Divergent Governance: How to Increase Overtaking and Mechanical Diversity

R

Richard James

Guest
Hello everybody,

My name is Richard James, I am a professional aerospace engineer and I have been a motorsport fan since childhood. Like almost all motorsport enthusiasts I have become disheartened by the mechanical sameness of most series and the lack of action and overtaking, especially in Formula 1. So I decided to put my aerospace root cause analysis skills to use and have a go at trying to solve the problems. The results of my efforts are here http://www.divergentgovernance.co.uk and I would like you to please let me know what you think.

As a heads up to what this idea has achieved so far; it is being debated by Indycar as a possible method to ensure mechanical diversity within the open aero rules for their 2012 car. I am also working very closely with an ex FIA technical consultant who has suggested it is vital to gain as much publicity for my idea as possible, hence the website and this post. The concept will be featured shortly in Racecar Engineering Magazine and it is to be validated by Cambridge University. So, in summary, it is slowly gathering momentum but it requires far more critical mass. This is a powerful voice for the fans to shout loud and clear about what they want to see. This is the reason I am asking you read it, get enthused, and to go forth and spread the word.

Really hope you like it,

Rich
 
Errmm, going to have to wait until the weekend before having time to read this Rich. As soon as I've found the time I'll try and offer some comment (assuming I understand it ;) )
 
What you need to do it lots of wind tunnel testing and CFD with one car behind the other and look at the behaviour of the following car and the efffect of the slipstream :thumbsup:

Then ensure that the cars have a push to pass like the IRL: 20 per race, 500 rpm boost, lasts for 15 seconds, 10 second recharge :)

The torque curve of the V8 engines is very short so its difficult to keep up momentum especially in slow corners so an engine with better torque will help. Bring back the V10's (and 12's)

Diversity can be acheived by bringing back the tyre war and having only 2 compounds for the whole year, really soft and really hard.
 
I'll attempt to digest and make comment... having skim read part one, i wouldn't disagree with much in the problem analysis, so I'm looking forward to your solutions. :goodday:
 
WOW!!!! Having now read the full paper (taking well over an hour) I must say firstly it’s an incredible piece of work. I’m currently sat at my desk surrounded by notes that were scribbled furiously while reading each paragraph and many of which were scribbled out again as each point was answered in subsequent paragraphs.

There is an awful lot of information to digest here in a single sitting however your over arching point makes perfect sense and it’s easy to see why most people will embrace this concept on reading it.

There are quite a few areas that are worth commenting on however I think I’ll hold off for the time being until more of us on here have read the paper and are ready for a discussion on any perceived pro’s and con’s.

Thanks a lot for posting it and I would imagine that there are a lot of us on here who will be debating this for a while to come. I look forward to that.
 
Just a note to say I too am reading and slowly digesting your work, Richard (1 hour in and a third of the way through...) and looking forward to joining in the discussion shortly.
 
It's good to hear you are all having a go, I know that if you read it thoroughly it takes about 3 hours (keep going Galahad). Can't wait to hear if it is understandable and, roughly, represents the majority consensus; and to just hear your general thoughts.

Rich
 
Right, I think I'm done and have had time to let it sink in. Apologies if this comes out as a bit of a muddle, I'll try to keep some sort of structure to it.

Firstly, I need to declare my position: I have followed F1 - and other forms of motorsport - since the early 1980s and continue to do so today. Much has changed over that period and my own tastes have evolved, but no matter how bad things have become, on occasion, I have never felt inclined to desert F1 completely. I recognise that as F1 has accumulated an enormous and lucrative casual and semi-enthusiastic fanbase it has lost a number of enthusiasts who are more fully engaged with all aspects of the sport, of which I see that you are one.

As such I don't perhaps see the future in such clearly defined terms as you. Intermittent tinkering with the regulations has served to maintain my interest over the years, increasing performance differentials over the short term and, crucially, upsetting the "natural order" of teams. Furthermore I do not yet believe we are at a point where the computer has eradicated the human element of design, or of innovation - though that date is approaching, it is a way off. Clearly, neither the teams with the biggest budgets, nor those with the biggest computers, do all the winning (in fact there are several examples of the reverse). It seems likely to me that the future for F1 will have to involve more frequent and more radical rule changes from year to year (or even month to month) to simply "stand still" in terms of unpredictability and interest. Nor do I think downforce is the devil, though it has been allowed to run way beyond control.

As regards your piece, I should congratulate you first on your vision and the clarity with which you express it. It is a sophisticated argument in many ways, even if built on some common sense foundations, but I remained interested and engaged throughout. As a whole it represents an outstanding manifesto and absolutely deserves the widest audience possible.

Moving on to the content, I absolutely share your analysis that the current regulatory framework means that F1 has a finite amount of time left before its problems consume it, but I don't think that a "tipping point" has yet been reached. Or to put it another way, you're even further ahead of your time than you may think you are! For as long as the sport is continuing to accumulate fans and increase income (which it assuredly still is) then discussions of the kind that have taken hold this year in Indycar are extremely unlikely to happen.

I also agree with the four key problems you identify and the root cause being design convergence. It is clear that given a long enough period without any rule changes, F1 would become a spec series through the back door as everyone would simply design the same car independently. This problem has been accelerated by the development of the rulebook, which instead of being top-down (imposed by the governing body) has too often been bottom-up (teams setting rules to please themselves, trapped in their own orthodoxy by financial constraint and, perhaps, by fear of the unknown). Thus the regulations have actively dissuaded innovation by closing off areas that a majority of teams are reluctant to investigate, for their own (no doubt internally sound) reasons.

As for the Parametric Governance model, what can I say? Conceptually it is extremely attractive and seems theoretically sound, though I struggle to grasp the idea of multitudinous perfect equilibria in 12+ dimensional space!

Tackling the "Landscape", I note the convenient spacing of the circuit markers in Figure 7, which illustrates the model's tendency towards diversity. Applying this to current F1, I feel that what we have in actuality is a single 'freak' circuit - Monza - whose characteristics are far removed from those of the majority of the others. A genuine question: Would such a situation lead to convergence at the lower end of the curve, or would the motive for divergence still be maintained?

As far as conformance over a season is concerned, you seem to have quite an open mind. It seems to me you have two options: to keep car specifications fully fixed from the first event, or to allow development and thereby potentially open the door for teams to build a car for each track. The latter, though, would surely result in a new form of convergence, would it not?

I wondered about the possibility of the governing body setting the "lines" slightly beyond the capabilities of current technology, leaving the teams' positions somewhat below the permitted optimum and thereby giving the teams incentive for in-season development to reach their maximum potential before the end of the year? For safety this would probably not be appropriate in all of the defined relationships but I wondered if you thought it a workable middle-ground between unlimited development and fixed specs - keeping costs under rather better control?

The question of adjustability for individual events is interesting, since suspension changes can have surprisingly big effects on downforce (through ride-height), tyre grip and wear, braking efficiency and so forth. I assume that the experts you have consulted have considered this and are happy that it doesn't undermine the entire model, though.

Having done some Game Theory in a past life, I am intrigued to consider how real actors would behave when presented with the model. Do you assume your rivals are going to, on average, congregate in the centre, and therefore pick an outside position? If you find that your first attempt leaves you crowded out by rivals, do you move to a new position with your second iteration - or stay where you are, knowing that your rivals will also want to move? And does this game eventually result in divergence, or ongoing (accidental) convergence through incompetence? The incentives for collusion between teams and/or the governing body would be immense.

I also wonder whether in reality there would be complete freedom to change your package of parametric choices dramatically - in the course of running a particular car to a particular regulatory package the team and driver will have acquired knowledge that would be lost if a radical change was to be made. In such a way divergence may occur much more slowly than expected, with evolutionary, incremental movement over a period of many years?

In summary, this excites me greatly and I believe in your idea. I am grateful to you for posting it here and look forward to the debate it will spark.
 
Richard,

I am still plouging through your document - what a breathtaking piece of work by the way - but two questions occur for now, with apologies if you have already answered them in the parts I haven't read yet:

1. Is there not a tendency to convergent car design anyway, that is nothing to do with the rules - i.e. the designers' natural instincts to copy the best of their opponent's ideas, either by simply looking and mimicking or by more, shall we say, underhand methods?

2. When it comes to maintaining divergence in the landscape, would it not be necessary to consider all the venues for a given season 'in the round', in order to provide a balanced mix of track characteristics, rather than simply adding or removing tracks according to commercial interests, as has been the case recently?
 
The concept of Game Theory raised by Galahad is an interesting one in this debate because it would suggest that the teams would design a car that would do them the least damage in the championship and therefore the most conservative in scope. For those unfamiliar with the prisoner’s dilemma which is at the heart of game theory it's basically two prisoners in a separate cells. If they both stay silent when questioned they both go free however if one blames the other for the crime then the other person is sent down and if they both blame each other then they are both sent down. So the best outcome would be that they would both remain silent and go free however for the individual the best out come is to blame the other person, since if he remains silent and the other person doesn't then he goes down and if he speaks out and the other person doesn't then he still goes free. The theory being that the best option is the one that does the least harm to the individual and not the group as a whole.

At the moment in F1 each team is trying to build the best car it can for itself to beat its rivals. Generally speaking if one team produces an innovation it's quickly copied by their rivals. This goes some way to maintaining the lack of performance gap as mentioned in the article. With Divergent Governance there still seems to be a risk that by matching your opponents design you are preventing him from gaining the upper hand.

Given an equal playing field, i.e. money, design talent, technology, driver talent and management there would very likely be several different solutions to the same problem however, the problem with F1 is that there isn't parity in any of the fields mentioned so these environmental factors are likely to have as big an effect on final car designs as the actual routes chosen. In the past in F1 we have seen radical designs appear in smaller teams however they have seldom had the finance or the depth of technical talent to progress these ideas to see them through. The haves and the have nots could potentially be more apparent in this new system since the reliance on design work would be far more important in order to explore the various avenues of development.

The concept of linking key areas of design to each other so that as you say, if you increase engine power you have to make a corresponding increase in weight is an absolutely fascinating one. The thing is, as shown in your simplest model of the oval track, with these links in place is there a risk that yes there are differing performance levels in various sectors of the track (the straights and the corners as you've written it) but at the end of the day the net difference between the two cars is zero? What happens if when the cars set off, car a is able to overtake b on the straight, car b overtakes a in the first corner, a then overtakes b on the next straight and b then passes a in the corner finally when they return to the same point then surely they will be exactly where they started relative to each other? Obviously that is the most basic and simple premise however for one car to win a race it has to be faster than the rest. If these cars line up in order of fastest first to slowest last, it doesn't matter about performance differentials in the end if it's still down to one car being overall faster than all the others.

That brings me back to my first point where teams will start out exploring one avenue of development until they discover that they are significantly slower than another team. They will then strive to copy what that team has done because it will move them past others in the same situation and bring them level with the team that is beating them. The remaining teams will then follow suit if they find themselves falling behind.

If you place a restriction on the development of the car then you could end up with the problem of a team making a mistake in the initial design being then being out of it for a considerable part of the championship. That's not going to go down well with fans, sponsors or the team itself.

The flip side to all this is that anything that can move motorsport design beyond the stagnation that it currently finds itself in and give designers more of a chance to show their stuff should be applauded and shouted from the highest mountain tops. I found while reading the paper that I'd thought of several questions along the lines of "a-ha, but what about........" only to find that a few paragraphs later the point had been answered.

In principle it's a fantastic idea and I don't want sound negative about it but every conceivable question needs to be asked to ensure that it's as valid an idea as possible if it's to have any chance of being pushed forward.

As I said in my earlier post, brilliant stuff.
 
cider_and_toast,

Have you heard of the Red / Blue game? I came across it a few years ago on a team-building day; it sounds very much like a version of the 'prisoner's dilemma that you describe.
Basically, we were split into about five teams of four. The game consisted of a number of rounds (let's say 12), during which each team had to choose one colour - red or blue. The umpire would go round to every team (which were all in separate locations, each out of visual or audible contact with every other team) and collect their selections.
Each colour attracted points, which were apparently weighted in favour of blue, roughly as follows: any team who chose blue in any given round would get two points and any team that chose red would get minus one point. But if all teams chose red, they would get one point each and if all teams chose blue every team would get nought. I may not have the exact scoring pattern correct, but you get my drift. Most teams would instinctively tend to favour choosing blue, especially since every individual had been asked to contribute £5 to the kitty before the game started, as it had the best chance of gaining good points. The kitty went to the winning team at the end of the game.
Now here's the rub: in the smll print of the rules was a clause that for the team with the most points at the end of the game to win, and collect the prize money, ALL teams must end the game with a positive score; otherwise the umpire kept the kitty.
Now one or two teams might have spotted the significance of this rule immediately, but several did not. But another rule stated that there were just two opportunities in the game for a representative from each team to meet and discuss tactics, say after rounds four and eight (remembering that there was otherwise no communication allowed between the teams - the umpire collected colour selections only at each round, and did not relay any information about other teams' choices) . Anyone could propose a meeting, but it could only take place if there was unanimous approval between the teams. Needless to say, in the game I took part in, no meeting ever took place and more than one team finished with negative points - so everyone (except the umpire) ended up losers.
 
Wow, what fantastic, comprehensive and intellegent responses. This is what I was really hoping for. I have got up early to simply rattle off a standard response or two, but I can see that to do you all justice I will have to go much deeper. I'm going to need far more coffee (not firing on all cylinders yet) so let me print off your inspired comments and I'll reply later, probably when the house has gone back to bed tonight.

I can take from your positive stances that I am at least heading in the right direction.

Thanks again,

Rich
 
Richard, a couple of other thoughts:

Firstly, I agree with cider_and_toast that the ability of teams to develop during the season is crucial. If a team were to be stuck with a car for a whole season that was as bad out of the box as the 2009 Mclaren, they would rightly feel that they may as well give up on the season. Similarly, if one team had a really dominant car, everyone else would be left fighting for second place. Also, as a spectator, I get almost as much pleasure and interest from watching the development race as I do the action on the track.

Secondly, I was glad to read your views on three things in particular, which I have been banging on about on more than one forum for ages; tyres, braking distances and downforce. I would be very much in favour of harder tyres, which I feel would improve things greatly due to better off-line grip from the lack of 'marbles', and reduced grip leading to longer braking distances and putting more emphasis back on driver input in cornering. When it comes to aerodynamic downforce, while I wouldn't necessarily go as far as to agree with you that it brings no benefits, I do believe that more needs to be done to reduce its overall levels, which would help reduce cornering speeds by placing more reliance on mechanical grip from the tyres and agin help increase braking distances to provide more overtaking opportunity for the 'late brakers'.

My reservation though, is this; if aerodynamic downforce is 'bad' and should therefore be reduced or removed, and if hard tyres are 'good' and should become the preferred choice, aren't these then examples of convergence rather than divergence? In other words, is there a risk that in a push to bring divergence in one way, we end up with convergence in another? And anyway, should it be taken as read that divergence is always good and convergence is always bad?
 
Right, the family have gone to a Halloween party, so I have some time. Firstly to answer some of the excellent questions posed by Galahad. I have included the first few words of each of Galahads paragraphs in italics and answered them in turn.

Firstly, I need

Always good to start with a compliment, you now have my undivided attention.

As such I don't

True we are not yet at the point where the computer has eradicated the human element, but it is plain to see that the computer is now providing answers to racing problems quickly and accurately, just as in aerospace. The differences between the solutions, to problems, that individual teams come up with are now quite small, and I think we are now at a decision point. Different people will always have different ideas about when the point has been reached, but is this important? The fact that it is either here now, not quite or been and gone, doesn’t alter the fact that a change is needed.
Regarding more frequent rule changes, this does upset the field (less and less as computers become ever more powerful) but, apparently, not as much as it upsets the teams as they feel compelled to throw money at rule changes.
I believe it is excessive downforce that is the problem. It is impossible to have non as this position is adjacent to lift, so the cars would be a little flighty to say the least. Lower levels of downforce needs to be controlled directly (parametrically) but remain in a relationship with other parameters so that it can contribute to the performance differentials and be a huge contributor to the visual diversity of a series.

As regards your piece

Gosh, even my toes are blushing. During the many late evenings, and very early mornings, spent putting this together, I wondered a thousand times if I was actually suffering some form of madness and was completely deluded. The positive responses I have since been receiving from the industry and the forums, have justified my efforts.

Moving on to the content

Again, I agree that the tipping point has not yet been reached, but would we see it coming or just find ourselves well and truly ‘in it’ so to speak. Formula 1 must avoid the tipping point at all costs, they are usually quite devastating for a business; as you say, look at Indycar.

I also agree with the four

There is very little to add to this statement. Perfect.

As for the

Yes, it is difficult with our 3D brains (built for a 3D world) to think in 4 or more dimensions, but just knowing that it is possible must suffice. Good luck to the design engineers! Isn’t this a great asset?

Tackling the "Landscape"

The tracks have converged for the reasons discussed on page 30 of the doc. and yes, the first few Divergent seasons will be run on a less than ideal landscape. However, once downforce has been successfully reigned in, the tracks can begin to metamorphose into the necessary varied backdrop for a diverse series. I’m convinced that the removal of one or two chicanes and the opening up of some radii will massively diversify the landscape. The subsequent adoption of more radical ideas such as returning Hockenhiem to a fantastic mix of speedway and stadium, plus re-commissioning some of the tracks not destroyed by Formula 1, will follow once the concept has proved itself.

As far as conformance over

I think I may have not made myself perfectly clear on this point. I also agree that, under a pure parametric system, allowing the teams to develop would result in a form of convergence (at each track) and this would not be good. The pure parametric system illustrates the absolute control of costs that the system attains, through addressing the root cause directly. Whether this degree of cost control is desirable is another question altogether. I personally prefer the solution discussed in ‘Formula 1 incorporation strategy’ (page 48). This utilises the fundamental mechanism that is the foundation of total cost control in a way that can absolutely direct, with supreme precision, where the spend race should be focused; the consumption of resource. This is the key benefit to true root cause identification, you then have the power to do with it what you like!

I wondered about the

This would work for certain relationships where both parameters are difficult to achieve, such as high power and low weight, but many relationships may result in convergence at one end if the other is too hard to reach, e.g. the inverse relationship between tyre allowance and downforce (page 47) could mean that if the number of tyre sets was too low most would go to the end with more tyres available. This is something that could be tweaked slightly each season to keep the carrot in sight but just, ever so, slightly difficult to reach. Also, if Divergent Governance begins to gather pace and some real life, actual race car designers become involved, my hope is that they can come up with relationships that I hadn’t thought about that would be challenging. Time will, hopefully, tell.

The question of adjustability

For a fully parametric framework, something similar to what is already applied at any spec series (like Indycar) would likely suffice. The exact degree has not been discussed yet as, at this stage, it would be irrelevant.

Having done some Game Theory

I cannot begin to tell you how excited I am by this aspect of the whole concept, it would be a completely new sphere of intrigue and interest within motorsport. Each new season car launch would create more buzz than a world cup final, and that’s before a wheel had turned in competition. My money would be on a varied field that oscillates, on a several year cycle, in many different directions. And the dream scenario would be that every now and then a team would show up, from nowhere, with a solution no one had even thought about. Take a look at the feature on the website regarding Cambridge University. An economist and a game theorist are about to start work proving the stability of diversity within the framework I have proposed.

I also wonder whether in reality there would be complete freedom

Yes, as per the answer above. It is for this reason the barriers to entry must be kept low so that new teams can easily bring new solutions. Also, teams that are doing well will not feel the need to experiment as much as the lower teams do. A team with a limited budget would be more hungry for the niche with the least near neighbours and spend more time experimenting to find it.

In summary, this excites me greatly and I believe in your idea.

Thank you for your very thoughtful comments and encouragement. I truly believe that Divergent Governance could be the game changer that has been needed for so long. It still requires much work but what I hope to achieve from these forum posts is fan power. If we really believe that this could be a great idea it needs to be heralded as such and shown to as many people as possible within the motorsport world.

Thankyou again Galahad.


Next is Chad

Is there not a tendency to convergent

Human nature will play the part you describe, but the rewards for diversity should soon prove this to be a poor strategy.

When it comes to maintaining

Ideally the tracks in a season will be diverse and challenging. As you point out, commercial consideration may mean, at least initially, that the tacks may not be ideal, but if the concept proves itself the will to alter the tracks accordingly will be met.



Hello Cider and Toast.

Re; Game theory. A perfect analogy, please take a look at the website for an introduction to the work about to be started at Cambridge.

At the moment in F1

Successful innovations are copied in a convergent framework as all teams are trying to answer the same question. The penalty for matching your opponent in a divergent series would be equally shared between the copied and the copier

Given an equal playing field

I’m hoping that the haves and the have nots would be less apparent, as there would be less advantage to spending huge amounts and more to innovative thinking, which is free. Also, the nature of Divergent Governance should reduce the tendency for a few teams to dominate at all tracks and allow teams on smaller budgets to do well at tracks that suit their design niche.

The concept of linking key

Yes, in that elementary example the lap time are identical, but that was necessary to illustrate the concept of how to achieve a performance differential within a level arena. Therefore the winner would be there on merit. From this basic building block we suppose a series where the theoretical, sum total of a teams best laps, for a whole season, must be equal. Thus, different cars will perform differently at different tracks, while maintaining a fair and level arena. The overtaking will come from the fact that each teams near niche neighbours will be trading advantage around the circuit, resulting in many battles up and down the grid.

That brings me back

When the relationships are developed the assumption will be that everybody will produce excellent designs that push known technology to the boundaries. This should result in a curve that rewards excellence wherever that may be. If a team is not doing so well it will be because they are too close to other teams and diluting their opportunity for points, or that they simply are not very good. Following the leaders design strategy will not help in either case.

If you place a restriction

Yes, this is dealt with on page 41. This is one of the reasons why the Formula 1 incorporation strategy enables development within the relationships for consumption of resource.

The flip side to all

Thankyou.

In principle it's a

I agree. Indycar will be debating this concept very soon but the more fans shouting for it the more impetus the will be to overcome the difficulties. As for Formula 1, we are at the bottom of a very large, steep, rocky and unforgiving hill, hope you’re up for it!!


The feature on the Red / Blue game is also interesting and relates to the comments on game theory.

So, if OK with you Chad I’ll go straight to your comment beginning;

My reservation though

I think that the range of diversity needs to reflect what is best for the sports show. Have aero downforce but the range needs to be from as low as is practical and safe, up to maybe no more than 1 or 1.5g. The same goes for grip, have a range of tyre allowance but the upper end still requires quite hard rubber.

And yes, convergence BAAAD.


That was very thought provoking, and I know it can sometimes take more than one reading for the kernel of the concept to be fully realised in the imagination, but once you find you can speak ‘Divergent’ all of these difficult elements drop into place.

Great stuff guys and hope to hear more thoughts soon.

Rich
 
Back
Top Bottom