Transcript
The moving target
Imagine a computer game that gets harder every time you get better.
You race a bobsleigh down a winter track. You make a near-perfect run. Your best time ever.
The next time you race, the computer’s rivals are faster. As if they had been watching. As if they had learned from you.
Improve again, and they improve again. In the end, they cannot be beaten.
Now imagine you are twelve. You do not speak the game’s language. You have a few words of English and a dictionary.
So you call your friends. You take turns. You pool your best results. And whoever is driving best that day gets the final runs.
Together, you beat the machine.
And one of those twelve-year-olds makes a promise. One day, a better game. A fairer one.
This is Slidensa — stories of ice, gravity and speed.
This episode is different from the others.
Most of this series is about the history and the physics of three sports: bobsleigh, luge and skeleton. This one is about the game that gives the series its name.
Slidensa is a game about all three sliding sports. It grew out of a single bobsleigh game from 1987. So there is an obvious question. Why does a game born from a bobsleigh game hold three sports?
And there is a second question, hidden inside that promise. What does “fair” mean, in a race against a clock?
The first answer lies on the ice. The second, in the end, leads to a club.
A game in a box
The game was called Bobsleigh. It was published in 1987 by Digital Integration, a small British company best known for flight simulators, for the home computers of the day — among them the ZX Spectrum.
Episode two of this series told its story in detail. Here is the short version.
It was two games in one. The first half was racing. You pushed, loaded and steered through six courses named after real sliding venues. They were compact, invented courses, not copies of the real tracks.
Three of the venues were in the Alps: St. Moritz, Innsbruck and Breuil-Cervinia. Two were in Germany: Winterberg and Königssee. And the sixth was Calgary, in Canada.
The second half was management. You began a season with a small budget and no bob. You bought a sled, and paid for runners, repairs, kit and training. You chased sponsors. And if your results were good enough, you were picked for the Olympics at Calgary.
That was well timed. The Calgary Winter Games were held in February 1988. A game released late in 1987 could put its players on the Olympic track before the real opening ceremony.
The game was full of small rituals. It opened with the theme tune of the BBC’s winter sports programme, Ski Sunday, playing over an empty results table. You chose one of eight nations, from Britain to Norway. Each venue had a map page: a white screen, a blue outline of the track, numbered corners and a chequered flag at the finish. And if you wanted to see a track first, the computer would drive a demonstration run.
There was the weather, too. Every venue had its air temperature and its ice, and those decided which set of runners suited the day. Choose the right set, and you had full steering. Every step away from it took away a quarter of your steering. Ignore the weather, and you could start a run with only a quarter of the control you needed, and never know why.
And the game could be shared. Two players could each take a nation of their own. The turns alternated: player one’s first run, then player two’s, then the second runs. Both raced in the same weather, and the results waited until both had finished. A player who went bankrupt was out, and the other carried on alone. To reach the Olympics, both nations normally had to finish the season in the top four.
Two friends, one keyboard, one shared goal. Keep that picture in mind.
The magazines liked it. Crash gave it eighty-eight percent. Sinclair User and Your Sinclair both gave it nine out of ten. Crash’s readers voted it the third best sports game of the year. It was later re-released on a cheaper label, and in 1988 it went into a ten-game compilation box from the publisher U.S. Gold.
A tape that travelled
All of this came on a cassette tape.
The Spectrum itself went on sale in Britain in April 1982. The Victoria and Albert Museum estimates that more than five million were sold around the world. And in the 1980s, a cassette could travel a very long way.
Bobsleigh certainly did. It had its own Spanish release, with Spanish instructions. In Italy, according to the ZX Spectrum database, it turned up on a newsstand tape under another name: Superbob.
How Western games were played behind the Iron Curtain is a story of its own, and the series will tell it.
For this story, one thing matters. A game made in Britain, with instructions in English, found players who could barely read them.
The boy with the dictionary
The game’s creator grew up on the other side of the Iron Curtain, where ZX Spectrum games were never legally sold.
His first Spectrum arrived around 1990. He was about twelve. One of the games was Bobsleigh.
Nobody in his group of friends really understood English. They got by with a few words and that dictionary. They played in turns, as a team, combining their best results to carry a single campaign to the Olympics. He calls it piggy-backing.
He finished the game many times, on a real Spectrum and later on emulators. It is still his favourite game of all time.
What he remembers is simple. The feel of the track. The crashes. Working out the money. And above all, the sound.
And the moving target.
The rivals really did copy the player. Each venue kept a table of record times. The rivals’ times were built from the best time in that table, plus a few seconds depending on the bob, plus a little randomness. And the player’s own record runs went into the same table.
So every record you set raised the bar for every rival after you. It was the game’s design, not a bug — a simple way to keep the computer competitive. Pretty naive, he says now.
The record tables were saved with the game, so reloading an older save brought the older, slower rivals back. But saving to cassette was slow and unreliable. Often it was easier simply to abandon a run.
For a group of twelve-year-olds, it felt unfair. Get better, and the game fought back. That was where the promise came from. One day, a better game. No bugs. A truly fair one.
Like many children of the Spectrum years, he wrote his own versions of the games he loved: first in the Spectrum’s simple language, BASIC, and later in Pascal. Bobsleigh was the one that would not fit. Its physics were too complicated to copy that way. Later attempts in modern game engines did not get there either.
Looking back, he sees the old game as a clever hack. It was never a real simulation of real tracks, or of real bobsleighs. It ran in forty-eight kilobytes of memory, and it cut every corner it could.
And yet the feel was right. It was good then, and it is still good now. So the thing to chase was never the code. It was the feel.
He describes it in a few short phrases. Sweaty hands. The feeling in your stomach when you have speed, on the descent. The fear of crashing. Going into a bend, trying to gain speed without going off the track.
He has never been down a real track himself. Like most fans, he knows the sport from television: the speed, the bends, the sound of a sled going past. A film about a bobsleigh team from Jamaica fed the enthusiasm, as it did for so many people. That film deserves an episode of its own, and it will get one.
One track, three sports
Television also answered the first question.
Watch a winter sliding event, and you notice something. The bobsleighs go down the track. On another day, the lugers go down the same track, feet first. Then the skeleton riders go down it, head first.
One channel of ice. Three sports.
“Why limit myself?” the game’s creator says. “I have track models. Let’s race them all.”
The earlier episodes of this series show how deep that connection runs. Episode one told how bobsleigh began, as sleds bolted together on Swiss roads. Episode four told how riders on the Cresta Run lay down and went head first. Luge has a story of the same age, in the sled races of the German-speaking Alps.
Three different stories, from the same valleys and the same winters. And since the 1970s, new tracks have generally been built to serve all three. Bobsleigh and skeleton start from the same house. Luge usually has its own starts, but rides the same bends.
The game’s first track is a good example. The track at Igls, near Innsbruck, was rebuilt for the 1976 Winter Olympics. According to the venue’s own history, it was the first artificially refrigerated track built for all three sports together. It became the model for almost every track since.
A game about only one of these sports starts to look oddly narrow — like a racetrack that allows only one kind of car.
How a sled rides
Three sports in one game raise problems a bobsleigh game never meets.
The first is control. A bobsleigh pilot steers with ropes. A luger steers with calves and shoulders. A skeleton rider steers with shoulders and knees. Episode three was all about those differences.
A game cannot ask a player to press a calf against a runner. So Slidensa gives every sled the same control: three states. Left, nothing, or right. That echoes 1987, when Bobsleigh also steered with two keys. And it keeps players equal: a keyboard, and one day a gamepad or a touchscreen, must all give the same three choices.
So the feel has to come from how each sled answers. How quickly does the steering build? How does the sled settle? How strongly does the bend pull it back?
The second problem is pace. The game has four sleds: the two-man bob, the four-man bob, the luge and the skeleton. In the real sport, on a comparable course, the luge is usually quickest, then the four-man, then the two-man. The skeleton is usually the slowest, or close to the two-man.
Slidensa keeps that order, in the ballpark of the real sport. The real gaps are written down as ranges, and every version of the physics is checked against them. If a change breaks the order, the sleds are retuned.
The tracks themselves are models, not replicas. They resemble real tracks without being surveyed copies, and the physics is tuned to feel good, not to reproduce real times to the hundredth. That is the 1987 lesson again: keep the real sport’s relationships, and chase the feel.
So how does a sled ride?
In Slidensa, the sled is a single point riding inside a fixed channel of ice, shaped like an upright U. On the straights it sits near the floor. In a bend it is thrown outwards, and it rides up the outer wall.
How far up? Until the slope of the wall balances the bend. Go faster, and the sled climbs higher. Nobody tells the sled where to be. The banking in the track’s data only says where it ought to settle at the speed the track was designed for. Arrive faster or slower, and it settles somewhere else.
Climbing is not free. Riding up the wall takes some of the sled’s forward speed, and coming back down returns it. And the outside of a bend is the longer way round, so a sled that rides too high travels further.
The runners turn the sled rather than scrubbing its speed away. Steering takes some sideways speed from the forward speed, and the runners then line the sled up with the run again, returning most of it, minus a little skid. Gravity, friction with the ice, the air and the brake do the rest. And one more thing slows a sled down: forcing it away from the line the bend wants to give it.
Put that together, and a single bend becomes a sequence of choices.
You come off a straight at speed. The bend arrives, and the sled is thrown outwards, up the wall. You could press against it and hold the sled low. That feels safe, but it costs you speed, because you are fighting the bend. Or you can let it ride: up the wall to where the bend and the slope balance, around the curve, and down again as the bend lets go. You correct it only when it wanders.
That second way is faster. It is also where the danger is.
Because on the way out of a bend, high on a steep wall, where the bend no longer holds the sled in, very little presses it into the ice. There, the runners can lose their grip. If the sled is not too high, it slips: it slams back down the wall and loses speed. If it is high on the wall, it tips over, and that is a crash. Over the lip at the very top, the run is over too.
That rule has an ancestor. In 1987, roll too far on a bend’s exit, and the bob bounced; roll much further, and it crashed. Slidensa works it out from where the sled actually is.
The start is physics too. Before the crew boards, the player pushes the sled away stroke by stroke, and each stroke is a running step. A bob crew pushes for about four seconds and around thirty metres, then jumps in. The luger, sitting on the sled, paddles away with the hands. And a crew that fails to get aboard in time has no run at all. The old game had that rule as well. Its message read: you did not get into the bob.
How close does all this come to the real sport? Close enough to check. On the game’s Igls, the autopilot takes the luge down in about fifty-one seconds. The four-man bob takes about fifty-two, the two-man a little more, and the skeleton nearly fifty-four. Top speeds run from about a hundred and twenty-four kilometres an hour on the skeleton to about a hundred and thirty-two on the luge.
All four land within about two and a half percent of real race times at Igls. That was never the goal. But it is a good sign when the feel and the stopwatch agree.
And in the hardest bends, the sleds pull about five times their own weight, for a fraction of a second, as on the real track.
A run you can trust
Underneath all of that sits the most important decision in the game. It is invisible, and it is about cheating.
In Slidensa, the race does not happen in the pictures.
It happens in a separate piece of software: the simulation. The simulation knows where the sled is, how fast it is going, how hard it is steering, whether it touches the wall and when it crosses the line. It knows nothing about graphics. The pictures, the camera and the sound only show what the simulation tells them, and they never change the result. Even the flying spray and the tumble after a crash are decoration, drawn from the simulation and never fed back into it.
The simulation is built to be completely predictable. It moves in fixed steps, sixty times a second. It keeps its own race clock and never consults the computer’s. At present it uses no randomness at all, and if randomness is ever added, it must start from a fixed seed, so it repeats exactly.
So a run does not have to be stored as a video, or as a time. It is stored as something much smaller, and independently verifiable. It is a list of what the player pressed, step by step. And it carries the version of the physics it was ridden in. A fifty-second run is three thousand of those steps, and the recording holds what the player pressed at every one of them.
Feed that list back into the same simulation, and the same run comes out again, to the thousandth. That is what the ghost of your best run is. Any copy of the game is designed to do the same, on a different computer, drawn by a different camera. Even the project’s broadcast-style videos, with directed cameras and a timing display, are drawn from these recordings.
The game already relies on this for its own records. Every record run is replayed from its inputs, and a replay that fails to reproduce the clock does not count. And because the physics will keep changing as the game is tuned, record tables are kept separately for each version. A new version starts new tables. The old records stay.
Online, the same idea becomes the anti-cheat. A player will not send in a time. The game will send the inputs. The server will run them through its own copy of the simulation, work out the time for itself, and treat that result as the one that counts.
A cheater can tamper with their own computer. They can change what the screen shows, or speed up the pictures. Scrambling what is sent cannot stop that, because the cheater controls the machine doing the scrambling. So the server believes neither the screen nor a time. It believes only its own replay.
That rules out faked times. Replay checking cannot prove that a human pressed the keys, though, because a program could produce inputs too. Defences against automated players will be a separate job.
And there is one honest caveat. Different web browsers have to produce exactly the same numbers, down to the last digit, and that has not yet been proven for every browser. Until it is, the plan is for the server to allow a tiny tolerance.
What exists, and what comes next
So what exists today, at the time of this recording?
The game runs in a web browser. It has one track, Igls, one of the old game’s six. The track plan, the roads and the gondola beside it come from open map data, and the height of the ground from an open terrain model. The mountain skyline is drawn separately.
It has the four sleds, each with its own start. It has a time-trial mode, where your best run races beside you as a see-through ghost. There are bronze, silver and gold medals to aim for, and a fourth for matching the autopilot. And there is a mode for learning the track, a descendant of the 1987 demonstration run, with one difference: the moment you steer, the sled is yours.
It is played on a keyboard for now.
There are three ways to watch a run. The default, the classic view, is a tribute to 1987: a narrow view of the track ahead, with the sled drawn large in the lower part of the screen. A near view sits low, just behind the front of the sled. And a chase view follows from behind. A racing line can be drawn on the ice, or hidden. The ghost can be switched on or off. And a stricter rule against braking early can be turned on, for those who want the purest test.
And it sounds right — which matters, because for a twelve-year-old in 1990 the sound was the best part. Slidensa has no recorded sound effects at all. Every sound is generated as the game runs, responding to the sled and its surroundings. Those sounds were tuned against real recordings from beside real tracks. A sled on ice turns out to be a broad roar. A luge’s steel runners ring at a low, steady pitch. And a bobsleigh rattles, about fourteen times a second. In the game, each sled has its own roar, its own ring and its own rattle. The crowd, the wind and the scrape of a wall are made the same way.
Then there is the plan, in four steps.
The first release is for one player: solo runs, in all three sports, and the slow, private business of getting faster. Next comes an online leaderboard, checked by the server’s own replays. After that, leagues, with seasons, divisions and recurring events. And the long-term vision is large races with many players at once, built on the same idea: the same simulation everywhere, and a server that checks the result. The network code for that is still to be written.
More tracks are planned too. Each will have its own model and fixed scenery, informed by open data and photographs where they exist. Before a track is accepted, every sled is run down it to check that the real sport’s order still holds. If it does not, the sleds are retuned. The track is never bent to fit.
The plan is to bring back all six venues of the 1987 game, with better layouts. That is more than nostalgia, because two of those tracks no longer exist. Calgary’s track closed in 2019. The track at Breuil-Cervinia, in the Italian Alps, closed in the early 1990s, and its ruins are still partly visible in summer. A third, Königssee in Bavaria, was wrecked by floods in 2021 and is being rebuilt.
A game can do something small for them. It can keep them.
One more tribute to 1987 may follow: the old map page, brought back as a track map with numbered corners that you can watch a run travel around.
There are no release dates in this plan. The game is in development. Online play, leagues and live races are plans, not yet built.
Before any of that can become a club, though, it needs a name.
A name that favours none
A game about three sports needs a name that does not belong to one of them. Most games like this reach for bobsleigh, because it is the best known.
The first favourite was Icenza. Short, cool, and built on the word ice. It was already taken.
Then came the combinations. Cool runs. Speed tracks. Ice tracks. Ice runnings. Mixes of ice, race and slide.
The one that stuck was built on the word “slide,” and that was the point. “Sliding sports” is the real collective name for bobsleigh, luge and skeleton together. A name built on “slide” favours none of them.
The name is Slidensa. It is pronounced SLY-den-sa.
Icenza belonged to the ice. Slidensa belongs to the slide — to all three sports at once.
Say it out loud, once. Slidensa. It sounds like a sled leaving the start — quick, low, and a little bit dangerous.
Don’t you think it’s a cool name?
A club, not a leaderboard
Back, then, to the promise of a fair game.
Offline, fair means you always have a chance. The computer’s rivals have limits of their own. They do not copy your best time. Weather and conditions change, but the rivals cannot run away from you.
Online, fair means the server believes its own replay, not your word. No online game is ever perfectly fair, but the aim is to come as close as possible.
And there is a deeper kind of fairness, which has nothing to do with rivals.
“The only truly fair goal,” says the game’s creator, “is to beat yourself.” To improve your own times. That is the real fun for dedicated players.
There is already a model for that: the Cresta club.
The St. Moritz Tobogganing Club calls itself a private club, founded in 1887. As episode four told, it began when the hotel that built the Cresta Run handed it over to a committee of the riders themselves. From the start, the people who rode the run were the people who ran it. For well over a century it has done one thing: helped people ride one run, and ride it a little better.
Look at how it treats a beginner. You book a practice session. Before your first ride, you get an introductory talk from the club’s Secretary. Senior members — the club calls them its “Gurus” — teach you how to ride. You get two or three rides on your first day, and you join as a temporary member for the season. You start from Junction, a third of the way down. Only when you are ready do you go from the top.
The club keeps its riders’ times. Its races carry names and histories: the Grand National, run since 1885, and cups given by members and named after them. It has its own colours, gold and scarlet. It calls the Cresta one of the last truly amateur sports. And its records can last: according to the hotel’s history of women on the run, one woman’s record from the top, set in 1911, stood until 2023.
And falls are part of it. The club says about one ride in twelve ends in a fall, more for beginners. It has a sense of humour about that, too. Fall off at the famous bend called Shuttlecock, and you become a member of the Shuttlecock Club, entitled to buy its tie.
But the real point is the atmosphere.
The Cresta is ridden in the mornings, from just before Christmas until the end of February or early March. A writer in 1913 explained why: riding starts after a night of frost has hardened the ice, and it stops when the sun softens the banks. So the day begins in the cold, in the blue shadow of the valley, with the run at its fastest.
The club holds more than forty races a season for its members. The fastest run ever from the top took under fifty seconds. Riders meet at the Sunny Bar of the Kulm Hotel. By one account, the first great celebration there followed the first Grand National, in 1885.
A Swiss hotel association adds a story. The winner of one of the club’s cups, the Morgan Cup, must fill it with champagne — no less than twenty-nine and a half bottles, the association says. And Cresta riders, the same account adds, have been known to tell one or two true, and perhaps not entirely true, tales of their daring.
Fiercely competitive on the ice, and an amateur club everywhere else. A record you chase for years. A tie you earn by falling off. A bar where everyone knows exactly how fast you were, and how you crashed. Riders come back winter after winter — not for Olympic medals, but because the run is the same and they are a little better than last year.
That is the spirit Slidensa is aiming for. An online club, with friends, and a game with enough depth, and enough difference between a good run and a great one, to keep people coming back.
What it could become
What could that club look like? These are dreams, not plans with dates, and some may never be built. But almost all of them come from the stories in this series.
A group of friends could share one team and one season online, and whoever is riding best that week takes the important runs — the piggy-backing of 1990, made official. Or, closer still to that childhood, two or more players could pass the keyboard on one computer and take turns, as Bobsleigh allowed.
Clubs of players could challenge each other. The Cresta’s first great race set ten riders from Davos against ten from St. Moritz. Davos won, on the run St. Moritz had built. Imagine a club with something to defend the following winter.
A newcomer could follow a learner’s path before open competition. And an experienced player could leave a ghost behind. A new rider would see exactly where a guru brakes, where they let the sled run, and where they hold the line.
Lost tracks could be raced again. Rebuilt from photographs, plans and memory — not as replicas, but as a way of remembering them.
A track could change with the seasons. The Olympia Bobrun at St. Moritz, as episode one told, is rebuilt by hand every winter from snow and water, never quite the same twice. A game could do the same, so that last winter’s best line is not quite this winter’s.
Any player could turn their own best run into a broadcast, and share it, as if it had been on television.
And clubs could be run by their members. They could set their own races and recurring events, with their own traditions. They could give cups named after their players. They could even keep a light-hearted honour for crashing at a famous bend — a Shuttlecock Club of their own.
They would need records that last. The idea is to keep every record and every ghost with the physics version it was set in. Then an achievement never disappears just because the game moved on. Years from now, a record from the first online season could still be there, with the name of the person who set it.
That is the Cresta lesson in one sentence. A club remembers.
You chase yourself
So, is a truly fair game possible?
Probably not quite. But the fairest thing in a sliding sport has always been the simplest: the track, the clock, and you.
Remember the picture from 1987. Two friends, one keyboard, one shared goal. The old game already knew that racing is only half the story. You race against the clock. But you race among friends.
Around 1990, a group of twelve-year-olds with a dictionary took turns at a game that copied their best times and got harder every time they improved.
Today, one of them is building a game where the rivals have limits, a submitted time is never simply trusted, and the time that matters most is your own.
The rivals no longer chase you.
You chase yourself.
Slidensa is in development. Sources, a transcript and the project are at slidensa.com.
Slidensa is an independent game about bobsleigh, luge and skeleton.