|
| |||
 |
The Best Season Ever: The Methodology | ||
| by Peter Goodchild and David Southworth, England | |||
|
1. INTRODUCTION
In this Annex we discuss in more detail the factors used and their weightings. We also discuss the factors we discounted. Once calculated, each factor has been standardised onto a scale from zero (representing the minimum value over the period) to one (representing the maximum). In this way the weightings adopted become meaningful measures of the importance of each factor. We have chosen the weightings to reflect our judgment of the strength of each factor in determining whether or not a season is exciting and, for simplicity, have ensured they total 100. We have made the assumption that Indianapolis is not a Grand Prix, and that its results would skew this study unnecessarily, were we not to expunge its participation in the 1950-60 championships. 2. THREE CATEGORIES OF FACTOR The 11 factors used can be split into three categories - Variety, Closeness of the Championship, and Competitiveness of Racing. A. VARIETY A1. Number of winners The number of different drivers who win races during a season is a good measure of the competitiveness of Formula 1. In a season with few races or drivers, you would naturally expect fewer winners, so this number has been divided by the expected number of different winners where all drivers are evenly matched, E:
E = DR - (D - 1)R where R = number of races, D = number of drivers. We have used the average number successfully qualifying. Weighting = 12.5 A2. Number of point scorers A similar measure of the sport's competitiveness. We have not modified this number, feeling that it is a reasonable measure of the strength in depth of the field. Although this slightly penalises early years, with fewer races, this is far less the case than with number of winners. Each race provides five or six (or, in the days of shared drives and points for fastest laps, sometimes more) point scorers, meaning that even a few races afford the opportunity for many drivers to score. Weighting = 10.0 A3. Maiden constructor wins A new constructor emerging, or an old one progressing, to take race wins can represent an exciting breakthrough and even a shake-up of the old order. Seasons in which two or more constructors emerge in this way are indeed rare, but certainly include extra spice. We have not excluded constructors who had race wins previous to 1950, or outside the Championship. Although there were inevitably more instances in the early days, it only took a handful of years for the order to settle, and for such occurrences to near their modern day significance. Weighting = 7.5 A4. Number of constructors competing A large number of constructors putting out cars is a good indicator of a healthy sport. There is interest and excitement in witnessing many smaller constructors competing against the odds and the might of Ferrari, McLaren and Alfa Romeo. Weighting = 5.0 B. CLOSENESS OF THE CHAMPIONSHIP B1. Number of lead changes in the Championship The Championship lead changing hands a number of times usually makes for an exciting season. To account for the length of the season, we have divided this figure by the total number of races. We have treated all instances of two drivers on equal points as a shared lead. A driver coming up to a shared lead then subsequently falling back, without ever taking an outright lead, is counted as one lead change. Weighting = 12.5 B2. Point at which the Championship was decided A nail-biting title decider is undoubtedly an ingredient for a memorable season. Conversely a driver wrapping up the title with races in hand, although noteworthy, does little for the heart rate in the latter stages of the season. We have expressed this as the percentage of the total number of races elapsed when the Championship was settled (eg a Championship undecided until the final race scores 100%; one settled in the sixth race of eight scores 75%). Weighting = 7.5 B3. Number of drivers in Championship contention at the last race This complements and refines the previous factor. When a driver is mathematically in contention but is unable to compete (such as in 1982, when Didier Pironi suffered a long-term injury), they have been excluded from this factor. Weighting = 7.5 B4. Driver and Constructor Championships won by different constructors This situation usually indicates two constructors genuinely battling for honours, as opposed to recording only a few good results. We have awarded a score of 1 when this occurs, and 0 otherwise. Prior to 1958, we have calculated a theoretical Constructors' Champion in the same way as was subsequently used - adding the points for the two best drivers of each constructor for each race. Weighting = 5.0 C. COMPETITIVENESS OF RACING C1. Number of lead changes during a race The amount of overtaking seen is one of the key factors of exciting racing. It is also very difficult to measure historically. We have taken the number of times the end of lap leader changes during each race, and used the race average for the season. This is by no means a perfect measure. Multiple lead changes within a single lap will be missed. A lead change caused by a spin or mechanical failure arguably makes for less exciting racing than one through a legitimate overtaking manoeuvre. A temporary change of lead through routine pit-stops certainly does, although the added excitement of possibly differing pit-stop strategies in part compensates for this anomaly. Careful inspection of the results from this measure have convinced us that, although imperfect, it does indeed provide a good indicator of race excitement. Weighting = 12.5 C2. Closeness of finishes We have measured the closeness of races through a points system - 5 points awarded for every car finishing within 1 second of the leader, 4 points for those within 2 seconds, down to 1 point for those within 5 seconds. We have taken the total points divided by the number of races. Weighting = 10.0 C3. Average grid position of race winners The rare occasions when a driver wins a race from way down the grid are invariably memorable ones. Even a winner coming from the second or third row usually indicates an interesting race. By taking the race average of the winners' starting grid positions over the season, this factor aims to measure that effect. Weighting = 10.0 3. DISCOUNTED FACTORS We considered and discounted many factors. Some of these were rejected in favour of better alternatives (eg Number of race winners starting below 6th on the grid, replaced by Average starting position of winners). Others we simply felt added too little to the measure of a good season (eg Maiden points for drivers or constructors). But there are two factors we were desperately sad to consign to the analytical trashcan. Both were excluded because of the subjectivity required in their measurement. Political Intrigue This factor would have been fun to measure. But then how on earth does one rank the FISA /FOCA wars of 1980-1982 against 1999's 'Bargeboardgate' or 1978's Fan Car scandal? And then what about the sportsmanship shown by Stirling Moss when he effectively gave away the 1958 World Championship, by protesting the exclusion of Mike Hawthorn at the Portuguese GP? All impossible to measure. Generally the inclusion of this factor would have worked in favour of recent years, when the ruling body, be it the CSI, FISA or the FIA, has been more proactive. Personalities We wanted to measure the degree to which the drivers (and teams) involved captured the imagination. But how would one objectively and statistically measure the Fangio/ Ascari/Farina rivalries or the Senna/Prost/Mansell/Piquet odyssey? We hesitated even to attempt this, as other studies have done, as we wanted to keep our ratings thoroughly objective. We toyed with a measure based on past championship successes of the leading championship contenders. However, we bowed to the inevitable, realising that this in itself was insufficient, as well as creating a significant anomaly in the early years of the championship which would have been difficult to correct. The inclusion of such a factor would have given a boost to some of the low scoring years - 1957, 1992 and particularly 1988 would all have profited. |
| David Southworth and Peter Goodchild | © 2000 Kaizar.Com, Incorporated. |
| Send comments to: david.southworth@dfee.gov.uk | Terms & Conditions |