Thursday, November 22, 2012

On Making Rules....

You may be aware that I am a member of the Super Touring Advisory Committee for the Sports Car Club of America. The STAC's responsibility is make recommendations to the Club for regulations for that category.

I wrote the following on a racing forum some years ago, and I thought it bore repeating, with some minor updating...

How to write a rule


We seem to get into a lot of arguments about various rules. “This rule says this” or “this rule says that” or whatever. And, typically, the basis of this argument is different strokes, different folks; one person read is another person’s cheat.
I see the root cause for all these arguments centering around one thing: poorly-written rules.
Yep, sorry if it pisses off the rulesmakers, but I think I can credibly argue that most disagreements center around either people reading into the rule what wasn’t intended (maybe based on a failure of the rulesmakers’ imaginations) or poor use of verbiage to describe the intent. So, here are Greg’s Tips to keep in mind when you’re trying to write a rule.

Tip #1: You can’t POSSIBLY think of all situations


Consider this: there’s a handful of you sitting around a table (or talking on a teleconference) trying to find the best way to write an allowance to the rules (e.g., struts, suspension bushings, engine mounts, whatever). Do you REALLY believe that you're smarter than those collective brains? Do you believe you can think of all possible permutations that the rest of the world can come up with? Of course not. The masses, as massive as they are, have a collective imagination that simply dwarfs your small collective's. Ergo, you are insignificant when it comes to thinking of all possibilities.

Tip #2: If It Doesn’t Say You Can, Then You Cannot


Glory be! That little bit of sunshine is your savior. The "IIDSYCTYC Rule" (GCR 9.1.3.D) is the one rule that can pull your butt out of a fire. You, as a rulesmaker, have the ultimate authority on allowances, because until and unless you allow it, it cannot be done, period! But, you absolutely must keep in mind...

Tip #3: The Roffe Corollary


The Roffe Corollary (attributed to a friend a former racer, George Roffe) states simply, "If it says you can, then you bloody well can!" Whereas The "IIDSYCTYC Rule" is your friend and savior, the George Roffe Corollary is your enemy and potentially your Achilles’ Heel. Many a rulesmaker has been humbled by seemingly simple words that opened massive doors to rules failures (e.g., remote reservoir shocks, splitter and undertrays, spherical suspension bushings, open ECUs, D Sports Racer records).

Tip #4: With Tips #1-3 in mind, describe only what you want to ALLOW, not what you want to restrict.


Remember, tip #1 says there is no possible way that a handful of guys on a telcon can think of all permutations, therefore when you use the magic words "unrestricted" or "open" or "can be replaced" and so forth you've now, with your magic wand, reverted that rule to The George Roffe Corollary. No longer is it IIDSYCTYC, it’s now "whatever you want to do within the restrictions listed below".
At this point, see tip #1, 'cause you just opened a biiiig hole (insert loophole-driving-truck-through reference here). Do you really think you can fill that hole sufficiently? Besides, there's no need to describe what you can't do, 'cause you've already got it: the IIDSYCTYC Rule.

Tip #5: See Tip #4.


If after writing your rule you still feel the need to start adding in restrictions, then your new rule isn't worded well; go back to Tip #4 and try again.

If you accept and understand your humility, and keep these simple tips in mind, I absolutely believe you cannot go wrong.

-----------------------------------------------------------------------------------------------------

October 2012 Addendum

All rulesmakers should take a half-hour and listen to this podcast. It really hits home toward the end...

The Cobra Effect
http://www.freakonomics.com/2012/10/11/the-cobra-effect-a-new-freakonomics-radio-podcast/
I think you start by admitting to yourself that no individual, no government, is ever going to be as smart as the people who are scheming against you. So when you introduce an incentive scheme, you have to just admit to yourself that no matter how clever you think you are, there’s a pretty good chance that someone far more clever than yourself will figure out a way to beat the incentive scheme.

November 2012 Addendum


I had a spirited debate with someone recently regarding regs, letter of the regs, the spirit of the regs. It's something that we like to lean on when trying to "interpret" a reg, and I suspect our direction-of-leaning tends to change with whatever side of the argument we're on.

In my opinion, the responsibility of the "spirit of the regulations" lies solely on the shoulders of the writers of the regulations. It is purely the responsibility for they, once they determine the "spirit", to codify that "spirit" into the "letter". Any subsequent failures of the reader of the regs to read them - or to successfully "intorturate" them - into something that was not intended is purely a failure on the part of the rules writers.

This came to the fore in recent events. I write this post the week after the first F1 race in Austin, the race where Ferrari intentionally broke the FIA seal on Massa's transmission so that he would get a 5-place grid penalty. By doing this it allowed Massa's teammate, Alonso, who was fighting for the championship, to get moved over to the "clean" side of the track for the start (and they did it immediately prior to the start so that their main competitor, Red Bull, could not pull the same trick and move him back over). As I write, many are arguing that while this was within the "letter" of the regs, it was not within the "spirit" and thus warrants penalty.

Poppycock.

Tony Dodgins wrote a good column on this issue recently in Autosport. He discussed this scenario and also compared it to the 1975 Glen race where Regazzoni - being lapped - blocked Fittipaldi so that Clay's teammate, Lauda, could gain an advantage (before Clay eventually got black-flagged) and subsequently allow Lauda to win the race. Dodgins noted that the argument about the "spirit" of the regs was brought up, that while Regazzoni broke no letter of the regs, he violated the spirit of the regs. Dodgins replies:

Rightly or wrongly, the fact is that anyone who believes in the 'spirit' of the F1 regulations is being naive. There are only the regulations. Period.
I concur with Dodgins. Remember what I wrote in prior posts about "failure [on the part] of the rulesmakers’ imaginations"? And the "you can't think of everything"? That's where these "spirit" arguments originate. No one person or group can beat the computing power of the masses, and no regulators can possibly figure out what all the masses can come up with. It's a Sysiphian task.

But that doesn't mean that regulators, due simply to their lack of foresight, should be intentionally made powerless to change the regs to meet the original intent ("spirit") and cover that new "intorturation". In fact, it is the duty of the regulators to re-think their "letter" positions and make a conscious choice to change that letter to meet the original intent; to do otherwise is an active allowance for the "spirit" to change and is a derelict of duty.

F1 regs were subsequently changed to ban Regazzoni's tactics; any driver that intentionally impedes the flow of another driver like Clay did in 1975 will not only get black-flagged but may get a significant start penalty in subsequent races -- or worse. And don't be surprised if a change to the "letter" of the regs about breaking seals on drivetrain components is forthcoming...

GA

Sunday, March 11, 2012

Automatic Transmission versus automatic transmission

Within the aviation community there's a spirited debate about the term conventional landing gear. Traditionally, the term refers to taildragger airplanes, those that have a tailwheel, like a Piper Cub, the Boeing B-17, and the Douglas DC-3. But since around the time of World War Two most aircraft have been manufactured with "tricycle landing gear" or a nose gear. Almost every airplane you see today is a tricycle with a nose gear; short of Homebuilt or other purpose-built aircraft (e.g., agricultural, aerobatics) you would be hard-pressed to find many airplanes with a tail wheel.

Yet despite the fact that most aircraft manufactured within the last half-century use tricycle/nose gear, the term "conventional landing gear" continues to apply to the tail wheel. This is confusing to people new to aviation; to them "conventional" is what they see out at the local airport every day like the Cessnas and Cirrus, or every airliner they'll board, all of which have nose wheels. Yet that traditional term persists.

So it is with automatic transmissions in cars today. Mention "automatic transmission" to me (I'm approaching a half-century myself) and I immediately think of the old GM Hydramatic transmission, with a torque converter, an internal pump, clutches and bands, a valve body, etc.Those are truly your father's automatic transmissions, the one where you moved the lever from "P" into "R" then "D" and then never touched it again except to put it back in "P" when you got home. Push the accelerator pedal to go; push the brake pedal to stop, no clutch pedal or shifter required.

On the other hand, mention "automatic transmission" to someone just learning to drive, and their mind may not necessarily turn to torque converters, valve bodies and such; they skip all those technicalities and go right to thinking "how do I go and stop, and do I have to touch the shifter again until I'm home?" Yet it's quite possible that these new drivers are learning in automatically-shifting cars that, from an engineering perspective, have much more in common with the old "4-on-the-floor" than they do with Granddady's TorqueFlite.

We still call these new technologies "automatic transmissions". But are they, really? What makes a transmission an "Automatic"?

Is simply the lack of a clutch pedal that makes something Automatic? If so, then what about the Formula One cars that do not have a clutch pedal, instead using an electronic button on the back of the steering wheel to activate the clutch? And what about before that manual button was required by the regs and they let the computer activate the clutch? Would you say that Formula One cars were/are using Automatic Transmissions?

Is an Automatic Transmission defined by a lack of a stick/shift lever that has to be rowed to manually select gears? If so, then that would again support the idea that Formula One cars use Automatic Transmissions, given their gears are selected using electronic "flappy-paddles" on the steering wheel instead of a shifter.

But if a Manual is all about having a clutch pedal and shift lever, then what about World Rally Championship cars that have sequential-shift gearboxes that don't use a clutch between shifts? If you don't have to use the clutch to shift, is that an Automatic? What about prior to 2011 when WRC cars were allowed to use those "flappy-paddle"selectors for their sequential-shift boxes instead of a gear lever; did Sebastian Loeb win numerous WRC championships with an Automatic Transmission?

But if the method of the engine engagement and how gear are selected  and a clutch pedal and shift lever are not needed to be considered Manual, where does that line cross to being considered Automatic? VW's DSG and BMW's SMG (as well as many other high-end super cars) use "conventional" flywheels with clutches, and internal gears and synchronizers. Yet none of them have that third pedal, all use computer controlled solenoids to engage/disenage the clutch and select gears. All have some form of manual over-ride giving the driver ability to select the desired gear, with computer backup to avoid human error such as the dreaded 3rd-to-2nd "upshift". But so did the Turbo Hydramatic...

If you contend that ones does need a clutch pedal, a shift lever, and one must be forced to move that lever and clutch pedal between shifts, is the Ferrari Enzo with the flappy-paddle an "automatic transmission" car? If gears/syncros, a shift lever, and a clutch - but no clutch pedal - are all required to be considered Manual, what about the old VW/Porsche  Sportomatic, where the conventional clutch was operated by a vacuum solenoid, actuated by vacuum switches mounted at the base of the shifter?

And how do CVTs (Continuously Variable Transmissions) fit into this mix? What about pure electric vehicles that may not have any gears at all?

Automatics have historically suffered from disdain within enthusiast communities because of their lack of sportiness. They typically had fewer gears than manual transmissions and the torque converter, pumps, fluids, bands all robbed efficiency from transferring power to the tires. Automatics also received disdain from motorists interested in efficiency, as they almost always had relatively-poor fuel economy for all the same technical reasons. But automatic transmissions were easier to drive and demanded less from the motorist, so their popularity has consistently increased.

Today, technology is blurring the lines between the traditional/conventional ideals of Automatic and Manual Transmissions. With the newest technologies of automatically-shifting transmissions, both the enthusiast and efficiency fans are finding these new designs meet both their needs, and more manufacturers are offering them to improve fleet fuel economy.

But are these new technologies really Automatic Transmissions? So what does define a transmission as "automatic"? Is it the technology in the box, or is it how the driver interacts with the machinery? Just as debates over "conventional" versus "tricycle" landing gear continue, I suggest the debate over what makes a transmission "standard" or manual or automatic or big-A Automatic will rage on. I also suggest that within the performance enthusiasts' mindset, the presence of a clutch pedal may be the milestone between automatic and manual transmissions (though they'll conveniently ignore their favorite super-car's flappy-paddle no-clutch-pedal gearbox). Regular motorists will likely make that same distinction, that being the absence of a clutch pedal and the ability of the car to pull away from a stop and shift without driver intervention.

And me? Being the motorist/enthusiast/racer/engineer that I am -- and a new, unabashed fan of Volkswagen's DSG -- my differentiation will go right back to the old Hydramatic transmission, with the emphasis on the "hydra" part. In my mind, if engine torque is transmitted through a hydraulic fluid via torque converter, bands, and valve bodies, I will forever consider it to be an Automatic Transmission. If power is transmitted through a flywheel, clutch, and meshing of gearsets, it's a Manual Transmission, regardless of how those mechanisms are engaged. I'll gladly let my street car automatically pull away from a stop and shift the gears for me, but I demand the ability to over-ride the computer and shift it myself.

These new technologies are pretty good, so good that someday I predict we will mourn the death of the good ole clutch pedal; those driving skills will eventually go the way of the manual choke and spark retarding, and the hand-cranked starter...

GA

Tuesday, January 17, 2012

Damn the Economy! Data Acquisition Engineers Needed

So I've got an opportunity to do some data acquisition for a pro team, guys that I consider personal friends. Problem is, with a 13-race schedule and a lot of testing, I'd need to devote more time off work than I've got available in vacation time (not to mention needing to keep some for my wife and otherwise-personal life). So I've been trying to find some good guys to suggest to the team for a full season of data engineering.

No one to be found.

As you are no doubt aware, data is big in motorsports, one of those former "unfair advantages" that, as they all do, has moved into "absolutely necessary". I was told by one data guy that I called that 2012 is going to be one of the largest Grand-Am fields in a long time, if not in history, and teams from the Continental series, Rolex GT, and DP are all looking for data guys (and this ignores all other series!) And there's just few good ones to be had. This guy I mentioned has been getting calls from Motec to support their customers, and I'm aware of other manufacturers looking for customer support for their products, too. There's only so many "few good men" available.

I've got my ideas on why this is the case. First, it's all contract work; except in the rare cases of the real top-end teams, data guys are part-time "travel crew", where they show up for the race and test weekends. The teams simply can't afford to keep someone like that on board full-time. As a result, most of these guys (like me) just do this part-time within their own lives, with "real jobs" during the week.

Second, the economy hasn't exactly been booming, so team racing budgets are strained. When that happens, data guys are typically low on the funding priority pole, after testing, development, so forth. Yes, we all know the value of a good data guy toward racing success, but we're in the minority. A lot of racing teams tend to be as short-sighted as other business. As a result, there wasn't a lot of work available so data guys made some career changes to do other things.

Third, the lifestyle of a data guy - hell, of racing in general - doesn't exactly appeal to everyone. Race schedules are tough, and you gotta be flexible.

Interested in the work? Here's what it takes to be a good data guy.

Personality and Communication Skills

First and foremost - more important than anything else - you must be able to work with people.

The job of a data guy (and I use that term generically; I know some data gals that kick ass) is to download, sort, and interpret data from the race car. But much, much more importantly than that, his job is to interact with all parties to tell them what that data means. To do that, you must be able to interact with personalities that spread the width and breadth of human nature. You'll be talking to mechanically-minded technicians working on and setting up the car, with team managers looking for things going wrong, with chassis and aero guys wanting to know where to change the setup, and with race car drivers (from dolts to supermen) that are looking for that edge. You have to be able to comprehend their position, understand their moods, and present the information to them on their level, using their language. The car may "talk" to the driver while on the track, but the data is the objective judge that has the facts. You are, in effect, the language interpreter between the race car and all the other people involved in that race car.

Further, you have to get along. Race teams are a group of people that, like a band, live and work closely, occasionally for long periods of time. You have to be able to get along with the roadies as well as the talent, and you have to tolerate all the foibles including Ian Faith's cricket bat.

If you can't work with people, knowing the ins and outs of all the hardware and being the team's biggest techno-nerd will get you absolutely nowhere.

Technical Ability

But, that said, knowing the ins and outs of all the hardware and being the team's biggest techno-nerd is very important, too. Knowledge of computers, software, electronics systems, data transfer/storage, are all important. You are going to be the guy that has to gather, correlate, and organize all the information. The first thing a driver is going to ask when he gets out of the car is "what's the data say?" and right behind him will be the crew chief asking the same question, with the team manager right on his butt heading your way. And they want it NOW.

You better know how to get to it.

Knowledge of Motorsports and Vehicle Dynamics, and an ability to "Visualize" the Data

There are a lot of books out there on data acquisition. I've got a few suggestions if you're interested. But when it all comes down to it, you can teach technical ability to get to the data, and you can teach how to run the software and make the pretty pictures, but I'm not clear that you can teach vehicle dynamics and truly express what the data is saying in terms of car control.

All those squiggly lines and tables of numbers mean nothing unless you can be the interpreter for what they are telling you. What does that line moving so quickly in that section mean? Why is that table data flat-lining and staying the same during the compression on the hill? Why is the data missing there? You can't just say "oh, look, that line is moving down and up fast while those two are moving, too" you need to be able to say "you are making a quick throttle lift right as you're turning in for South Bend, causing corner weight transfer which is upsetting the chassis. This could be why you're losing 10 mph through there versus your co-driver." You have to take your personality and communication skills with your technical ability and knowledge of vehicle dynamics and put it all together into a reasonable presentation of what's going on. After all, the only thing people REALLY want to know is just that: "what's happening?"

Do you naturally talk with your hands on an invisible steering wheel with your right foot on an invisible throttle? Having a motorsports background, even an amateur one, is not absolutely required but it goes a long way toward being able to interpret the data into real-world scenario.

Humbleness

You're not in a competition with these guys, you're all on the same team. You don't have anything to prove; if you're doing your job right, then people will know. You are the silent member of the band, the trombonist or bass drum player way in the back of the symphony, the one providing the support, not the one out front of the stage waving the violin around. If the symphony wins, you've done your job, no need to bang the bass drum louder so the audience notices you. Sure, you'll give your opinion, but most of the time they'll ask you for it if they want it (and if they do trust you for your opinion, that's a clear sign Yer Doin' It Right). But in the end, it'll be you, the data guy, that has to convince the Pilote Prima Donna that he is, in fact, lifting where he thinks he's not...but in a way that he believes it (and maybe in a way that convinces him he thought of it. See "Personality".)

Business Acumen

Let's face it: this is not a full-time job. However, with the right promotion, time management, and business sense, it might be a lot of part-time jobs rolled together into a more-than-full-time job. Parlayed right, you can get a lot of work out of this.


At the same time, understanding how the motorsports business works helps in a lot of decisions about going forward. It's nice and all to say "you gotta buy those new pimpy shocks" but it's better to present it as "those $2000/corner shocks should give the drivers confidence to gain up to a second per lap, which would put you on the podium next race".

Confidentiality

If you do work with multiple teams, you must, must, must maintain a Chinese Wall between all of these teams. I've not seen too many teams that require signing of Confidentiality Agreements (amazing) but to a good data guy those are totally unnecessary. Your reputation is your value, not what some team manager thinks he can glean out of you for what you did for the other team. Just like talking bad about other people behind their backs, your revealing other people's data or information to a team, even for the best of reasons, immediately implies to them that you're willing to do it the other way.

Their asking is just part of motorsports, you should refuse outright, without condition, and walk away if they insist. I personally don't even bring data from other teams with me to the track, just in case someone gets access to my PC.

Flexible Work Schedules

It's motorsports, and all that that implies. You'll be called to fly to the track "tomorrow", and you'll be working dawn to well into the night. Your work isn't done until the car is on the trailer and way down on the highway.

If you're a Monday-through-Friday, 9-to-5 kinda guy, this just ain't your game.

So What's the Benefits?

It's motorsports, and all that that implies. As a contract employee you'll get either per-hour or per-diem pay. You'll get your expenses paid to go to some pretty damn cool races and race venues. You'll, hopefully, be part of a successful motorsports program that wins races and championships, you'll get to bask in the glory of success, and you'll be able to tell people, "yeah, I was part of that team".

And you'll get to hang with some pretty damn cool folks.

Data acquisition is, in my mind, a rare piece of motorsports where there's a high demand but a low quality supply. Unlike being a driver, you'll get reimbursed for your efforts, and if you're better than the rest you'll gain a favorable reputation and actually rise through the ranks based on that talent. Maybe it's not as satisfying as being a driver - it's certainly not going to gain you the fame and access to the winner's trophy podium - but in the end you're still a key piece of that success and you're far better off going home with reimbursement for your efforts.

So, who wants to be a data guy...?

Greg