Both reference sites sell a position.
Outbid.lol and TopFloor.company run the same loop: pay to hold the top slot, lose it when someone pays more. TopFloor's contribution was a spatial metaphor, height as status and "buried" as the consequence. That is one lever pulled once.
What they sell is a position: a discrete integer slot, held indefinitely once purchased, unchanged between page loads. Three dead spots follow.
- No event. A takeover is instantaneous and invisible. Nobody watches, they check.
- No counterplay. Once outbid, your only move is a bigger number, later.
- Deterministic whale win. Deepest pockets hold first place permanently, and the player has exactly one decision available: how much to pay.
The fix is not animating the leaderboard. It is changing what is being sold.
Why the score is a duration and not a height.
The first draft of this prototype scored you on live altitude in kilometres. Nothing was ever published that way and no player was ever affected: it was replaced during the build, before this went anywhere. It is written up because the reason it was wrong is not the obvious one, and it is the decision the rest of the design rests on.
The problem was not that kilometres are arbitrary. Arbitrary units make excellent status objects: golf's −7, a sub-four mile, ELO 2800, Ironman's 140.6. Arbitrariness is not the bug.
A live altitude is an instantaneous scalar with a 4.7 second median hold, so by the time anyone reads it, it is not merely stale, it is false. There is nothing to report because nothing was completed. A rank integer works for Outbid not because it is business-native but because it is a state you were in, and it stays true long enough to screenshot.
A duration does. It is finished the moment you post it and stays true forever. It is also the only axis here with any dynamic range: the entire live altitude band, from an exposed leader at 81.5 km to a fully towed craft at 122.8 km, spans 1.5×, which no relabelling stretches into readable separation. Durations run from a 3 second spike to a 30 second hold, a 10× spread a stranger reads instantly.
So kilometres stayed, demoted to telemetry. They sort the race, draw the picture and give the tow its geometry, and they stop pretending to be the score.
Position is the integral of spending over time.
Not an auction. A continuous two-axis race where money is fuel rather than a price. Everything below falls out of three lines.
density(alt) = exp(-alt / 42) air thins as you climb thrust = $/sec * TPD * density(alt) * tow * jink * (1 - 0.42 * wind) dv/dt = thrust - G - DAMP * v first order, tau about 2.9s
exp(-alt / H) is the barometric formula,
and air density genuinely does fall off exponentially with height. That single property
is what makes each extra dollar buy less climb than the last, which is what produces a
ceiling money cannot cross. Everything around it is set for playability rather than
accuracy. The scale height is 42 km against Earth's 8.5, gravity is roughly a hundred
times Earth's, and dollars-to-thrust has no physical equivalent at all. Air at 110 km
here still has nine percent of sea-level density; in reality that altitude is vacuum,
and no air-breathing engine works above about 30 km. Altitude in this game is a
readable interface for diminishing returns on spending, not a claim about the sky.
01Gravity is universal
Stop spending and you sink. There is no pay-once-and-sit. Every craft on screen is climbing or falling at all times, which is what makes the field worth looking at.
02Height gets exponentially more expensive
Thrust efficiency decays with altitude, so holding a high position costs far more per second than a low one. Measured cost to hold, at zero vertical speed:
| Altitude | Cost to hold |
|---|---|
| 0 km | $13 / min |
| 60 km | $52 / min |
| 100 km | $135 / min |
03Money alone stalls at 104 km
At maximum burn, thrust equals gravity at 104.4 km. A hard wall. No balance lifts you past it on its own, and that is what makes a whale insufficient.
04The only way past the wall is a tow
Sit in the wake 1.7 to 22 km below another craft, in their lane, while they are burning, and your efficiency multiplies by up to 1.55. Sit closer than 1.7 km and you are in turbulence instead, which costs you.
| Route to altitude | Ceiling |
|---|---|
| Solo, maximum burn | 104.4 km |
| Allied through a pact or crew | 116.7 km |
| Full lane tow | 122.8 km |
| Redline spike, transient | 133.5 km |
| Too close behind, turbulent | 91.7 km |
05Lane steering
Getting into a wake is a positioning skill, not a purchase. Leaders can jink sideways to shake a tow hunter, but steering costs up to 15% efficiency, so shaking someone has a price. Every contest becomes a readable dogfight rather than a number comparison.
06Redline and burning out
Redline doubles your burn for a spike. Heat builds over six seconds until the engine burns out: zero thrust for 7.5 seconds while you tumble, and you restart hot. Critically, heat only sheds at low throttle. Without that rule you can chatter the redline just under the limit forever and hold a permanent thrust bonus, which quietly buys back the static first place this game exists to prevent. Cooling has to cost you height.
07The wind, or why first place decays
Lead the field and there is nobody above you to tow off, so you take the whole climb alone and your thrust bleeds away. Fully in the wind, roughly 26 seconds at the top, a leader sustains only 81.5 km, which is 23 km below their own solo ceiling. They sag back into the pack and someone else takes it.
08Syndicates and pacts
Allies tow each other regardless of lane, and the lower partner pledges a share of their burn upward: they are billed in full, and a fraction arrives as the other's thrust. The relay leaks with distance and reaches zero at 48 km, so a coalition has to climb together to lift anyone high, which drains all of them at once and means a carry can never outrun its own support. Defection is one keystroke and strands the partner mid-climb.
This is the only counter to a whale the physics permits, which makes it a real social layer rather than a bolt-on feature.
The Line sits above what money can reach.
Its placement is the entire pitch.
| Solo, maximum burn, sustained | peaks at 105.2 km, never passes 106 |
| Solo redline-spam, the most reckless thing money can do | touches 108 for 9 seconds out of 600 |
| The Line | 110 km |
| Towed off a rival | 122.8 km |
It was chosen by sweeping it, not by feel. Across one 600 second session:
| Candidate line | Whale's time on it | Occupied | Holders |
|---|---|---|---|
| 106 km | 107.5 s | 113% | 5 |
| 108 km | 39.0 s | 70% | 5 |
| 110 km | 0 s | 43% | 3 |
| 112 km | 0 s | 23% | 3 |
110 is the lowest altitude the highest spender never touches while the board stays alive. It peaks between 109.8 and 110.2 km: a permanent, visible near-miss.
One board you buy into, one you cannot.
- Live hold. Whose clock is running right now. Usually nobody's, which is the point.
- Longest holds. The record, in seconds. It never decays, and it is the board a whale cannot buy into.
- Uptime above 60 km. The ladder every paying player's clock actually runs on, won by patience rather than spikes. This is what most buyers are really purchasing.
Collisions are predicted, not reported.
Neither reference site has one. Both show a state. APEX integrates every craft's trajectory forward twelve seconds, with density recomputed at each step and fuel drawn down as it goes, then fires an alert the moment two paths are going to intersect.
TopFloor tells you that you got buried. APEX tells you that you are about to be, puts a countdown on it, and hands you a throttle. That converts loss aversion from a static regret into a live, temporal, actionable threat, which is the difference between a thing you check and a thing you watch.
The same integration draws the trajectory arcs on screen, so the picture and the alerting system are literally the same maths. It also means each craft's arc visibly flattens at its own ceiling: the central constraint is something you see rather than read.
Atmosphere and ascent, drawn as flight telemetry.
- Burn rate is plume length and brightness. You can see who is spending.
- The tow is a translucent wake cone; a craft in lock shimmers green.
- Prediction is a dashed forward arc with a terminal dot at the ceiling that burn buys.
- The Line is a gold rule across the sky with the airspace above it tinted, the longest-ever hold hanging off its end, and a running clock beside every craft above it.
- The wind is a widening gold ring only the leader wears.
- Burning out kills the plume, goes dark red, and tumbles.
Each one exists to prove a claim.
| Rival | Policy | Claim it proves |
|---|---|---|
| MERIDIAN | Whale. Near-max burn, huge balance, ignores towing. | Money alone never reaches the Line. |
| PARALLAX | Tow hunter. Seeks wakes, steers into them, bursts when locked. | Technique beats spend. |
| NULLSET | Sniper. Loiters cheap, then commits everything to steal a peak. | Timing beats both. |
| SYNDICATE K | KILO, TALLY and ORBIT stack in a chain and pledge to whoever is highest. | Coalitions beat whales. |
| LONGHAUL | Grinder. Holds about 62 km on minimum sustainable burn, forever. | Uptime is a separate game. |
They react, too. Hold the record long enough and MERIDIAN and PARALLAX form a counter-syndicate against you, announced on screen.
Measured behaviour.
From 300 to 600 second headless runs of the shipped build:
Time on the Line by strategy, with a competent pilot flying the player:
| Strategy | On the Line | Peak | Note |
|---|---|---|---|
| YOU | 32.3 s | 115.1 km | towing a rival, redline only when locked on |
| KILO | 29.6 s | 119.5 km | syndicate carry, two crewmates paying for it |
| PARALLAX | 26.4 s | 117.6 km | tow hunter |
| NULLSET | 16.2 s | 120.8 km | sniper |
| MERIDIAN | 1.3 s | 110.2 km | the whale. Outspends the board, never holds. |
| LONGHAUL | 0.0 s | 71.2 km | never contests altitude, plays the uptime ladder |
Across runs the whale peaks between 109.8 and 110.2 km and its longest continuous stay above the Line is 1.3 seconds, always alone rather than towed. It has never held it. Every craft that does hold it does so by using another player.
Two bugs, both invisible on screen, both fatal to the thesis.
Burning out was resetting heat to zero
Which made blowing your engine a free reset rather than a punishment. The optimal solo line became redline-until-you-blow, forever: measured at 526 seconds above 108 km out of 600, from a craft whose sustainable ceiling is 104.4. Restarting hot fixed it. The same craft now manages nine seconds.
The wake was too short, not too narrow
A competent pilot could hold a tow only 2% of the time, which made the Line unreachable by skill and left pacts as the only route. The instinct was to widen the lane target. Classifying every failed tick said otherwise:
tooFar 69% nearest craft above was beyond the wake's reach tooClose 12% noneAbove 10% laneMiss 2% <- the thing intuition wanted to fix notBurning 2% TOWING 3%
Lengthening the wake from 9.5 km to 22 km took the player from 0 seconds on the Line to 32, while the whale stayed at zero. Widening the lane band, the intuitive fix, moved the number by one percentage point and let the whale in.
What is real and what is not.
The economics, the ceilings, the physics and the rival strategies are the real proposed design, and every number here is measured from the code rather than asserted.
Balances are simulated. No payment is taken, no payment is possible, and there is no backend. A production version would need real money handling, identity, an anti-collusion story for pacts, and a serious look at whether continuous real-money burn is a responsible mechanic at all.
That last one is a genuine open question, not a formality. The difference between compelling and predatory here is spend caps and session limits, and they belong in the product before it takes a cent. "APEX" is also a working name and would need a trademark check.
One commitment worth writing down while it is still cheap to make. A product shaped like this has real pressure to fill an empty sky with house craft and keep the meter running, because an empty board earns nothing. The meter floor in the build is the structural answer: below four craft airborne, nobody is charged. If house craft ever fly here they will be labelled as such, and a record set against them will not count.
Why this is a different game.
TopFloor sells a seat with a view. Its player makes exactly one decision, how much to pay, and afterwards there is nothing to do and nothing to watch. Outbid sells a receipt.
- Your score is a clock, so it is finished when you post it and stays true forever, which a live altitude never could be.
- The Line sits above the ceiling of maximum spend, on purpose and by measurement, so the highest-spending strategy has held it for zero seconds.
- The only way over is to borrow a rival's thrust, which you lose the instant you pass them.
- Leading actively drains you, so nobody can perch on it either.
- Every threat is predicted with a countdown, so there is a spectator layer rather than a state to check.
The test held throughout: strip the visuals entirely, leave only the numbers, and it is still a different game. The tables above are that test, and the number that matters most in all of them is a zero.