TASVideos

Tool-assisted game movies
When human skills are just not enough

Submission #5429: Mothrayas's PSX TOCA Touring Car Championship "max points" in 10:37:18.72

Console: Sony PlayStation
Game name: TOCA Touring Car Championship
Game version: unknown
ROM filename: TocaChampionshipRacing.img
Branch: max points
Emulator: BizHawk 1.11.8.2
Movie length: 10:37:18.72
FrameCount: 2267283
Re-record count: (unknown)
Author's real name: Wout van Poppel
Author's nickname: Mothrayas
Submitter: Mothrayas
Submitted at: 2017-03-31 23:50:45
Text last edited at: 2017-04-19 18:06:31
Text last edited by: Spikestuff
Download: Download (1361181 bytes)
Status: published
Click to view the actual publication
Submission instructions
Discuss this submission (also rating / voting)
List all submissions by this submitter
List pages on this site that refer to this submission
View submission text history
Back to the submission list
Author's comments and explanations:
Some say that he's actually born in Switzerland...and that he might be a machine from two decades into the future. All we know is...

This movie is submitted on March 31st, 2017 (UTC time), to celebrate the 20th anniversary of the start of the first race of the 1997 British Touring Car Championship - the tour that is simulated in this game. It also happens to be close to a certain other date.


(Link to video)

Goals:

About the game

TOCA Touring Car Championship is a racing simulation game, developed by Codemasters and released in November 1997 in Europe and August 1998 in North America (there titled "TOCA Championship Racing" on the box-art, although the in-game screens still say TOCA Touring Car Championship). It is the first game of the TOCA series of racing games, which later evolved into the Race Driver series and later on the Grid series.

The game is modeled after the 1997 British Touring Car Championship season, featuring the car manufacturers, cars and drivers of that season (excluding the independent drivers), and featuring all the racing circuits that were driven in that season. This totals out to 8 cars from 8 different manufacturers/teams, 16 drivers/racers (2 per car manufacturer/team), and 8 circuits (9 if counting two different track layouts of Donington Park as separate circuits).

Besides the usual single race and time trial modes, the game features a championship mode which, depending on the length setting, involves going through all the races of the BTCC season, in the original order, and with the original lap counts. This movie plays through the "long" championship setting, which includes all 24 races from 12 weekends across all 8 venues with their original lap counts - in other words, it clears the same amount of road as an entire year's worth of a British Touring Car Championship season (although outside of the actual races, the movie only does a single qualifying lap per race). This means that this movie does 24 races that are realistically scheduled to last around a half hour or more. The TAS brings this down by several minutes per race - suffice to say, opposing cars get lapped a lot.

Tricks/Notes

There's really not much for tricks in this movie, other than just plain driving fast, taking corners optimally, braking where necessary, etc., so I'll mostly put my notes about the game here.

Delta timing

Driving physics and controls are absurdly inconsistent. What this means is that sometimes the same steering input can have significantly different results even under very similar or even seemingly identical circumstances, leading to a lot of random undercutting corners, overcutting corners, or other such issues.

What also plays into this is that the game does not run at a consistent framerate, and uses delta timing for its physics. What this means is that if the game lags more, it will speed up the in-game physics to compensate. (Most modern games also do this). This normally helps to get a consistent gameplay experience without experiencing lag or slowdown as in older games when the framerate drops.

However, this game has a wildly inconsistent framerate, so the delta time also shakes about a lot. This depends on such things as what track you're on, how many cars are in the race, how many cars are in the race in proximity of the player, how many cars are on camera, how close or far they are on camera, where you are on the track, how much of the track is visible, what camera mode is used, and besides all that, a lot of sheer unpredictability.

This not only affects the controls as above, but also affects the physics in weird and inconsistent ways. This means that if you try to take the exact same turn the exact same way between laps, you can get completely different results, and suddenly spin out where a turn previously was possible, or the other way around. Because of this, it's practically impossible to get consistent lap times.

Weather

The weather of a track is preset before a race starts - either by the player (in single race or time trial modes), or pre-determined by the game (in championship modes). Various different weather types exist.

Sunny/Cloudy/Overcast

No in-game effect. Only the skyboxes are changed.

Most championship races feature weather of one of these types, so I'm not going to list them all.

Storm

Storm weather places dark clouds over the track, has the player cars turn their light on, and has light rainfall fall over the track. Due to the rainfall, the track is more slippery, and makes cars more prone to spinning out. Extra careful braking is required.

Also, in this (as well as other weathers with precipitation), AI Audi cars do much better at qualifying. Interestingly, that doesn't actually convert to better racing results - AI Audis sometimes still actually struggle with the rain conditions, maybe due to an overly confident AI that makes mistakes.

In championship mode, this weather is featured in venue 2 (Silverstone #1) and venue 11 (Brands Hatch #2).

Rain

Rainy weather also adds a rain effect, but it is actually more intensive, both in appearance and effect. This makes

In championship mode, this weather is featured in venue 6 (Donington National) and venue 10 (Thruxton #2).

Snow

Compared to Rain, the rain particles are replaced with a lesser amount of snow particles. This adds to the slipperiness of the track by even more than Rain.

This weather does not appear in Championship Mode.

Fog

Foggy weather coats the air with a layer of fog, severely reducing draw distance and overall visibility. However, it does not affect track physics. With a cheat code, the fog can take on various colors for a disco effect, making it a lot more fun.

This weather does not appear in Championship Mode.

AI notes

The AI of this game is sort of known for actually averting rubber-band AI. Based on limited tests I've done, I'm actually not sure if that's completely true or not, but at least even if it does exist, the effect is relatively minimal. This is actually a kind of double-edged sword; it's beneficial to the player when they get ahead, because they can then open up a very substantial lead that allows them more room for error. It's also relatively easy to force the AI to slip up, and get a very significant lead that way. On the other hand, it means that if the player themself makes a mistake, the AI will take off forward instead, and the player may be left unable to catch up. With tool-assistance thrown into the mix, the best case scenario happens for the playe, in that they can get ahead very far. In races that average at about 25 minutes, the player gets ahead by at least 5 minutes in most of them. This means that the AIs will get lapped a lot.

The AI does get upgraded or downgraded depending on a few difficulty settings. Single race mode has Easy, Normal or Hard options for the opposing AIs, and Championship mode has Standard and Expert options (don't ask me why different game modes have different ways of measuring difficulty). The TAS runs Championship Mode on Expert mode, naturally. Besides the setting, AIs also gradually improve as the championship goes on

AIs function generally the same as player cars, although they do cheat quite a bit, especially at higher difficulties. They don't reach impossibly high top speeds, but do sometimes have better than normal acceleration, have significantly better traction, and sometimes are able to use impossible motions to stay on the track. Most notably, the Audi AIs are the biggest cheaters in the game - if one starts in 1st or 2nd in a race, they get a massive acceleration boost at the start, significantly faster than is normally possible, even compared to a player using the same car. This means that even in the TAS, if an Audi gets second place in qualifying, it will speed up ahead at the start of the race, leaving the player second (until overtaking him, usually in the first turn). This is most visible in the second set of Thruxton races.

The AIs also cheat at qualification times, in that the given times are simply random times within a certain range for each driver, and are not actually simulated. This is most clear with the fact that most AI racers do not even come close to their qualification times for finishing any lap.

Pre-race setup

Before starting the championship mode, the movie goes to the controller settings mode, to set up analog controls and set controller sensitivity to high. This is done to allow a greater range in steering, at the cost of finer precision control - but a TAS can get fine precision control anyway, because, well, TAS. (Do not try racing with these settings in real-time, as the slightest flick of the stick will send the car spinning to a halt or careening off to the side). Analog acceleration is optional but is not set, because there is no need for it and it does not provide an advantage. The rest of the button layout is left at default.

When starting up championship mode, the setting is set to Long Championship (for full completion) and difficulty is set to Expert (the highest available difficulty for this mode). These should all be straightforward.

On the name selection screen, I figured I'd go with the title of a popular tame racing driver. It fits well with the idea of a mysterious driver who nobody knows or refers to by name, and who has no personality or emotion other than being a badass racing driver who would clean sweep a whole racing season, by large, if he felt like it.

While selecting the car, the gear shift setting is set to manual. It doesn't actually provide much over the default auto gear shift, but the extra control is nice, and a perfectly controlled TAS has no problem with the added difficulty in controlling the vehicle anyway.

Lastly, before the start of the first venue introduction, I pause the game to put the game music volume at maximum. This costs barely any time as the game is still busy loading things anyway, and being able to listen to the game's soundtrack more clearly is a worthwhile entertainment boost.

Car selection

Like was said before, there are 8 cars in the game (9 or 10 (depending on the game version) if counting cheat vehicles, but those cannot be used in Championship mode anyway). There are also 16 cars in a race, two for every manufacturer.

The cars are as follows, in the order they appear in the car selection screen (and how they are also ordered internally). The racers that drive them are listed besides them, also in internal order. When selecting a car, the player will replace the first listed racer.

  1. Honda Accord (James Thompson / Gabriele Tarquini)
  2. Audi A4 (Frank Biela / John Bintcliffe)
  3. Vauxhall Vectra (John Cleland / Derek Warwick)
  4. Volvo S40 (Rickard Rydell / Kelvin Burt)
  5. Ford Mondeo (Paul Radisich / Will Hoy)
  6. Nissan Primera (David Leslie / Anthony Reid)
  7. Peugeot 406 (Tim Harvey / Patrick Watts)
  8. Renault Laguna (Alain Menu / Jason Plato)

Selecting the best car for the championship is a surprisingly deep process, as the cars on the whole do not differ that much from each other. Regardless, there are still notable differences, although some cars are similar enough to be grouped together into tiers. So here's a tier list in ascending order:

  • Bottom tier
    • Vauxhall Vectra
    • Ford Mondeo
    • Nissan Primera
    • Peugeot 406

These cars are all very similar to each other, but they're all plainly worse than the other four cars. They don't accelerate as fast, have no special grip or cornering advantages, they're just plain bad options. They have the same maximum speed as all other cars at 146 MPH, but they just have a harder time reaching it.

  • Audi tier
    • Audi A4

The Audi A4 is a special kind of car (and the only one in its own tier). This is due to its four-wheel drive, and its 95kg weight penalty (the weight penalty was introduced for the 1997 season after Audi entered the BTCC scene in 1996 with their star driver Frank Biela immediately dominating the competition). The four-wheel drive provides its advantages in traction and handling, especially in rainy or other poor-grip conditions (emphasised in the game by the Audis getting greatly boosted qualifying times, usually 1-2ing the qualification rounds if not counting the player). However, the weight penalty makes the car more sluggish-feeling than the others. Conveniently though, the AI Audis in this game compensate for this by cheating like hell with acceleration at higher difficulties, the effects of which are clearly visible even in this TAS.

Ultimately, the handling may give the Audi a slight edge in rainy tracks, but overall across a whole championship the advantages are not big enough to make up for the disadvantages compared to the higher set of cars.

  • Top tier
    • Honda Accord, Volvo S40, Renault Laguna

These are the Big Three cars of the game. They accelerate faster than the others and reach the peak speed of 146 MPH faster than the other cars.

The differences between these cars themselves are minor, and it's hard to get a consistent read on them (due to the generally inconsistent nature of the game). The differences that I have been able to locate are that the Honda Accord is consistently very slightly faster at acceleration, while the Renault Laguna has better cornering abilities (and the Volvo S40 has the Laguna's speed but slightly worse cornering). Ultimately, the cornering is more important to retain speed throughout the race tracks - especially those with tough corners or hairpins - so the Renault Laguna wins out between the three. The Honda Accord might provide a slight performance gain at Thruxton (it being the fastest track), but overall it loses out.

It's also of note that the Renault Laguna is driven by Alain Menu, who won the 1997 British Touring Car Championship in a dominant fashion (after getting 2nd place for three successive seasons in 1994, 1995 and 1996), and is also consistently the strongest AI in the game (followed by Jason Plato and James Thompson). This also makes Alain Menu's Renault Laguna a logical choice for being considered the best option, while selecting it also puts him out of the competition.

Ultimately, the Renault Laguna is the best overall option for the championship; the Audi might have a slight advantage in the rainy courses, and the Honda could make a case at Thruxton, but overall taking into account all the races, the best option is the Renault Laguna, so that is what is picked for this run.

The races

The lap count given for each race counts only the laps of one race, and does not count qualifying laps, nor does it count the fact that each venue has two races at once. Therefore, while e.g. Donington GP says 18 laps, the actual number of laps driven in the run is (18+1)*2 = 38 laps.

Races 1 and 2: Donington Park (Grand Prix Circuit)

  • Location: Derby, England, United Kingdom
  • Length: 2.51 miles
  • Lap count: 18
  • Weather: Cloudy

It's the first race of the game, and already we're thrown into the longest circuit of the season with some of the trickiest turns of all the tracks.

The most striking part of the track layout is the hairpin shape near the end of the track (which is known as the Melbourne Hairpin). While most of the track can be passed reasonably well without having to brake a lot (or at all, in some positive circumstances), the hairpin is a very definite pace killer.

Right before the hairpin though, there's a little skip at the esses, where rather than weaving through the track (which requires significant braking/slowdown), one instead passes the dirt patch from the left side instead, through the grass. This saves a good amount of time every lap.

This track does include an unique pit stop trick which no other track in the game features; the pit stop in this track allows you to cut the corner here, avoiding the final sharp turn (Goddards) and, despite the pit stop slowing the car down, reducing distance enough to save a few seconds on reaching the finish line. Note that this is only beneficial to finish a qualifying lap or finishing a race; otherwise, the pit stop keeps you driving at reduced speed for a far too long time to be worth it.

At the start of the race, Frank Biela (who starts out in second place) overtakes me. This happens because the Audis, and Biela in particular, are cheating bastards in this game who actually inch ahead on the starting grid as long as you're not looking, and accelerate significantly faster than the player is capable of. This gets even worse in later races as the AI advances. This is a recurring trend, but at least the AI is overtaken by the first turn anyway.

Races 3 and 4: Silverstone (International Circuit)

  • Location: Northamptonshire, England, United Kingdom
  • Length: 2.26 miles
  • Lap count: 20
  • Weather: Storm

It's only the second venue of the game, and already we're thrown into the big daddy circuit...in stormy weather. Silverstone is another winding and tricky circuit, and the rain only makes it worse.

The qualification results show the Audis getting markedly better performances than usual - this is because of the wet conditions. The game flatly gives the Audis an advantage when calculating qualifying times in these conditions. And, as usual, the Audi who qualified in second races off faster than actually possible at the start of the race.

The tricky turns of the course combined with the rain reducing traction make for a lot of challenges to progress through the course at a proper speedy pace.

Races 5 and 6: Thruxton Circuit

  • Location: Hampshire, England, United Kingdom
  • Length: 2.36 miles
  • Lap count: 20
  • Weather: Overcast

Thruxton is the fastest circuit in the season, which is a lot of fun. Besides the esses early on in the track, almost all of the stage is just plowing forward at full speed in a course with mostly very light turns.

The "club" at the end of the track normally requires some careful motion (unlike most of the rest of the track), but a real daredevil of course just shoots straight past it, nearly hugging the posts and walls. The slightest missteer in this part would mean smashing into a wall and losing all your speed if not just outright flipping the car over. (The player character can recover from that in some gravity-defying action, but AI racers are not as lucky; if they end up on their side or upside down, they eventually are retired from the race).

Races 7 and 8: Brands Hatch (Indy Circuit)

  • Location: Kent, England, United Kingdom
  • Length: 1.20 miles
  • Lap count: 38
  • Weather: Cloudy

Brands Hatch is the shortest track of all the venues here, but this is compensated for by the lap count. Combined with the fact that the same venue is revisited later on in the season, that means Brands Hatch has a total of 156 laps run through it. That hairpin at the start (Druids) gets kind of old after a while. At least the music is good here.

Races 9 and 10: Oulton Park (Fosters Circuit)

  • Location: Cheshire, England, United Kingdom
  • Length: 1.66 miles
  • Lap count: 27
  • Weather: Sunny

The second big turn in the track is really awkwardly positioned, in a way that doesn't really allow traversing it in a quick manner - the turn is just too sharp for that. So instead, let's make things even harder by plowing through the grass there instead! Hey, it's faster, and that's all that matters, right?

Fun fact: behind the direction signs before the second turn is actually a fully drivable version of the Oulton Park International Circuit, even though it never is used in this game. It even has the Oulton Lake at the end of the hairpin (although it's blocked by an invisible wall). I assume this is a development holdover from 1996, back when the Oulton Park International Circuit was still on the BTCC calendar.

Races 11 and 12: Donington Park (National Circuit)

  • Location: Derby, England, United Kingdom
  • Length: 1.96 miles
  • Lap count: 25
  • Weather: Rain

We're back to Donington Park again - now in the rain. The good news is that at least we don't have to do the hairpins again, and the new esses can still be cut through. The bad news is that every other turn is still there, and the rain makes it significantly worse. Rain is actually surprisingly even worse than the storm at Silverstone, as there's even less traction on the road now. This makes turns such as the first one (Redgate) significantly harder, and more braking is required at some places now.

Races 13 and 14: Croft Circuit

  • Location: Darlington, England, United Kingdom
  • Length: 2.13 miles
  • Lap count: 25
  • Weather: Cloudy

Croft was a newly revamped circuit and was newly added to the BTCC schedule in 1997. I assume this debuting factor is the reason why the lap count is out of line comparatively with the lap length, resulting in a race that's a fair few miles longer than all the others (53 miles, as opposed to the usual ~45 miles), and that ends up lasting five minutes longer than the average of the other races.

At the first semi-hairpin turn, I make use of the fact that an old part of the circuit is still there on the outer side of the turn, and it still counts as tarmac, so I make use of it to make the turn much smoother for myself so that I can pass it at a significantly higher speed than is otherwise supposed to be possible.

Races 15 and 16: Knockhill Racing Circuit

  • Location: Fife, Scotland, United Kingdom
  • Length: 1.30 miles
  • Lap count: 35
  • Weather: Sunny

Knockhill, based in Scotland, is a nice hilly track. It has some turns that look tricky but can still be passed at a high speed, which makes it an entertaining track to look at from a viewing perspective.

Races 17 and 18: Snetterton Circuit

  • Location: Norfolk, England, United Kingdom
  • Length: 1.96 miles
  • Lap count: 23
  • Weather: Overcast

Snetterton is a nice fast-paced circuit, featuring some long straight action to reach full speed at.

Races 19 and 20: Thruxton Circuit

  • Location: Hampshire, England, United Kingdom
  • Length: 2.36 miles
  • Lap count: 20
  • Weather: Rain

Welcome back to Thruxton...in the rain. Combining the fastest circuit of the season with weather that screws you over at the slightest misstep is a recipe for annoyance.

Races 21 and 22: Brands Hatch (Indy Circuit)

  • Location: Kent, England, United Kingdom
  • Length: 1.20 miles
  • Lap count: 38
  • Weather: Storm

More wet conditions here. Another 78 passes through Druids do not get any more fun or refreshing if they now have to be done in wet conditions as well.

Races 23 and 24: Silverstone (International Circuit)

  • Location: Northamptonshire, England, United Kingdom
  • Length: 2.26 miles
  • Lap count: 20
  • Weather: Cloudy

And back to Silverstone - unlike the previous revisits, this time around we get to have a dry track after a wet first one, rather than the other way around. That actually makes this revisit more convenient for once, instead of more annoying.

Results

At 24 races with each 15 points for first place plus 1 point for starting on pole position, I get the maximum possible point count of 384. Jason Plato in second place gets 288 points despite not winning a race or pole - more points in fact than what Alain Menu won the 1997 BTCC with (which was 281 points).

The final points table is as follows:

Result Placing Car Pts
1st (player) Renault Laguna 384
2nd Jason Plato Renault Laguna 288
3rd James Thompson Honda Accord 224
4th Gabriele Tarquini Honda Accord 170
5th Frank Biela Audi A4 141
6th Rickard Rydell Volvo S40 138
7th David Leslie Nissan Primera 85
8th John Bintcliffe Audi A4 80
9th Kelvin Burt Volvo S40 63
10th Tim Harvey Peugeot 406 14
11th Anthony Reid Nissan Primera 10
12th Derek Warwick Vauxhall Vectra 5
13th Will Hoy Ford Mondeo 3
14th John Cleland Vauxhall Vectra 2
15th Patrick Watts Peugeot 406 1
16th Paul Radisich Ford Mondeo 0

Other notes

  • The actual rerecord count is unknown, but is at least 90k-100k.
  • This movie beats the fastest known and recorded real-time full championship run of this game (by Mushy) by 2:16:38.48 (two hours, sixteen minutes, thirty-eight seconds and forty-eight centiseconds). This is counting in-game time, and not counting qualification laps (which the real-time playthrough skipped).
  • This movie is improvable with some extra polishing. An estimated 15-20 minutes could likely be saved. That would take an astronomical amount of effort, though - you could probably better invest that in making a new SM64 120 star run from scratch, running the entire Metroid Prime trilogy, perfecting Ocarina of Time, or improving Metal Force.
  • I posit that this is, as of the time of submission, the longest complete/uncut fully video-recorded TAS in existence.

Screenshot

Thanks

Thanks to Masterjun, ThunderAxe31 and Warepire for encouragement and interest in the run as it was unfolding with the occasional update on IRC. Also, other people I forgot.

Thanks for watching!


Masterjun: Almost 11 hours!? Who even starts TASing when you know is going to end up being hours and hours of just driving laps?? The answer is Mothrayas, apparently. Judging! o:

Masterjun: Quick reminder: A TAS where the player is driving around for over 10 hours on maps that are not nearly diverse enough for the time spent on them can not get into Moons!

So we'll talk about Vault. The game is certainly non-trivial and the goal of full completion is clearly defined and makes sense. But really, everyone planning this should know how a run would look like and could calculate how long a run would be and how tedious and tiring it would be to TAS everything manually.

  2016-10-11 20:17:41 <Mothrayas> 11 hours toca touring car championship tho
  2016-10-11 20:17:54 <Mothrayas> (note to self: find a way to somehow do that for next april fools)
So here we are now, a run that passes the Vault criteria, a run that took way too much effort, a run that isn't optimized to frame levels, even contains speed-entertainment tradeoffs, but still beats the RTA record by over 2 hours.

If this isn't a valid speed record completion for this site, then I don't know what is.

Accepting to Vault as a full completion. (Also replaced the submission file with the optimized end input.)

Spikestuff: Ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha, ha.... Publishing. (Note: I'm using my older machine to deal with this, which hasn't missed a beat.)


Similar submissions (by title and categories where applicable):