown in the interface is the reliable way to know exactly how much time is left rather than guessing from your local date.

What changes at a reset is the state of your rating and the state of the leaderboard. Your rating is re-baselined for the new season, the uncertainty around it resets toward a wider margin (because the system is starting a fresh estimate for the new season’s field), and the top-250 lists in each rank tab are repopulated for the new season. That last part is important for how you read the board after a reset: immediately after a season turns, many rows are hidden because their uncertainty is too high to publish, and the board visibly repopulates as players play public FFA games and the system narrows their uncertainty. This is the hidden-by-uncertainty rule in action, and it is the reason a freshly-reset leaderboard looks sparser than a mid-season one. Two habits make the reset less disorienting. First, before a reset ends, play a few public FFA games while the season is still active so that your standing heading into the boundary is based on recent play rather than on matches from weeks ago. Second, after the reset, expect your placement to shift and your board to be incomplete for a while, and treat the first week of the new season as a re-calibration period where your rating and the top lists are still settling. The eight-week length is long enough that you have a real arc to climb within a season — enough time to move several divisions or a full base — but short enough that the ladder does not become a permanent fixture that never resets, so the top of each season is a fresh field of players competing for the top 250 of their rank.

The win screen, game history, and the leaderboard tabs

The win screen is the first place you read your result, and FFA Rating gives it a distinct layout because it has to show both your match outcome and your rating movement at once. The screen separates your result from your placement on the ladder, and it shows the rating change for the game, the expected value as a percentage, and the season context — the dates and the remaining time — so you can see the match outcome and where it put you without leaving the screen. Game history carries the same information across all of your public FFA matches, so you can scroll back through recent games and watch how your rating and the ± have moved over a stretch of play rather than relying on the single number from the last match. The history view is the best place to judge whether you are actually climbing over a week or a month, because a single game’s rating change is noisy and a streak of changes is the real signal.

The leaderboard is the second place you read your standing, and it is organized as a set of rank tabs rather than one flat list. There are eight tabs, one per tier base, from Iron through Grandmaster, and each tab shows the top 250 players in that base for the current season. The board opens directly on your rank — the tab for the base you are currently in — and it pins your position within that list, so you can see where you sit against the other players in your base without scrolling. Two behaviours are worth knowing. First, because each tab shows only the top 250, a base with more than 250 rated players does not show everyone; you are looking at the head of your base, and your pinned position tells you whether you are inside that head. Second, the board hides rows whose uncertainty is too high to publish, which is the same hidden-by-uncertainty rule you saw at a season reset: a player whose rating the system is not yet confident about does not appear on the board until their uncertainty narrows. The practical read of the leaderboard is therefore: the tab you open on is your base, the pinned row is you, the top 250 is the visible field in that base, and any gaps you see right after a reset are players still waiting for their uncertainty to narrow, not missing data. Together the win screen, the history, and the tabbed leaderboard are the three surfaces you will actually use to track your FFA climb, and the win screen is the per-match read while the history and the leaderboard are the longer-horizon reads.

Two decision frameworks: calibration play and mid-season recovery

The two decisions that matter most in FFA Rating are what to do during the ten calibration games and what to do when you drop a division mid-season. Both are about protecting your rating, and both have a concrete plan with numbers.

Framework A: playing the calibration window well. Your goal in the first ten games is to let the system measure your real skill, so the placement you get is a fair starting rank rather than a lucky or unlucky one. The concrete plan: in the first few games, play your standard public FFA routine — the same map reading, the same economy pacing, the same endgame decision you would make in any public FFA — so the system sees your normal performance, not a cautious or an over-aggressive version of you. The number that matters here is the uncertainty: you want the ± to start narrowing across the ten games, which is the signal that the system is building a stable estimate. If you win early and the rating jumps, do not change your play to “protect” the number, because a changed playstyle is exactly what corrupts the estimate; keep playing normally and let the ten games complete. If you lose early, do not tilt or change your approach either, because the estimate updates game by game and a bad first game is not your placement. By the tenth game the placement tier appears, and the right response is to note the rank and the ± together: a placement with a still-wide ± means the system will keep adjusting for a while, so the first few games after placement are where you confirm whether the starting rank was fair. The scenario: you finish your ten calibration games and are placed at Gold II with a ± of about 70. That is a solid mid-ladder placement, and the wide margin means the next several games will refine it. Your plan is to play three or four more public FFA games with your normal routine and watch whether the point estimate holds around the Gold II value or drifts; if it holds, Gold II is your real base and you climb from there, and if it drifts up or down, the system is still settling and you should not plan a long arc around a single placement number.

Framework B: recovering after you drop a division. Dropping from, say, Gold I to Gold III is a two-division slip, and the recovery is about getting the rating back without burning games. The concrete plan: after the drop, check the expected value on your recent losses to see whether you were losing to stronger fields (which is normal) or to weaker fields (which is the real problem). The number that matters is the expected percentage on a loss — if you lost a game the model expected you to win by a wide margin, that is a data point worth reviewing, because it means you underperformed your rating; if you lost to a higher-rated field, the loss is consistent with your rating and the drop is the system working, not you collapsing. The recovery habit is to play a short, focused block of public FFA games — five to eight — with the goal of winning against fields at or below your rating, which is where the rating change is most favourable, and to bank Gold and keep your land connected so that the losses in that block are few. The scenario: you are at Gold I and drop to Gold III after three games. You check the expected values and see that two of the three losses were against higher-rated fields (expected 35% and 40%) and one was against a lower field (expected 65%). The two losses against higher fields are normal and the rating drop for them is modest; the one against the lower field is the real cost. Your recovery plan is to play six games targeting mid-to-lower fields, keep your core land connected so you are not the one breaking, and aim for four or more wins in the block; that combination of wins against at-or-below-rating fields plus a few more games is what moves you back toward Gold II and then Gold I, because each favourable win nudges the point estimate up and the narrowing ± confirms the recovery is real.

Failure modes and the counterplay for each

Players new to FFA Rating make the same four mistakes, and each has a concrete counterplay. The failures are not “I lost a game” — they are specific reading errors about the rating, the uncertainty, and the season, and they are all avoidable.

Failure 1: reading the point estimate and ignoring the ±. This is the most common first mistake. A player sees 1613 and treats it as a fixed, known rank, then is confused when the number moves a lot over the next few games. The ± is the whole reason the number moves, and ignoring it is how you misread your own standing. The counterplay is the “read the pair” rule: always look at the point estimate and the ± together. A wide ± means the system is still uncertain and the number will move a lot; a narrow ± means the system is confident and the number will move little. When your ± is wide, do not plan a long arc around the current point estimate, and when it is narrow, trust the estimate more. The mental reframe: the rating is a range, not a point, and the ± is the width of that range.

Failure 2: assuming the board is complete after a season reset. Right after a reset, the leaderboard is sparser than usual because rows with too-high uncertainty are hidden, and a player who reads the sparse board as “everyone else is weak” or “the top 250 is empty” makes a wrong call. The counterplay is the “reset is a re-calibration” rule: for the first week after a reset, expect hidden rows and a shifting top 250, and do not compare your position to the mid-season board. The gaps are players waiting for their uncertainty to narrow, not missing competitors, so the board repopulates as the season’s games are played.

Failure 3: changing your playstyle during calibration to chase a higher placement. This corrupts the estimate, because the system is trying to measure your normal skill and a changed playstyle gives it the wrong data. The counterplay is the “play normally through ten games” rule: your job in calibration is to perform your real routine, not to optimise the number. A cautious first streak or an over-aggressive tilt both place you at the wrong rank, and the estimate takes games to correct, so the cheap fix is to just play your normal FFA from game one to game ten.

Failure 4: planning around the wrong midnight. Because the season resets at UTC midnight, a player who plans around their local midnight ends up a season-boundary off, and that matters for both the pre-reset habit and the post-reset re-calibration. The counterplay is the “UTC is the boundary” rule: convert UTC midnight into your local time before you plan, and use the season countdown in the interface as the authoritative remaining time. The common thread across all four failures is that FFA Rating is a statistical system and the mistakes are all reading errors about what the system is and is not telling you. The counterplay to all four is the same habit: read the pair, respect the uncertainty, trust the season boundary, and play your real game.

Mode and map adjustments, and how to read your ladder

The final layer is adaptation: how your FFA play should flex by map and by the season phase you are in, and how to read your climb now that FFA has a real ladder. The map adjustments are where FFA knowledge stops being generic and starts being a decision tool, and the ladder-reading is where the mode’s whole purpose becomes concrete.

Map adjustments. Public FFA spans the game’s full map pool, including the new Atlas maps v36 added, so the map you land on sets the tempo of the FFA match and, by extension, the quality of the rating data point it produces. On a high-land, compact map — a continent where most of the territory is solid and connected — the match is a spread-and-accumulate game: you place your first city on the widest bridge, spread into the biggest pockets, and play toward holding the most connected land. On a water-heavy map with internal seas and broken archipelagos, the match is a pinch-and-eliminate game: you own the land bridges early, target the smaller players in sequence, and force eliminations so the field shrinks. The adjustment that matters for the rating specifically is this: a rating data point is most informative when your play on the map is representative of your real skill, so on a map you know well you produce a cleaner signal than on a map you are learning. The practical habit is to watch how you perform across different map families and note where your win rate and your rating change diverge — if you are consistently worse on water-heavy maps, that is a real gap the rating is measuring, and closing it is the fastest way to move the point estimate up.

Season-phase adjustments. The eight-week season has a natural shape, and your play should flex to the phase. In the first week after a reset the board is re-calibrating and the top 250 is shifting, so the goal is to play enough public FFA games to establish your starting standing and to let your uncertainty narrow — quantity matters more than a single perfect game here. In the middle of the season the board is stable and the top 250 is set, so the goal is to climb within your base: play at-or-below-rating fields to bank wins, keep your land connected, and push toward the next division. In the final stretch of the season the goal shifts to consolidating: if you are near the top of your base’s top 250, protect that position; if you are below it, use the last games to close the gap, because the next season’s board is a fresh field and a strong finish is the best heading into the reset. The habit to build across the season is the same as the other ranked ladders: treat each public FFA game as a data point, read the rating change and the ± after every match, and track your rank over the whole season rather than over a single game.

Reading your ladder. The last adjustment is the one that changes why you play: your FFA results are now a real, persistent ranking, which is the gap the community kept pointing at for years. The practical change is that you can climb FFA the same way you climb 1v1 and 2v2 — by your rating, not by a screenshot — and you can measure your FFA progress the same way, by watching the rank number and the ± move across the season. The mode is new and the ladder is still building, so the top of each base is dominated by the players who have been the best at public FFA, because they have the map knowledge and the strategy already. If you have been a strong public FFA player, this is your home: the clean, skill-sorted version of the mode you already play. For the shared ladder rules that apply to all ranked modes — the ELO-style rating, the free-match budget, and the rating protection — see Ranked Mode & Rating System. For how to read the win conditions across every mode, see Win Conditions. For the midgame habits that carry over from public FFA into this rated context, see FFA Midgame Strategy.

Related content

Continue reading
Adaptive Build Order: Spend the Next Gold on the Constraint

Use OpenFront v34.3 costs, troop growth, route payback, and stop signals to choose City, Port, Factory, defense, or a liquid reserve without copying a fixed script.