What has changed

A VFR aircraft you put in a hold can be released again.

Telling a VFR arrival to hold worked; telling it to resume own navigation did not, and neither did a landing clearance, a straight-in or an approach. The aircraft orbited and every command bounced. Resume own navigation now hands a VFR pilot back their own flying, which is what the instruction means: they were never on a filed procedure to rejoin, and that was exactly why the release was being refused. Reported by a player who sat with one circling for an hour.

Reported by a player

Readbacks are no longer spoken long after the clearance they answer.

On a machine where the voices are slow to generate, the radio could fall far enough behind that a readback arrived more than half a minute after you issued the clearance, describing a picture that had already moved on. The radio was supposed to drop audio that had gone stale rather than speak it late, and it checked for that in only one of the two places it mattered. It now checks again at the moment it would speak. The words were always on the strip the instant the clearance was issued, so nothing is lost but a voice that had stopped being true. Reported by a player.

Reported by a player

An aircraft now holds the altitude you assigned until it is established.

The approach clearance has always said "maintain three thousand until established" and the aircraft then ignored it, descending to the same height as everything else. That made parallel approaches nearly impossible to separate, because every arrival ended up at one altitude no matter what you told it. Now it holds what the clearance named, so two approaches can be stacked, and the glidepath takes over once the aircraft is established. Reported by a player working the St. Louis parallels.

Reported by a player

Working approach only, you can no longer clear an aircraft to land.

That is tower's job, and the approach position exists to separate the two. The game let you do it anyway and the menu told you to, which meant the one seat built to draw that line did not draw it. Now the landing clearance is not offered on that seat and the aircraft goes to tower with F, which is the ending that already existed. Nothing changes on the normal seat, where you work the runway and the radar together.

Working approach, departures no longer call you holding short of a runway you do not control.

Tower launches the departures; on the approach-only seat they should reach you already airborne, and most did. Light aircraft did not. They appeared on the runway calling holding short, waiting on a takeoff clearance from a controller who does not own the ground, and the only way to move them was to issue one. Reported from the field, and it happened at eight of the thirteen approach fields but at neither of the two the tests were checking, which is why it took a player to find it.

Reported by a player

The command panel is shorter, and it leads with what the aircraft needs next.

Selecting an airborne arrival put 61 buttons on screen. It is 36 now, and nothing was taken away. Hold at a fix hid its nineteen buttons behind an expander, the way direct to fix already did, and the two are one Hold section instead of two headings for the same decision. Speed offers four common values with a note that any speed can be typed, the same way the heading grid already worked. And for an arrival the approach and landing clearances now come first rather than sitting under three grids of vectoring controls, which is the order the phone layout already used.

The fix lists are in alphabetical order.

Hold at fix and direct to fix both listed the field's fixes in the order the data happened to carry them, which reads as no order at all: with forty buttons you were scanning the whole grid for a name you already knew. Both are A to Z now. Both, not just the hold list that was reported, because two grids of the same fixes disagreeing about where a name lives would be worse than neither being sorted.

Working approach, you can now actually send an arrival to the tower.

On the approach seat an arrival that established on the ILS went ready for the tower, flashed its badge and asked for the frequency change on the radio, and there was no way to give it: the menu offered no button and typing F was refused. The handoff only happened because a timer eventually did it for you. The badge now reads TWR F rather than RDY F, the command menu offers Contact Tower, and F works. The shift header also says which seat you are working, so a remembered APPROACH ONLY that does not apply to a Ranked shift is visible rather than a surprise.

US spelling throughout, and a check that keeps it that way.

The changelog, the command reference, the in-game help and the per-airport pages had drifted into a mix of US and British spelling, in a few places carrying both spellings of the same word on the same page. Fifty-three of them are corrected. A check now runs on every build over everything a player reads. It leaves published names alone, which matters more than it sounds: one of the charted fixes at Hutchinson is called SPELT, and correcting it would invent a waypoint that does not exist.

An approach clearance can no longer capture the parallel runway.

Clearing an arrival for one runway of a parallel pair could put it on the localizer for its twin. The aircraft flew a correct-looking approach to the wrong strip of pavement, which cost at least one player a separation loss and a scored session: the game punished somebody for its own mistake. The clearance now holds the runway it names, and a check runs the old failing scenarios on every build so it cannot come back quietly.

Traffic calls named the aircraft ahead as UNDEFINED.

A follow instruction could go out on frequency describing the traffic ahead as undefined when the aircraft type was not one the phrase knew how to say. It falls back to naming it as traffic now. The command menu also stopped scrolling sideways on narrow windows, and the WHY? explanation now appears whenever there is a note to read rather than only when something has gone wrong.

The briefing's first line was spoken as a single word.

A change to how transmissions are trimmed for length made the first line of the turnover briefing come out as one word, which sounded like the audio had failed to start. It reads in full again. Installing the game to a home screen also works now, which lets the browser keep the voice model between visits instead of discarding it, and the copy no longer promises a download that was never going to happen.

The data-block keys are in the help panel now, and stop stacking messages.

Walking a data block around its target left one message on screen per press, five of them piling up over the scope. It is one running confirmation now, replacing itself as the block moves. The keys were also missing from the in-game help: that list only showed keys that stage a command into the bar, and these act on the scope instead, so they appeared on the website and nowhere in the game. There is a scope keys line under ? now.

Moving a data block did not work if you selected the aircraft by clicking.

Clicking an aircraft puts its callsign in the command bar and moves the cursor there, ready for you to type. The comma and period keys were being ignored whenever the cursor was in that bar, so the two things you had to do were in each other's way: you could not move a block without selecting an aircraft, and selecting one stopped the keys working. The period went into the command bar instead. They now move the block wherever the cursor is, and never end up typed into a command. Reported by a player within hours of it shipping.

Reported by a player

Nine fields never told you which departures were in use.

Departure procedures that serve every runway at a field are published with a marker meaning exactly that. The briefing understood a procedure tied to one specific runway and did not understand that marker, so it read those procedures as serving no runway at all and left them out. Kansas City lost all six of its departures that way, Scott all four, nine fields in total: the briefing named the arrivals and went quiet, as though the field published no departures. The aircraft were flying them the whole time, which is why it went unnoticed for so long. All nine now name their departures, and Saint Louis picked up four it had also been dropping.

Southwest was flying under the wrong callsign at Wichita and Tulsa.

The scope showed WN3649 at Wichita and SWA3649 at St. Louis for the same airline: WN is Southwest's ticketing code, SWA is the one used on the radio. Corrected at both fields, and Flexjet's name is now spelled the same way everywhere.

Flying above a departing heavy is no longer called a wake conflict.

Wake turbulence sinks: the vortices trail behind an aircraft and settle below its flight path, which is why the standard technique is to stay at or above a heavy's track. The check ignored that and measured the altitude gap in either direction, so a light aircraft overflying a heavy that was climbing out beneath it was treated exactly like one following underneath, and that could end a shift on a conflict which cannot physically happen. Wake spacing now applies only when you are at or below the aircraft ahead. Below and behind a heavy it is unchanged, and it starts applying again the moment a climbing departure reaches your altitude.

The lobby map: half the airports could not be clicked.

Twenty of the forty fields handed your click to a neighbor when you clicked them directly — St. Louis went to Cahokia, Kansas City to Topeka — because the invisible target and the pin both grew with the map, so zooming in could never separate a cluster. Pins and targets now hold their size on screen, a click resolves to the nearest field rather than to whichever was drawn last, dragging the map pans it again, and the wheel scrolls the page until you deliberately engage the map.

Reported by a player

Two pilots stopped refusing instructions they should have accepted.

A departure whose filed cruise altitude sat below a crossing restriction on its own departure procedure announced it was unable to make the crossing before it had even left the ground, with nothing you could do about it. Separately, a VFR arrival that had asked for the pattern was offered runway-entry buttons that the tower then refused as "not a VFR arrival". Both are fixed.

The shift briefing fits on screen without scrolling.

Taking the position meant scrolling past a paragraph that read the same at every field. The generic description is gone (it lives in Help), the briefing is tighter, and the busiest opening still fits above the button.

Established arrivals now accept realistic final-approach speed assignments.

Jets used to reject every assigned speed at or below the speed they were already flying on final. The higher unconfigured speed floor still applies while they are being vectored, then drops to the aircraft's configured approach range once it is established, so controllers can build spacing without asking an airliner to fly faster than it already is.

Approach charts stop at the runway instead of running past it.

Open the charts panel on an approach and the drawn route carried on beyond the runway to a fix miles the other side of the field — Springfield's RNAV 13 ran fourteen miles past the airport. That tail is the missed approach, which the published data includes and which we were drawing as part of the approach itself. Every approach now ends at its runway. What aircraft actually flew never included it, so only the chart was wrong.

Springfield's crossing runway is where it actually is.

Runway 13/31 at Abraham Lincoln Capital sat about a thousand feet east of its real position, so both RNAV approaches to it aimed a few hundred feet to the side of the pavement you could see. One threshold had been typed in with a wrong longitude and the other was calculated from it, which made the runway the right length and the right heading while both ends were wrong together. Both ends now come from the published survey.

An approach you cannot make from where you are is flown as a join, not a botched cut-in.

Clearing an arrival for the approach from too far beside the final course used to be flown as a direct cut toward the centerline, with far too little course left to reach it, so the aircraft crept toward the runway and arrived beside it instead of on it. The pilot now recognises that the intercept cannot close in the distance remaining and builds the geometry first: it flies out to a join point on the final course at the right energy and joins there, showing JOIN on the data block until it is genuinely established. The clearance is still accepted. Giving the aircraft a workable intercept angle is still the controller's job, and a shorter, cleaner vector is still the better answer.

The radio no longer falls a minute behind you.

On a busy board, if your machine could not generate speech as fast as the transmissions arrived, the backlog grew and every line was still eventually spoken — so pilots read back instructions you had given a minute earlier, describing a picture that had already moved on. A transmission that has waited longer than the frequency could plausibly have been busy is now dropped instead of spoken late, and the time that would have gone into generating it goes to catching up. Nothing is lost: the readback text is already on the strip the moment it is issued. An emergency is never dropped, however long it has waited. Reported by a player at Tulsa.

Runway numbers no longer print on top of each other at parallel fields.

On any field with a bank of parallel runways, the numbers at the thresholds fused into one unreadable smear when you zoomed out to the sector view. Seattle's three sixteens came out as "16R16L16C"; St. Louis ran 30L and 30R together; Chicago and Atlanta were worse. The numbers now lie along the pavement the way they do on a chart, and a parallel bank fans out just far enough to separate, in the same left-to-right order as the runways themselves, so the leftmost number belongs to the leftmost strip. Zoomed in close, where they were never crowded, nothing moves. The approach labels on the extended centrelines know where the runway numbers are now too, so the two sets stop crossing when you have an arrival selected.

Browsing the airport list can no longer cost you a record.

The server counted every page view and every leaderboard you opened against the same small budget it used to accept a new high score. Open enough airport cards from one office or phone network and the next record submitted from that network was rejected, silently, because the archive is sent without waiting for an answer. Reading and submitting now have separate budgets, reading got a much larger one, and the boards are cacheable so a busy day is absorbed before it reaches the server at all.

Approach labels on parallel runways are readable again.

Select an arrival and each landing runway names its approach out at ten miles. On a field with parallels those names printed on top of each other: Louisville showed "ILS 17R" and "ILS 17L" as one unreadable smear, and Seattle's three 16s were worse. Each label now runs along its own extended centerline the way a chart prints it, spaced so no two collide, and a true parallel bank reads as one label — "ILS 16L/C/R". Runways that only look like a bank are never folded together: St. Louis still shows 29 as its own approach, because 29 is the crosswind and does not point where 30L and 30R do.

Louisville's Ohio River is drawn from federal survey data now.

The Ohio arrived from a world-scale map dataset and it showed: too wide, too blocky, and folded over itself west of the field. It now comes from the USGS National Hydrography Dataset, the same survey the government uses, so the meanders and the big horseshoe bend upstream are where they actually are. It was the first river to be sourced this way; the rest followed the same day.

An IFR go-around can no longer be reversed at the runway threshold.

Once the pilot begins the missed segment, another landing clearance waits until the aircraft has flown clear for a new approach. VFR aircraft retain their visual resequencing behavior.

The Missouri is one river from Kansas City to the Mississippi.

Ten fields draw the Missouri and they used to draw it from three separate traces that disagreed by up to sixteen miles where they met. Jefferson City and Columbia disagreed by twenty-three. There is one river now, and one Arkansas behind Wichita and Tulsa, so the water is in the same place on every scope that shows it. Nothing about a field you work changes; the map is simply telling the truth in more places.

Bloomington-Normal now gives you a board you can actually work.

Central Illinois Regional shipped with too many aircraft allowed on frequency at once for a field with crossing runways and no published departure procedures. In testing it saturated and fell over before anything reached the ground. The limit is lower now, and the field runs the way the rest of the Steady tier does.

The Missouri River is now the same river on every scope that shows it.

Kansas City International, Downtown, Johnson County and New Century each drew their own Missouri and their own Kansas River, and the four did not line up: fly one field, then another, and the water had moved by up to two and a half miles. Columbia Regional was worse, with the Missouri drawn on the wrong side of the field, more than twenty miles from where Jefferson City drew the same stretch. All six fields now take that water from one shared map, placed through each field's own reference point, so a bend you learn at one airport is in the same place at the next one. Hillsdale Lake, which Johnson County and New Century both show, is shared the same way.

Your longest shifts were never being submitted at all.

A shift was only ever sent to the records when it beat that field's SCORE. Your longest shift is very often not your highest-scoring one, so the long ones simply went nowhere — there was no record of them to build a board from. A shift is now submitted when it beats either the field's score or the field's longest shift, and a shift that beats both is still filed once, as one result.

Your record is now the score you FINISHED a shift on, not the highest number you passed through.

The score moves in both directions, because a deal, a proximity bust, a forced go-around and a failed emergency all take points back. So the highest number a shift ever showed and the number it ended on are different results. The HIGH beside your score was keeping the highest, which is a score nobody ever finished a shift with, and it disagreed with the leaderboard, which has always recorded the final score. Two things went with that: a later run that was genuinely better than the one on the board set no record and earned no personal best, and the last minutes of a shift cost you nothing. Now the score is taken once, when the shift ends, and the same number goes to your HIGH and to the board. A deal at minute fifty is a real deal.

A claimed callsign can no longer be lost to a momentary server hiccup. If the name registry cannot be read, claims are refused for a moment rather than answered wrongly.

Callsigns are held on the server: the name is bound to your browser and to a token that proves it is yours, which is what lets your name appear beside a score on the public board. If the file holding those claims could not be read for any reason, even for an instant, the server treated it as though nobody had ever claimed anything, rebuilt what it thought the list should be from the scoreboard, and wrote that guess back over the real one. Everybody's proof of ownership would have been erased in that write, permanently, and anyone whose name had traveled to a new browser on a transfer code would have been locked out of their own name with no way to get it back. Nothing on the server said a word while it happened. No player ever hit this and no live claim was affected: it was found in an audit and fixed before it could fire. A registry that cannot be read is now treated as unknown rather than empty. Nothing is rewritten, nothing is guessed, the failure is logged loudly where we will see it, and name claims and availability checks answer "try again in a moment" until the file reads cleanly, which they do automatically. A score submitted during such a moment is still recorded in full; only the name beside it is held back, because the server will not print a name it cannot prove belongs to you.

You can no longer put two aircraft on the same runway pointed at each other. A clearance that would do it is refused: "unable, opposite direction traffic on that runway."

16R and 34L are not two runways. They are one strip of concrete worked from opposite ends, and the game did not know that. Runway occupancy was decided by comparing the runway's name, so a departure rolling 16R and an arrival landing 34L looked like traffic on two completely different runways. Nothing warned you, nothing refused the clearance, and the two aircraft closed head-on down the same pavement at a combined three hundred knots. It came to light at Denver, where one flow was set up to depart 16R while landing 34L, but the flaw was never really about Denver: the same hole was open at every airport in the game, on every runway, to anyone who typed the two clearances by hand. Runway occupancy is now decided by the pavement rather than by the name. An aircraft on either end of a runway occupies that runway, which also closes a quieter version of the same gap: an aircraft holding in position at one end no longer looks invisible to an arrival landing toward it from the other. What has deliberately not changed is ordinary sequencing. Two aircraft using the same end one after another is normal tower work, an arrival clearing the runway ahead of a waiting departure is the everyday picture, and all of that still behaves exactly as it did, including clearing an arrival to land behind a departure that is still on its roll. Only the head-on case is refused, and it is refused in both directions: a takeoff or a line-up into traffic committed to the far end, and a landing clearance into traffic committed to the near one. Real opposite-direction operations do exist at real facilities, but they are a deliberate, coordinated procedure with its own rules, and the game does not model them. Rather than pretend to run one, it declines. A soak-test law now watches for this on every field, every runway, every tick of every test session, and the law is itself tested by staging the Denver case and confirming it is caught.

Training no longer holds you on the blip-drag until you perform it.

The turn lesson used to advance on a scope drag and nothing else, so a player on a trackpad, or one who simply typed the clearance the way a controller does, had already done what the card asked and the card would not move. Any heading now counts: the command menu, a typed C 090, or the drag. The card also carries SKIP THIS STEP, which passes the gesture alone and keeps the lessons after it. The drag is still taught, because it is the fastest route to a fine heading and doubles as a protractor.

Arrivals descending via a STAR no longer show up underneath a crossing restriction they have not reached yet.

At Atlanta, inbounds on the ONDRE ONE and CHPPR ONE were checking in at 11,000 with a charted 13,000 crossing still ahead of them; Kansas City, Denver and Seattle had the same thing on other arrivals. A published minimum now holds the aircraft up until it crosses the fix that carries it. Where the charted minimum sits above the top of the modeled altitude band, the aircraft holds at the top of the band, which is as high as this airspace goes.

The AUDIO button now actually opens the audio panel.

Clicking AUDIO looked like it did nothing: the panel was opening inside the top bar, which crops anything taller than itself, so the volume slider and the voice picker were drawn where nobody could see them. The panel now opens below the bar, under the button, at every window size.

Display settings, including the airspace map layers, are now a plain click away in the ⋯ menu.

They used to be behind a right-click on THEME, a button the top bar drops as soon as the window narrows and a gesture phones do not have, so the map-layer switches could be unreachable. ⋯ is now always in the bar and its DISPLAY & MAP LAYERS row opens the same panel with an ordinary tap. Right-clicking THEME still works when THEME is showing.

The display panel no longer runs off the edge of the screen.

Long option names pushed the panel wider than the window, so Text size "Large" and the "Neighbor airspace" layer switch were cut off at the right edge and could not be pressed. Options now wrap inside the panel and it is always fully on screen.

The airspace boundary now draws in its own violet ink so it no longer muddles the range rings.

Boundary lines and range rings are both dim circles around the field, and in the same color family they read as one confusing set. The airspace layer has its own color on all four themes while staying quieter than the aircraft, runways and data blocks. The class is still carried by the dash pattern, solid B, long-dash C, short-dash D, so nothing depends on color.

A cleared VFR arrival can now be ordered to go around while it is still maneuvering to join final.

Go Around is available before the strip reaches FINAL and for pattern aircraft on base or final. When two self-navigating VFR pilots meet, one now keeps a stable right-of-way while the other yields instead of both maneuvering indefinitely.

Repeating a VFR landing clearance now releases a present-position hold.

A hold issued after the original clearance could remain latched until the controller changed runways. Re-clearing the same runway now returns navigation to the pilot without restarting a visual setup already in progress.

Reported by a player

Small-field IFR menus no longer offer an instrument approach the runway does not publish.

The menu now says the runway is visual-only and leads directly into the point-out-field and visual-approach sequence. Landing clearance remains separate.

Repeating a VFR landing clearance no longer makes the pilot abandon and restart an in-progress visual setup.

An IFR landing clearance no longer accepts a momentary runway heading while the aircraft is still assigned a diverging vector.

The aircraft must have a workable final instead of turning away and claiming FINAL only at the threshold.

Extended VFR pattern traffic no longer turns an avoidance maneuver inward toward the runway.

Pattern aircraft now remain on their assigned rectangular legs instead of accepting generic free-roam avoidance headings. Releasing an extension authorizes the approach, but the pilot continues downwind until its predicted base track is clear of visible traffic.

VFR arrivals now wait until the airport is in sight, then build enough room to complete the landing you cleared.

The landing button appears after the pilot reports the airport in sight. Once cleared, a close-in or upwind aircraft flies away or around as needed, uses nearby visible traffic to choose a sensible visual entry, and holds one stable see-and-avoid maneuver when another track is predicted to conflict. Two pilot-owned VFR tracks no longer create a controller loss of separation; a controller vector or hold can put that responsibility back on the player, and wake turbulence still applies. The pilot slows and descends, establishes final, and lands without immediately going around or falsely asking whether it was cleared. A medical or fuel emergency may keep its speed, but no longer freezes that pilot-owned route or descent before touchdown. It may use the published final as an approximate visual reference, but it does not fly the instrument procedure, fixes, markers, or altitude restrictions.

Reported by a player

A visual arrival now shows FINAL only after it is genuinely lined up and converging on the runway.

A shallow diagonal could cross near the extended centerline and be labeled FINAL while it was still offset and drifting away. The visual final now waits for a tighter, closing alignment, drops the label if the aircraft turns away, and keeps steering decisively onto the centerline without hiding a properly aligned arrival in INTC.

A landed aircraft now rolls out down the runway and only calls ready for the ground once it has slowed, instead of stopping dead where it touched.

A landing arrival used to freeze on the spot the wheels touched and immediately read as ready to hand to ground, even while its speed was still bleeding off. It now travels down the runway as it decelerates, the way a real rollout does, comes to a stop on the pavement, and only becomes ready for the ground frequency once it has slowed to taxi speed. The time it spends occupying the runway is unchanged, so it still holds the next same-runway departure exactly as before.

A visual arrival with no landing clearance yet now holds over the field instead of flying off into the distance.

A VFR arrival that came in past the far side of the runway could get pinned to the runway heading and fly straight off the map. It now tracks to the field and orbits overhead until it receives a runway and landing clearance, then builds its own visual setup.

Visual-flight traffic is no longer offered instrument clearances it cannot fly.

The game's VFR traffic does not fly published instrument procedures. Its command panel no longer offers direct-to-a-fix, holding at a fix, RNAV or ILS clearances, or the IFR visual-approach exchange. Holding present position and the runway landing clearance stay.

VFR traffic in the pattern flies its own base to final and lands on the runway centerline again.

Closed-traffic aircraft doing touch-and-goes had started touching down off to one side of the runway, on either side, sometimes far enough that the mislaid aircraft read as blocking the runway and sent the next arrival around. Cleared for the option, a pattern aircraft now leads its base turn and rolls out lined up on the extended centerline, flies the visual approach down to the numbers on its own with no vectoring from you, and any approach that does not line up goes around rather than landing beside the pavement.

Airport labels that sit close together no longer overlap, and reveal themselves cleanly as you zoom in.

Around a busy metro like Kansas City several fields sit within a mile or two of each other, so on the lobby map their code labels used to pile on top of one another. The map now works out which labels have room in the view you are looking at and shows only those, keeping the busier field's label and holding the crowded ones back until you zoom in far enough to give each its own space. Point at any dot and its label always shows.

Two departures off parallel runways that turn away from each other are no longer failed for a loss they never had.

The real rules let parallel runways launch almost together when the two departures diverge by fifteen degrees or more. The game was holding every pair to full separation regardless and ending the shift on a correctly flown departure. It now applies the reduced minimum the divergence earns, so a properly separated pair reads as safe.

A handful of navigation fixes that sat slightly off their charted position have been moved to where the charts put them.

One fix at Seattle (HELZR) was drawn away from its charted position, so a direct-to steered aircraft to the wrong point. The same check corrected similar errors elsewhere and now stops them from returning.

Reported by fishbait322

Arrivals no longer brake early while they are still miles from lining up on final.

Coming into Seattle from the north, an aircraft still flying its arrival route could start slowing for the runway well before it was established, because a distance check ignored how far off the final course it still was. It now holds the published arrival speeds until it is actually turning onto final.

Reported by fishbait322

An aircraft cleared to one runway of a parallel pair no longer lands on the other one.

It could settle onto the neighboring centerline, fly it toward the ground and swing across at the last moment. Parallel-runway approaches now remain committed to the runway they were cleared to use.

Reported by fishbait322

Jets arrive at a realistic descent speed and slow down where the chart asks them to.

Everything used to enter the sector at 250 knots, which is the limit below 10,000 feet, so a published crossing speed above that could not be met at all. Aircraft now work out where to begin slowing, so they are stable at the restriction instead of dumping speed the moment they pass the fix before it. A charted speed also survives an altitude you assign, which is what the orders say.

Reported by fishbait322

Aircraft cleared for an approach, and traffic in the pattern, are watched for separation.

An aircraft stopped being watched the moment you cleared it for the approach, and VFR pattern traffic flew whole circuits with nothing checking it. Both now remain real traffic for separation purposes.

The strip and the scope name the departure procedure an aircraft is flying, not the fix it leaves the sector on.

A departure on the JHART SEVEN read as a TWILA departure, because TWILA is the exit gate rather than the procedure. The radio and the chart button had it right already.

Reported by fishbait322

An approach join now flies the fix it was cleared direct to.

Clearing an aircraft direct to a fix and then clearing the approach could send it to a different fix on the procedure. It flies the one you named.

Reported by fishbait322

The bug report box pauses the session while you type, and gives the time back when you close it.

Traffic used to keep moving behind the box, so reporting a problem cost you the board.

Reported by a player

Departures flying a SID fly the published path from rotation.

They held runway heading too long and then joined the nearest leg, which skipped the runway-aligned first fix on many transitions. Found in the field.

A departure that turns away from its route now says so on the radio, and only turns when a parallel departure justifies it.

A lone departure could turn away from its only fix without a word. Found in the field.

Browsing the chart viewer no longer fights the automatic focus on the selected aircraft, and the pane can be navigated properly.

Reported by a player

Requesting an airport gives an unmistakable confirmation and will not send twice.

A phone could submit the same request repeatedly because nothing on screen confirmed that the first one had worked.

Scope text and aircraft data blocks now render reliably across supported browsers.

This fixes the blank data blocks that some players were seeing.

Data blocks can no longer come up blank.

The game now detects and repairs an empty text render before it reaches the scope.

The briefing button no longer sticks on TUNING.

If audio cannot start, the control now returns to LISTEN instead of waiting forever.

Fixed garbled voices in Firefox and a crash in Safari, both from field reports.

A bad voice load is now caught before distorted audio can be played.

The tutorial's approach step clears whether you click the button or press the key.

Runway flow arrows point down the runway, in the direction of travel.

They were drawing the reciprocal, which is the approach course rather than the direction anything moves.

The theme and updates buttons move into an overflow chip when the top bar runs out of room.

The menus and buttons follow the theme, which fixes an unreadable command menu in DAYLIGHT.

Final tracking converges onto the centerline all the way through touchdown.

Removed a spurious unable-restriction call at the end of a descend-via path.

New arrivals never expect a runway that is closed.

A missed approach holds runway heading and waits for you instead of routing itself somewhere.

No session starts in a loss of separation.

An automatic handoff says contact Center, not radar contact lost.

Aircraft fly with honest inertia: level-offs taper, turns are bank limited so a fast ship arcs wide, and the takeoff roll and landing rollout take real distance.

A cleared aircraft no longer dives at the approach. Descent rates are capped per type.

Fixed arrivals flying outbound past the field while still tagged FINAL.

Two established parallel arrivals no longer false-trip the radar minima against each other.

Fixed a parallel-departure deadlock using tower visual separation and a divergent climb.