Race and Vehicle Simulations

In the world of simulations, and perhaps none more so that of race simulations, the word “hardcore” gets bandied around a lot. With each new title announced or released, almost as soon as news has broken, people are categorising the title along the spectrum of how hardcore it is or, as if often the case, is not.

With racing simulations, we have titles like Grid and Dirt firmly occupying the “arcade” end of the spectrum (I won’t go as far as the likes of Mario Kart; I think you’ll understand I am talking of a spectrum within a spectrum here). In the middle ground, we have the likes of the Gran Turismo and Forza series which are increasingly labelled under the now-popular “sim-cade” title. And of course, at the top of the spectrum we have the likes of rFactor, the Game Stock Car series and iRacing.

Within these sub-genres, there is also a spectrum; Forza is, by many, regarded as a more serious simulation than GT, and netKar Pro, especially in original trim, was undoubtedly up there as the most hardcore of the hardcore racing sim titles. Whilst there are some empirical measurements, for want of a better word, of just how realistic a title is (in-sim performance vs real world; track accuracy; level of detail of the modelling of various aspects and so on), this categorisation is largely a subjective matter, and you probably don’t need me to tell you that much debate ensues over just what the pecking order is.

Some titles seem particularly prone to such debate. Live For Speed had a devoted user-base that other titles could only dream of, and there is no doubt that for many LFS really was the sim of all sims. But despite all the clever and fancy physics modelling going on within the title (it was to my mind the first sim to get a proper handle on tyre flex, for example), many saw it simply as “arcade”. Perhaps if the conversation were taking place now rather than 5 years ago it would be “sim-cade”. Who knows? Quite why there was such difference of opinion remains something I am not too sure about. More recently, it is unlikely to have escaped the attention of sim racing fans that SMS’s Project CARS seems to have picked up the mantle of most-debated-on-the-sim-scale. Just recently things really seem to have escalated in this area, but that’s another story for another day.

Whilst some would argue heatedly over whether title X is a sim, sim-cade or even arcade, there are some titles that generally seem to be above such debate, at least in wider circles. A good case in point is iRacing, a title which has, it’s probably fair to say, held the title of king of the sims for a good chunk of the time that has passed since its release. Now I’m not getting confused here; let’s make it clear that not everyone likes it. For some it’s the price of content and monthly subscription rates, for others it’s the race structure and somewhat hit-and-miss appeals system. But whilst these are factors that do not really affect iRacing’s claim to being classed among the upper end of the sim spectrum, there are also those who take issue with elements of the physical simulation. I’ve seen the term “IceRacing” used numerous times for example, with a number of people clearly not happy with the perceived level of grip on offer from Kaemmer’s black round things. There’s also the infamous (and perhaps somewhat overblown) “two-footed-magic-save-hack”. Whilst there is real-world theory to support it as a physical phenomenon, the ensuing discussion over on the iRacing forums led to the discovery of an undeniable physics flaw relating to the application of torque on the chassis by the drive train. Do these things mean it isn’t a sim? To nearly all simmers, no. It simply means it isn’t perfect. There are flaws. There are errors. There are quirks. But it is still very much a sim.

When we talk about arcade, sim-cade or hardcore simulation when describing the likes of Grid, iRacing, Project CARS and the rest, most of the conversation seems to centre on the interaction between driver, car and track. It’s about whether or not input X in car Y in situation Z has an output that matches our expectations of real life. For some it’s as simple as “Can you do doughnuts?”, while others take a more astute approach: “Does increasing the throttle from 30% to 75% at 2500rpm in 3rd gear exiting an off-camber corner cause the rear end to push out the correct amount?”. Either way, we have expectations (justified to varying degrees by our individual experience, knowledge and understanding) of what should happen, and we judge what actually happens against those expectations. For many years people have stated the old mantra “As simulations become more realistic, they will converge to being the same product”. Whilst there is an obvious logic to this statement, I would contend it is undeniable that the current crop of sims are some way from this singularity. Take a similar car out in iRacing, netKar Pro or rFactor, and were it possible to transplant their physics engines into a common interface, I dare say anyone accustomed to these titles will likely be able to tell which one they are playing. All sims are flawed, all have issues, and all are different. But despite the differences, and whilst one might have their own pecking order for iRacing, nKP and rF, I doubt many would classify one or the other as anything other than a racing sim. But something recently has caused me to ponder this.

In choosing the title for this website, we wanted to make it clear that we were interested in car sims (like those mentioned so far in this piece) as they are our bread and butter. But also, as can be seen in Bob’s piece on SCS’s Scania Truck Simulation, we have interests that spread beyond these confines. The car sims will undoubtedly make up the majority of our content, but every now and then we might like to spread our wings a little. You’ll probably have noticed that the allusion to car sims within our title is the “Race simulations” part. Not “car sims”, but “race sims”. Why? I’m not quite sure to be honest, but all of the car simulations mentioned above simulate cars on tracks, so why not? Isn’t it just semantics? Maybe, but maybe not.

What has caused to me question of lot of the titles we call “racing simulations” is a title which I am sure many would not. The title in question is iGP Manager. For those who aren’t familiar with iGP, in a nutshell it is an F1 management title. You pick suppliers, sponsors, drivers, staff. You develop your car, practice for a race and determine car setup for your drivers, define strategy ahead of qualifying and the race, and then you leave your drivers to do their best come race day. During the races you can leave them to it, or you can watch and monitor certain statistics such as live timing, tyre temperatures and fuel use, and in your managerial role you can take on race engineer duties and alter strategy on the fly. In a sense it’s an incredibly simple title. Whilst there are a number of decisions to be made outside of the races, and the number of possible permutations will have a large number of zeros on the end, the strategy for success is a largely linear one: get the best drivers you can afford and attract to your team, secure as much sponsorship income as possible, and develop your car as best as you can. This is glossing over things, but I hope you get the idea.

Race strategies are also pretty simple. Or so I thought. On the surface, choosing the strategy is a fairly straightforward process: soft tyres are quicker than hard tyres but don’t last as long; pit stops obviously take time and so you want as few as possible. You want to strike the optimum balance between pace on the track and time in the pits. That’s about it. Yet choosing a strategy can be agonising. Sometimes that strategy you’ve spent so long tweaking turns out to be a real dud as you watch your nearest competitor head off into the distance. Other times the roles are reversed, and that winning strategy you settled on seems so obvious that you can’t believe that your competitor messed it up so badly. Perhaps most galling of all is when you think you’re in contention, running right up at the sharp end, only to see your competitor pit a lap earlier or later and stealing a march on you. Sometimes when you fuck up it’s better to really fuck up. Actually, no, scrap that. Most galling is when you are running in a strong position and run out of fuel in the pit lane on the final lap (unlucky, Jon).

So if it is so simple, why is it so difficult to get right? In short: it’s so difficult to get right because it isn’t so simple. The relative durability and pace of the different tyre compounds are not fixed, easily predictable parameters. Instead, they are complex variables interdependent on one another and a whole host of other parameters. The first thing to consider is the compound impact of the abrasiveness of the track surface and effects of the circuit layout on the tyres (conveniently supplied as a “Tyre wear” rating for each track in the circuit information). The next key factor is temperature, which is determined from real-world data (Jon looked out of the window ahead of last season’s round at Silverstone, whilst league member Suresh handily provided us with a weather report from outside of his door in Singapore prior to qualifying). This is updated by the minute, so if the clouds roll over on a sunny day you better hope the drop in temperature doesn’t adversely affect your setup too much. Conversely, if you did your setup work in cooler temperatures, you can find yourself nervously watching the ambient temperature hoping the sun doesn’t appear.

This is only the tip of the iceberg. The harder tyres take longer to warm up than the soft variants, so time lost at pit stops isn’t restricted to time spent in the pit lane. Whilst with a bit of forethought this can be taken into account and the relative pace of the different compounds can be guesstimated over the length of a stint, what is nigh on impossible to account for is what your competitors will do. Whilst the time lost getting a set of the harder tyres up to optimum operating temperature might be offset by time gained from ultimately running longer in the stint, if that dropped time sees you lose track position, crucial ground can be lost trying to regain it. Similarly, a set of softs might see you gain an initial advantage, but if you’re stuck behind someone on cold hard tyres and you’re not past by the time they are up to pace, that advantage has diminished if not turned to disadvantage. With a number of laps until your next stop, knowing a competitor ahead is pitting imminently, should you tell your driver to push to gain a short term advantage on the overlap in the knowledge that in doing so you are sacrificing longer term pace by pushing the tyres too hard? Should you sacrifice some overall race pace to try and get the jump in qualifying? Combining the changing weather with the fact the track changes over the course of a race, the length and order of the individual stints takes on even greater importance.

Even on a circuit where tyre choice is a no-brainer, strategy is anything but. Monaco is a prime example: the low speeds, relatively cool temperatures, and tight and twisty layout make soft tyres the only choice. Doing the calculations, a 3-stop strategy is the way to go. Easy. Except again, it isn’t. In our league’s last visit to the land of extravagance and £20 pints of beer, one of our league front-runners made a seemingly bad strategy call. He opted for a sub-optimum 4-stop strategy, and at half race distance was languishing down in the middle of the pack a long way from the sharp end. However, as the leaders started to encounter the dreaded backmarkers, he started to make up ground. As the leaders were slow to react and continued to see time pour away, he closed in. As the race played out, his punt on (sorry, Tero, “inspired decision to adopt”) an alternative strategy saw his driver proceed to the front of the running order, and ultimately hold on for a narrow win. A sub-optimum strategy turned out to be the best when all stochastic factors came into effect.

So where am I going with this? Well let me first pose a question: is iGP a simulation? After a few races, I’d have said no. Putting the races to one side, as a management title to ask if it is a simulation is bit like asking if Themepark is a themepark management simulation. This isn’t meant in any way as a glib comment or a slight against iGP: I think it is a great title and it has provided me with a lot of entertainment. It is also in a sense relatively simple (though it is worth noting it is in constant development and numerous features have been added since release and are continuing to be added); for me it is a management game, and a fun one at that. It obviously isn’t a driving simulator, since you can’t drive the cars. And whilst there are many factors in play, and their servers are doing all sorts of calculations during the races (which, unlike some other similar titles of this sort, are all calculated in real time), there is nothing on the scale of the conventional racing sims when it comes to physics engines and tyre modelling. To call it a vehicle simulation would be, I think, something of a stretch. However, one thing I won’t hesitate to call iGP is a racing simulator.

Perhaps the highest praise I can offer to iGP in this respect is that playing it has had a massive impact on my viewing of real world F1. Whilst I always found it easy to sit back and admire the nous of Ross Brawn as he called seemingly perfect strategy after perfect strategy during the Schumacher years, and similarly sat there and was astounded by some of the dreadful decisions made by Ferrari in recent seasons, I’ve always taken it a little for granted. It’s something which, on the surface, is pretty simple. Just like forming a pre-race strategy in iGP appears. It’s easy to look at someone getting it right or wrong after the event and think how obvious it is how things panned out. But now I watch the races with what I feel is an added insight and interest. I find myself paying far more attention to the splits and deltas, visible tyre wear and traffic gaps. I’ve found myself thinking “Well I wouldn’t have done that in iGP” as McLaren seem to try there best to lose a race; I have increased admiration for Perez as he gets more life out of a set of soft tyres than anyone else; I’ve thought “Been there, done that” as Kimi drops down the order like a stone as he tries to survive to the finish line on rubber that is a long time past its best.

Whilst iGP represents the world of F1, in many ways there are things which are quite inaccurate (or, perhaps more fairly since it isn’t an official F1 title, are different). For starters, you are not required to run both compounds of tyre in a race. Additionally, the implementation of KERS does not operate like in real world F1; instead of charging and use governed on a per lap basis, instead the KERS charges over the course of the first few laps, and this supply lasts you for the duration of the race. Were such equivalent discrepancies between reality and virtual representation present in a conventional racing sim, the title would no doubt be ridiculed by the community and consigned to the arcade end of the sim spectrum. But in iGP, it doesn’t really matter. Whether you call it “F1” or “Super Magic Coloured Dot Racing”, the title succeeds in capturing the essence of the sport and, “inaccuracies” or not, it does a fantastic job of simulating the racing.

So to bring this back to where I began, how does iGP fit in with the conventional sim racing titles we know and love? Despite its “simple” nature, I find it hard but to conclude that iGP as a racing simulator is right up there. It isn’t hard to imagine it being far more “hardcore”; a full bank of telemetry data and a few bucket loads of extra detail and functionality would soon see to that. But when we talk about racing simulators (with the emphasis on racing) as we are used to calling them, I think many could learn a lot from iGP. iRacing’s tyre temperature foibles might have an affect on the behaviour of the car over the course of a lap, and undoubtedly detract to some degree from its successes as a simulator of the included vehicles. But where the impact is really felt is when it comes not so much to the driving per se, but rather the racing as a whole. Should the focus of accuracy and detail in a simulation be millimetre perfection in track layout when something so fundamental to the racing as proper tyre temperature, wear and life-cycling appears to be so relatively primitive? The Williams FW31 could do with a phone book-sized manual to guide you through setting it up properly, yet tyre management, such a huge part of F1 (and not just in the current Pirelli era), forms such a relatively small part of the driving experience. iRacing is in constant development, and Kaemmer’s ever-evolving New Tyre Model is slowly but surely improving all the time. But it feels to me like sometimes the focus isn’t quite in the right place. It’s not just iRacing, and it’s not just tyre temperatures.

Assetto Corsa stands as possibly the most highly anticipated title in sim racing circles. With both the online component and the welcome inclusion of AI, it is clear that a large part of the AC experience will be in the racing. Whilst Stefano has shown in nKP and FVA what he is capable of (including some very good tyre modelling; heating and wear included), in AC I see another area which is going to be very important: brakes. Much of AC’s line-up is comprised of standard road cars. Whilst they might be sports cars, something that plagues nearly all road cars when on a track is brake fade. In reality, anything over a few laps of most circuits in an E30 M3 is going to require some serious attention to brake and tyre temperatures. The presence or absence of well modelled brake fade will have a big impact on what it is like to drive such a car in a sim, but an even bigger impact when it comes to racing. Brake fade brings with it a fairly major issue in sim racing, namely a lack of feeling in a normal sim brake pedal. Even a fancy hall sensor will make no odds in this respect. But does that mean it shouldn’t be included? I say no, others may disagree.

As much of a fan and advocate of racing sims as I am, the truth is I spend a very small percentage of my seat time actually racing. I’ve often wondered why this might be the case. I’ve had enough experience to know that, generally, open servers in various titles are a big no-no. But even in organised racing environments such as over on the iRacing servers, the appeal which racing has held for me has always felt somewhat limited. I’ve had some experiences on iRacing that are sure to put many off (waiting the best part of an hour for a race to start only to be taken out from behind in the first corner, for example), but that doesn’t account for why I feel far more inclined to run practice laps or time trials than to engage with other racers. And whilst far from the most experienced of racers, I’m experienced enough to know how to behave on track, and am comfortable enough dicing with others. Indeed I have enjoyed some fantastic racing in a number of titles, including iRacing, over the years.

Some time ago I came to the conclusion that it isn’t really racing simulators I am interested in, it is driving simulators. It isn’t the racing that interests me, but more the pleasure and challenge of driving the car itself. Whilst I think there is some truth to this, I’ve also come to the realisation that this isn’t quite the full story. I have realised that what I don’t like is having to race at 100%, 100% of the time. In probably any title, you can go into a race, back things off a bit, and pick up places as others overstep that fine line between on the edge and over it. But I don’t want to compete purely on attrition, I want to race. Real racing, in many forms at least, is not about going all out, all of the time. Yet until racing simulators improve in some of the key areas of car preservation (engines, transmission, brakes, tyres), provided you can maintain full on attack without mistakes, too often it is possible to sustain this approach where in real life the machinery would not allow it. Ask your driver to go all out in iGP and soon your tyres will be toast and components on the car will deteriorate at a higher rate, just as in reality. Racing sims far too often seem to do too little to reflect this fact.

There are a whole host of things that could help to remedy this. As I have discussed before, one the greatest appeals for me amongst the proposals for Racing Legends was that of car ownership. If a sim car weren’t simply a collection of polygons and data values reset each time it were loaded, but instead had some persistence of existence, with the player having some semblance of ownership (and responsibility), it might have a big impact on one’s approach to how it would be treated. Similarly, if abusing a sim car carried with it greater repercussions, in line with the real world, these things would matter a lot more. Think of a Tamagotchi with wheels.

I worry sometimes when I write a piece like this that it will be perceived as overly negative. I think it’s fair to say (and I have been told on numerous occasions) that I am very good at seeing the bad points in something. Whilst I might not say so as often,  I do see the positives too. I think iRacing, for example, is a superb simulation, and it does many things very, very well. But I also think the sim racing landscape is potentially on the brink of changing. Whilst a number of facets of rFactor are really showing their age in comparison to some of today’s top titles, ISI seem to be making great strides with rFactor2 in many of the areas that help qualify iGP for me as such a success when it comes to simulating racing. The advanced new tyre model in conjunction with the modelling of the ambient environment and “live track” conditions already make it quite a different experience from the competition. No longer does “What time have you run at this circuit” really mean anything; to glean any real significance from the answer also requires knowledge of the time/temperature and how rubbered in the track was. As with all titles, rF2 (in its beta form, it is worth stressing) has divided some in the community. Many love it, some appear to loathe it. However, whether or not you think it is the best simulation of a given car or not, and whichever way the remaining development goes, there are enough things in place to suggest it could significantly move the goalposts in terms of simulating actual racing. Similarly, in our recent interview, Ash McConnell mentioned the recent inclusion of heat transfer by Gregor Veble in his constraint based physics engine powering ORC. This is a very real feature that, depending on implementation and the car in question, could potentially have significant effects on how one needs to manage the car. Such advances really could be a game changer for sim racing.

Just as techniques improve in all areas of software development, the modelling of tyres, aerodynamics, transmissions and so on are all incrementally getting better. Whilst things sometimes take a step backwards (see Jon’s interview with Stefano about removing features from nKP, for example), overall it’s probably reasonable to say that the simulations we drive and enjoy are moving forwards, and in the process getting closer and closer to reality. However, there are some areas that are arguably some way behind the curve, and that only now are starting to catch up. GTR introduced sim racers to the world of live track technology yet it has been sadly, and perhaps strangely, absent ever since. Tyre models have improved undoubtedly in many areas, but some aspects of this seem to not to have been given the level of attention they arguably deserve.

Perhaps the biggest question mark hangs not over the heads of the developers on whom we depend to provide the things we want, but over those of us simmers. Do people want to race a car that suffers brake fade after just a couple of laps? Whilst less realistic, would it be more fun if it didn’t? Do sim racers want to have to worry about tyre temperatures and condition to such a degree that it becomes as important as any other factor when racing? In the clamour for realism, and the instant denouncement of its absence, maybe we in the sim community need to be more honest with ourselves and with each other about exactly what it is that we want. I feel and believe that I want absolute realism, or at least as much as can be shoehorned into a PC program. But would I want to run laps of the Nurburgring in a car that is begging for mercy half way around? Would it be enjoyable? Some might say that one of the many benefits that simulated racing holds over its real life counterpart is that we can avoid such things, but then isn’t that what the sim-cade genre is for; to maintain elements of realism, but water it down for enjoyment’s sake? Perhaps the answer is to stop looking at things in such simple terms and accept it isn’t as black and white as something being a sim or not. The introduction of “sim-cade” into the simmer’s lexicon helps to some degree, but the fact is it isn’t a three-stage spectrum, it is a continuous one.

We might slowly be heading towards convergence, but with so many elements of simulated racing still left to be given the proper attention they demand, there’s still a long, long way to go to until we reach the fabled singularity. I’m looking forward to enjoying the progress.