Ranking F1 Drivers

Last week tennis fans were aware that Switzerland's Roger Federer
had broken the record for most consecutive weeks ranked number one
on the ATP computer.

http://www.atptennis.com/5/en/rankings/entrysystem/

Federer has now been ranked number one in the world for 162 weeks.
He bettered the previous record set by Jimmy Connors in 1974-77.

Women's tennis has a similar ranking system.

http://www.sonyericssonwtatour.com/2...es_numeric.asp

Several other sports operate ranking systems along similar lines. For
example:

Golf http://www.pgatour.com/r/stats/current/186.html
Soccer http://www.fifa.com/en/mens/statisti...b-2007,00.html
Cricket http://www.icc-cricket.com/icc/rankings/lg.html
Snooker http://www.snooker.org/rnk/

Formula 1 doesn't have (or need) a ranking system, but it might be interesting
to create one.

Actually there is such a system available on the FORIX website - it's called
the Wildsoft F1 Ranking - but, you have to subscribe to FORIX to be able to
see it.

http://forix.autosport.com/stats.php?l=0&c=1205

The Wildsoft system was devised by Dimitry Cherepovich, and is calculated over
the last five years of F1. Only data from GPs within the last five years are
included, so the 2002 Australian GP has just been "dropped" (actually, the
rankings haven't been updated since December, but I'm sure it will be soon).
Currently the system has 88 GPs in its database. Points are awarded for the
usual F1 achievements - wins, pole positions, fastest laps, etc., but also
include other criteria such as distance raced.

I wondered if I could "borrow" some of the features of the Wildsoft system
and produce my own version.

The first change I decided on was to drop the five year period and switch
to the last 100 GPs instead. This is easier to manage mathematically, and
I always prefer things to be simple. Two other "rules" I have adopted from
Wildsoft are that no points are awarded UNLESS a driver actually starts a
GP; and that any driver who gets disqualified also fails to gain any points.

Increasing the number of GPs in the database to 100 means that data is
used going back to the 2001 Austrian GP (13 May). That race was won by David
Coulthard, but clearly it should have less "value" than Felipe Massa's
Brazilian victory last October. Therefore the older races count for less
points than more recent ones and races in the current season are more
valuable than races last season (or older). Thus my rankings are based on
performances over the last five years 10 months, but if the number of
races per season continues at its present level the window will shrink.

It should be obvious that Michael Schumacher will be ranked as number one,
but my first question to ponder is, how long before Fernando Alonso (or
someone else) overtakes Michael. The 2007 drivers will all be adding to
their points totals as each race of 2007 is completed, but MS (and other
retired drivers) will steadily lose points as the oldest GPs are removed.

OK, preamble over, here are my current rankings.

GPs - number of GPs counted for each driver (from the last 100)
PP - number of Pole Positions (in the last 100)
W - number of wins (in the last 100)
FL - number of Fastest Laps (in the last 100)
Points - these are my points, allocated for various things, not just
the wins, poles, f.laps
(+/-) - the change in points as a result of the most recent race
Seasons - the seasons in which the driver competed (last 100 GPs only)

Numbers in brackets () in second column indicate the change in position
as a result of adding the most recent GP.

First race (GP #1) - Austrian GP, 13 May 2001
Last race (GP #100) - Brazilian GP, 22 Oct 2006

Driver GPs PP W FL Points (+/-) Seasons
================================================== =================
1 M Schumacher 100 32 44 33 6505.19 (-17.71) 2001-06
2 F Alonso 81 15 15 8 5055.20 (+54.67) 2001/03-06
3 K Räikkönen 98 11 9 19 4035.21 (+62.04) 2001-06
4 R Barrichello 98 10 8 12 2958.73 (-28.39) 2001-06
5 JP Montoya 86 13 7 12 2647.00 (-42.15) 2001-06
6 G Fisichella 98 2 3 1 2242.73 (+4.01) 2001-06
7 J Button 95 3 1 - 2177.36 (+23.94) 2001-06
8 (+1) F Massa 70 3 2 2 2075.99 (+242.54) 2002/04-06
9 (-1) R Schumacher 91 5 5 5 1884.79 (-31.35) 2001-06
10 J Trulli 97 2 1 - 1677.53 (-27.04) 2001-06

11 D Coulthard 99 1 3 4 1491.48 (-26.85) 2001-06
12 (+1) N Heidfeld 94 1 - - 1270.37 (-2.11) 2001-06
13 (-1) M Webber 86 - - - 1125.09 (-16.75) 2002-06
14 T Sato 66 - - - 865.16 (+9.52) 2002-06
15 J Villeneuve 76 - - - 799.31 (-17.82) 2001-06
16 C Klien 46 - - - 675.66 (-9.48) 2004-06
17 T Monteiro 36 - - - 632.16 (+16.05) 2005-06
18 C Albers 36 - - - 553.22 (+17.39) 2005-06
19 (+1) V Liuzzi 22 - - - 385.46 (+19.79) 2005-06
20 (+1) P de la Rosa 38 - - 1 379.82 (+21.24) 2001-02/05-06

21 (-2) O Panis 61 - - - 374.79 (-11.74) 2001-02/04
22 S Speed 18 - - - 335.82 (+20.52) 2006
23 N Karthikeyan 19 - - - 283.00 (-3.86) 2005
24 N Rosberg 18 - - 1 281.78 (-0.46) 2006
25 C da Matta 27 - - - 254.46 (-5.99) 2004
26 A Pizzonia 20 - - - 201.14 (-3.83) 2004-05
27 (+2) R Doornbos 11 - - - 195.65 (+21.87) 2005-06
28 (-1) Z Baumgartner 20 - - - 194.89 (-3.62) 2004
29 (-1) H-H Frentzen 36 - - - 173.53 (-7.19) 2001-03
30 (+1) R Kubica 5 - - - 160.10 (+22.77) 2006

31 (-1) G Bruni 18 - - - 154.77 (-2.83) 2004
32 (+1) P Friesacher 11 - - - 121.04 (-1.74) 2005
33 (-1) J Verstappen 28 - - - 120.17 (-5.58) 2001
34 G Pantano 14 - - - 109.24 (-2.09) 2004
35 J Wilson 16 - - - 93.51 (-2.45) 2003
36 (+4) S Yamamoto 7 - - - 91.29 (+23.02) 2006
37 (-1) E Irvine 28 - - - 89.07 (-6.07) 2001-02
38 (-1) R Firman 14 - - - 87.54 (-2.36) 2003
39 (-1) F Montagny 7 - - - 80.93 (-0.90) 2006
40 (-1) M Salo 17 - - - 76.72 (-3.70) 2002

41 T Glock 4 - - - 61.13 (-1.03) 2004
42 R Zonta 7 - - - 57.00 (-1.30) 2001/04-05
43 (+1) A McNish 16 - - - 50.50 (-2.45) 2002
44 (-1) M Häkkinen 11 - 2 2 49.55 (-7.25) 2001
45 N Kiesa 5 - - - 48.68 (-1.13) 2003
46 Y Ide 4 - - - 48.04 (-0.57) 2006
47 M Gené 3 - - - 45.45 (-0.92) 2004
48 A Yoong 14 - - - 42.17 (-2.35) 2001-02
49 A Wurz 1 - - - 40.66 (-0.60) 2005
50 E Bernoldi 22 - - - 36.37 (-3.09) 2001-02

51 J Alesi 12 - - - 20.18 (-3.53) 2001
52 A Davidson 3 - - - 7.89 (-0.27) 2002/05
53 (+1) T Enge 3 - - - 6.58 (-0.61) 2001
54 (-1) T Marques 8 - - - 5.99 (-1.37) 2001
55 L Burti 8 - - - 4.54 (-1.44) 2001

I'll aim to post updated rankings after each GP.

I am working on compiling the data for 1950-2001 so I can see who
was "ranked number 1", or who was a "top 10 driver" all the way
back to 1950.

--

Brian W Lawrence
Wantage,
Oxfordshire, UK
Brian_W_Lawrence@msn.com

Brian Lawrence wrote:
> SNIP!
>
> I wondered if I could "borrow" some of the features of the Wildsoft system
> and produce my own version.
>
< CHOMP>

Nice one, Brian.
I look forward to seeing these.

On Tue, 6 Mar 2007 13:09:03 -0000, "Brian Lawrence"
<Brian_W_Lawrence@msn.com> wrote:

>Two other "rules" I have adopted from
>Wildsoft are that no points are awarded UNLESS a driver actually starts a
>GP;

Fair enough.

>and that any driver who gets disqualified also fails to gain any points.

For ease of calculating, fair enough. To be accurate, though, only
disqualifications for something that altered performance should be
ignored. Perhaps that's a pedantic argument; I dunno.

>Increasing the number of GPs in the database to 100 means that data is
>used going back to the 2001 Austrian GP (13 May). That race was won by David
>Coulthard, but clearly it should have less "value" than Felipe Massa's
>Brazilian victory last October. Therefore the older races count for less
>points than more recent ones and races in the current season are more
>valuable than races last season (or older).

I don't get this. Why do you consider that there is a decreasing level
of ability from drivers and teams as time passes ?

Given that you've done the work and input the data, why remove race
results beyond 100 ? It makes the results less meaningful.

If you're going to have a cutoff, to be fair, you need to exclude or
mark drivers who have had results excluded.

This highlights the problem for me -

> Driver GPs PP W FL Points (+/-) Seasons
>================================================= ==================

>10 J Trulli 97 2 1 - 1677.53 (-27.04) 2001-06

>44 (-1) M Häkkinen 11 - 2 2 49.55 (-7.25) 2001

So Trulli is 34 times better than Häkkinen ?

<Sissons>Yeeessssssssssss.

Daytona

"Daytona" <me@privacy.net> wrote:

>>Increasing the number of GPs in the database to 100 means that data is
>>used going back to the 2001 Austrian GP (13 May). That race was won by David
>>Coulthard, but clearly it should have less "value" than Felipe Massa's
>>Brazilian victory last October. Therefore the older races count for less
>>points than more recent ones and races in the current season are more
>>valuable than races last season (or older).
>
> I don't get this. Why do you consider that there is a decreasing level
> of ability from drivers and teams as time passes ?

I don't. Driver (and team) ability is variable, some improve, some deteriorate,
some stay the same. Rather, what I'm ranking is individual race performance
over a finite period (100 races). In that context, as I said in the OP,
Coulthard's win in race #1 doesn't have as much value as Massa's in race #100.
The fact that DC won that race five years ago has less relevance to a ranking
of drivers in March 2007 than Felipe winning in Brazil does.

> Given that you've done the work and input the data, why remove race
> results beyond 100 ? It makes the results less meaningful.

As I also said in the OP, I plan to add all of the data back to 1950, so I
will be able to post all-time rankings, although those would not be weighted
in favour of more recent performances.

> If you're going to have a cutoff, to be fair, you need to exclude or
> mark drivers who have had results excluded.

I could certainly mark drivers whose careers began before the cutoff. I may
well do so.

> This highlights the problem for me -
>
>
>> Driver GPs PP W FL Points (+/-) Seasons
>>================================================ ===================
>
>>10 J Trulli 97 2 1 - 1677.53 (-27.04) 2001-06
>
>>44 (-1) M Häkkinen 11 - 2 2 49.55 (-7.25) 2001
>
> So Trulli is 34 times better than Häkkinen ?

No, not at all.

However, over the last 100 GPs, Trulli is ranked #10 (from 97 of 100
races) while Häkkinen is ranked #44 (from 11). After Australia Trulli
will probably have 98 of 100, but Häkkinen will definitely have only 10.
After 11 races of the season Häkkinen will no longer be in the list.

If it was a measure of all-time F1 ability Mika would be many places
above Trulli. But it's not that, it's a ranking of performance over
the last 100 GPs. No more, no less.

--

Brian

Brian Lawrence wrote:
> "Daytona" <me@privacy.net> wrote:
>
>>> Increasing the number of GPs in the database to 100 means that data
>>> is used going back to the 2001 Austrian GP (13 May). That race was won
>>> by David Coulthard, but clearly it should have less "value" than
>>> Felipe Massa's Brazilian victory last October. Therefore the older
>>> races count for less points than more recent ones and races in the
>>> current season are more valuable than races last season (or older).
>>
>> I don't get this. Why do you consider that there is a decreasing
>> level of ability from drivers and teams as time passes ?
>
> I don't. Driver (and team) ability is variable, some improve, some
> deteriorate, some stay the same. Rather, what I'm ranking is
> individual race performance over a finite period (100 races). In that
> context, as I said in the OP, Coulthard's win in race #1 doesn't have
> as much value as Massa's in race #100. The fact that DC won that race
> five years ago has less relevance to a ranking of drivers in March
> 2007 than Felipe winning in Brazil does.
>> Given that you've done the work and input the data, why remove race
>> results beyond 100 ? It makes the results less meaningful.
>
> As I also said in the OP, I plan to add all of the data back to 1950,
> so I will be able to post all-time rankings, although those would not
> be weighted in favour of more recent performances.
>
>> If you're going to have a cutoff, to be fair, you need to exclude or
>> mark drivers who have had results excluded.
>
> I could certainly mark drivers whose careers began before the cutoff.
> I may well do so.
>
>> This highlights the problem for me -
>>
>>
>>> Driver GPs PP W FL Points (+/-) Seasons
>>> ================================================== =================
>>
>>> 10 J Trulli 97 2 1 - 1677.53 (-27.04) 2001-06
>>
>>> 44 (-1) M Häkkinen 11 - 2 2 49.55 (-7.25) 2001
>>
>> So Trulli is 34 times better than Häkkinen ?
>
> No, not at all.
>
> However, over the last 100 GPs, Trulli is ranked #10 (from 97 of 100
> races) while Häkkinen is ranked #44 (from 11). After Australia Trulli
> will probably have 98 of 100, but Häkkinen will definitely have only
> 10. After 11 races of the season Häkkinen will no longer be in the list.
>
> If it was a measure of all-time F1 ability Mika would be many places
> above Trulli. But it's not that, it's a ranking of performance over
> the last 100 GPs. No more, no less.

But 100 GPs is much too long a period to be a guide to current performance -
and much too short to be a guide to all-time performance.
--
John Briggs

"Brian Lawrence" <Brian_W_Lawrence@msn.com> wrote in message
news:5557fjF22h7t0U1@mid.individual.net...
> Last week tennis fans were aware that Switzerland's Roger Federer
> had broken the record for most consecutive weeks ranked number one
> on the ATP computer.
>
> http://www.atptennis.com/5/en/rankings/entrysystem/
>
> Federer has now been ranked number one in the world for 162 weeks.
> He bettered the previous record set by Jimmy Connors in 1974-77.
>
> Women's tennis has a similar ranking system.
>
> http://www.sonyericssonwtatour.com/2...es_numeric.asp
>
> Several other sports operate ranking systems along similar lines. For
> example:
>
> Golf http://www.pgatour.com/r/stats/current/186.html
> Soccer
> http://www.fifa.com/en/mens/statisti...b-2007,00.html
> Cricket http://www.icc-cricket.com/icc/rankings/lg.html
> Snooker http://www.snooker.org/rnk/
>
> Formula 1 doesn't have (or need) a ranking system, but it might be
> interesting
> to create one.
>
> Actually there is such a system available on the FORIX website - it's
> called
> the Wildsoft F1 Ranking - but, you have to subscribe to FORIX to be able
> to
> see it.
>

excellent idea I look forward to seeing the full ranking. It would be great
if you make a website out of it as well.

	#1
Brian Lawrence	Ranking F1 Drivers Last week tennis fans were aware that Switzerland's Roger Federer had broken the record for most consecutive weeks ranked number one on the ATP computer. http://www.atptennis.com/5/en/rankings/entrysystem/ Federer has now been ranked number one in the world for 162 weeks. He bettered the previous record set by Jimmy Connors in 1974-77. Women's tennis has a similar ranking system. http://www.sonyericssonwtatour.com/2...es_numeric.asp Several other sports operate ranking systems along similar lines. For example: Golf http://www.pgatour.com/r/stats/current/186.html Soccer http://www.fifa.com/en/mens/statisti...b-2007,00.html Cricket http://www.icc-cricket.com/icc/rankings/lg.html Snooker http://www.snooker.org/rnk/ Formula 1 doesn't have (or need) a ranking system, but it might be interesting to create one. Actually there is such a system available on the FORIX website - it's called the Wildsoft F1 Ranking - but, you have to subscribe to FORIX to be able to see it. http://forix.autosport.com/stats.php?l=0&c=1205 The Wildsoft system was devised by Dimitry Cherepovich, and is calculated over the last five years of F1. Only data from GPs within the last five years are included, so the 2002 Australian GP has just been "dropped" (actually, the rankings haven't been updated since December, but I'm sure it will be soon). Currently the system has 88 GPs in its database. Points are awarded for the usual F1 achievements - wins, pole positions, fastest laps, etc., but also include other criteria such as distance raced. I wondered if I could "borrow" some of the features of the Wildsoft system and produce my own version. The first change I decided on was to drop the five year period and switch to the last 100 GPs instead. This is easier to manage mathematically, and I always prefer things to be simple. Two other "rules" I have adopted from Wildsoft are that no points are awarded UNLESS a driver actually starts a GP; and that any driver who gets disqualified also fails to gain any points. Increasing the number of GPs in the database to 100 means that data is used going back to the 2001 Austrian GP (13 May). That race was won by David Coulthard, but clearly it should have less "value" than Felipe Massa's Brazilian victory last October. Therefore the older races count for less points than more recent ones and races in the current season are more valuable than races last season (or older). Thus my rankings are based on performances over the last five years 10 months, but if the number of races per season continues at its present level the window will shrink. It should be obvious that Michael Schumacher will be ranked as number one, but my first question to ponder is, how long before Fernando Alonso (or someone else) overtakes Michael. The 2007 drivers will all be adding to their points totals as each race of 2007 is completed, but MS (and other retired drivers) will steadily lose points as the oldest GPs are removed. OK, preamble over, here are my current rankings. GPs - number of GPs counted for each driver (from the last 100) PP - number of Pole Positions (in the last 100) W - number of wins (in the last 100) FL - number of Fastest Laps (in the last 100) Points - these are my points, allocated for various things, not just the wins, poles, f.laps (+/-) - the change in points as a result of the most recent race Seasons - the seasons in which the driver competed (last 100 GPs only) Numbers in brackets () in second column indicate the change in position as a result of adding the most recent GP. First race (GP #1) - Austrian GP, 13 May 2001 Last race (GP #100) - Brazilian GP, 22 Oct 2006 Driver GPs PP W FL Points (+/-) Seasons ================================================== ================= 1 M Schumacher 100 32 44 33 6505.19 (-17.71) 2001-06 2 F Alonso 81 15 15 8 5055.20 (+54.67) 2001/03-06 3 K Räikkönen 98 11 9 19 4035.21 (+62.04) 2001-06 4 R Barrichello 98 10 8 12 2958.73 (-28.39) 2001-06 5 JP Montoya 86 13 7 12 2647.00 (-42.15) 2001-06 6 G Fisichella 98 2 3 1 2242.73 (+4.01) 2001-06 7 J Button 95 3 1 - 2177.36 (+23.94) 2001-06 8 (+1) F Massa 70 3 2 2 2075.99 (+242.54) 2002/04-06 9 (-1) R Schumacher 91 5 5 5 1884.79 (-31.35) 2001-06 10 J Trulli 97 2 1 - 1677.53 (-27.04) 2001-06 11 D Coulthard 99 1 3 4 1491.48 (-26.85) 2001-06 12 (+1) N Heidfeld 94 1 - - 1270.37 (-2.11) 2001-06 13 (-1) M Webber 86 - - - 1125.09 (-16.75) 2002-06 14 T Sato 66 - - - 865.16 (+9.52) 2002-06 15 J Villeneuve 76 - - - 799.31 (-17.82) 2001-06 16 C Klien 46 - - - 675.66 (-9.48) 2004-06 17 T Monteiro 36 - - - 632.16 (+16.05) 2005-06 18 C Albers 36 - - - 553.22 (+17.39) 2005-06 19 (+1) V Liuzzi 22 - - - 385.46 (+19.79) 2005-06 20 (+1) P de la Rosa 38 - - 1 379.82 (+21.24) 2001-02/05-06 21 (-2) O Panis 61 - - - 374.79 (-11.74) 2001-02/04 22 S Speed 18 - - - 335.82 (+20.52) 2006 23 N Karthikeyan 19 - - - 283.00 (-3.86) 2005 24 N Rosberg 18 - - 1 281.78 (-0.46) 2006 25 C da Matta 27 - - - 254.46 (-5.99) 2004 26 A Pizzonia 20 - - - 201.14 (-3.83) 2004-05 27 (+2) R Doornbos 11 - - - 195.65 (+21.87) 2005-06 28 (-1) Z Baumgartner 20 - - - 194.89 (-3.62) 2004 29 (-1) H-H Frentzen 36 - - - 173.53 (-7.19) 2001-03 30 (+1) R Kubica 5 - - - 160.10 (+22.77) 2006 31 (-1) G Bruni 18 - - - 154.77 (-2.83) 2004 32 (+1) P Friesacher 11 - - - 121.04 (-1.74) 2005 33 (-1) J Verstappen 28 - - - 120.17 (-5.58) 2001 34 G Pantano 14 - - - 109.24 (-2.09) 2004 35 J Wilson 16 - - - 93.51 (-2.45) 2003 36 (+4) S Yamamoto 7 - - - 91.29 (+23.02) 2006 37 (-1) E Irvine 28 - - - 89.07 (-6.07) 2001-02 38 (-1) R Firman 14 - - - 87.54 (-2.36) 2003 39 (-1) F Montagny 7 - - - 80.93 (-0.90) 2006 40 (-1) M Salo 17 - - - 76.72 (-3.70) 2002 41 T Glock 4 - - - 61.13 (-1.03) 2004 42 R Zonta 7 - - - 57.00 (-1.30) 2001/04-05 43 (+1) A McNish 16 - - - 50.50 (-2.45) 2002 44 (-1) M Häkkinen 11 - 2 2 49.55 (-7.25) 2001 45 N Kiesa 5 - - - 48.68 (-1.13) 2003 46 Y Ide 4 - - - 48.04 (-0.57) 2006 47 M Gené 3 - - - 45.45 (-0.92) 2004 48 A Yoong 14 - - - 42.17 (-2.35) 2001-02 49 A Wurz 1 - - - 40.66 (-0.60) 2005 50 E Bernoldi 22 - - - 36.37 (-3.09) 2001-02 51 J Alesi 12 - - - 20.18 (-3.53) 2001 52 A Davidson 3 - - - 7.89 (-0.27) 2002/05 53 (+1) T Enge 3 - - - 6.58 (-0.61) 2001 54 (-1) T Marques 8 - - - 5.99 (-1.37) 2001 55 L Burti 8 - - - 4.54 (-1.44) 2001 I'll aim to post updated rankings after each GP. I am working on compiling the data for 1950-2001 so I can see who was "ranked number 1", or who was a "top 10 driver" all the way back to 1950. -- Brian W Lawrence Wantage, Oxfordshire, UK Brian_W_Lawrence@msn.com

	#2
Knight who says \NI!\	Re: Ranking F1 Drivers Brian Lawrence wrote: > SNIP! > > I wondered if I could "borrow" some of the features of the Wildsoft system > and produce my own version. > < CHOMP> Nice one, Brian. I look forward to seeing these.

	#3
Daytona	Re: Ranking F1 Drivers On Tue, 6 Mar 2007 13:09:03 -0000, "Brian Lawrence" <Brian_W_Lawrence@msn.com> wrote: >Two other "rules" I have adopted from >Wildsoft are that no points are awarded UNLESS a driver actually starts a >GP; Fair enough. >and that any driver who gets disqualified also fails to gain any points. For ease of calculating, fair enough. To be accurate, though, only disqualifications for something that altered performance should be ignored. Perhaps that's a pedantic argument; I dunno. >Increasing the number of GPs in the database to 100 means that data is >used going back to the 2001 Austrian GP (13 May). That race was won by David >Coulthard, but clearly it should have less "value" than Felipe Massa's >Brazilian victory last October. Therefore the older races count for less >points than more recent ones and races in the current season are more >valuable than races last season (or older). I don't get this. Why do you consider that there is a decreasing level of ability from drivers and teams as time passes ? Given that you've done the work and input the data, why remove race results beyond 100 ? It makes the results less meaningful. If you're going to have a cutoff, to be fair, you need to exclude or mark drivers who have had results excluded. This highlights the problem for me - > Driver GPs PP W FL Points (+/-) Seasons >================================================= ================== >10 J Trulli 97 2 1 - 1677.53 (-27.04) 2001-06 >44 (-1) M Häkkinen 11 - 2 2 49.55 (-7.25) 2001 So Trulli is 34 times better than Häkkinen ? <Sissons>Yeeessssssssssss. Daytona

	#4
Brian Lawrence	Re: Ranking F1 Drivers "Daytona" <me@privacy.net> wrote: >>Increasing the number of GPs in the database to 100 means that data is >>used going back to the 2001 Austrian GP (13 May). That race was won by David >>Coulthard, but clearly it should have less "value" than Felipe Massa's >>Brazilian victory last October. Therefore the older races count for less >>points than more recent ones and races in the current season are more >>valuable than races last season (or older). > > I don't get this. Why do you consider that there is a decreasing level > of ability from drivers and teams as time passes ? I don't. Driver (and team) ability is variable, some improve, some deteriorate, some stay the same. Rather, what I'm ranking is individual race performance over a finite period (100 races). In that context, as I said in the OP, Coulthard's win in race #1 doesn't have as much value as Massa's in race #100. The fact that DC won that race five years ago has less relevance to a ranking of drivers in March 2007 than Felipe winning in Brazil does. > Given that you've done the work and input the data, why remove race > results beyond 100 ? It makes the results less meaningful. As I also said in the OP, I plan to add all of the data back to 1950, so I will be able to post all-time rankings, although those would not be weighted in favour of more recent performances. > If you're going to have a cutoff, to be fair, you need to exclude or > mark drivers who have had results excluded. I could certainly mark drivers whose careers began before the cutoff. I may well do so. > This highlights the problem for me - > > >> Driver GPs PP W FL Points (+/-) Seasons >>================================================ =================== > >>10 J Trulli 97 2 1 - 1677.53 (-27.04) 2001-06 > >>44 (-1) M Häkkinen 11 - 2 2 49.55 (-7.25) 2001 > > So Trulli is 34 times better than Häkkinen ? No, not at all. However, over the last 100 GPs, Trulli is ranked #10 (from 97 of 100 races) while Häkkinen is ranked #44 (from 11). After Australia Trulli will probably have 98 of 100, but Häkkinen will definitely have only 10. After 11 races of the season Häkkinen will no longer be in the list. If it was a measure of all-time F1 ability Mika would be many places above Trulli. But it's not that, it's a ranking of performance over the last 100 GPs. No more, no less. -- Brian

	#5
John Briggs	Re: Ranking F1 Drivers Brian Lawrence wrote: > "Daytona" <me@privacy.net> wrote: > >>> Increasing the number of GPs in the database to 100 means that data >>> is used going back to the 2001 Austrian GP (13 May). That race was won >>> by David Coulthard, but clearly it should have less "value" than >>> Felipe Massa's Brazilian victory last October. Therefore the older >>> races count for less points than more recent ones and races in the >>> current season are more valuable than races last season (or older). >> >> I don't get this. Why do you consider that there is a decreasing >> level of ability from drivers and teams as time passes ? > > I don't. Driver (and team) ability is variable, some improve, some > deteriorate, some stay the same. Rather, what I'm ranking is > individual race performance over a finite period (100 races). In that > context, as I said in the OP, Coulthard's win in race #1 doesn't have > as much value as Massa's in race #100. The fact that DC won that race > five years ago has less relevance to a ranking of drivers in March > 2007 than Felipe winning in Brazil does. >> Given that you've done the work and input the data, why remove race >> results beyond 100 ? It makes the results less meaningful. > > As I also said in the OP, I plan to add all of the data back to 1950, > so I will be able to post all-time rankings, although those would not > be weighted in favour of more recent performances. > >> If you're going to have a cutoff, to be fair, you need to exclude or >> mark drivers who have had results excluded. > > I could certainly mark drivers whose careers began before the cutoff. > I may well do so. > >> This highlights the problem for me - >> >> >>> Driver GPs PP W FL Points (+/-) Seasons >>> ================================================== ================= >> >>> 10 J Trulli 97 2 1 - 1677.53 (-27.04) 2001-06 >> >>> 44 (-1) M Häkkinen 11 - 2 2 49.55 (-7.25) 2001 >> >> So Trulli is 34 times better than Häkkinen ? > > No, not at all. > > However, over the last 100 GPs, Trulli is ranked #10 (from 97 of 100 > races) while Häkkinen is ranked #44 (from 11). After Australia Trulli > will probably have 98 of 100, but Häkkinen will definitely have only > 10. After 11 races of the season Häkkinen will no longer be in the list. > > If it was a measure of all-time F1 ability Mika would be many places > above Trulli. But it's not that, it's a ranking of performance over > the last 100 GPs. No more, no less. But 100 GPs is much too long a period to be a guide to current performance - and much too short to be a guide to all-time performance. -- John Briggs

	#6
Bryan	Re: Ranking F1 Drivers "Brian Lawrence" <Brian_W_Lawrence@msn.com> wrote in message news:5557fjF22h7t0U1@mid.individual.net... > Last week tennis fans were aware that Switzerland's Roger Federer > had broken the record for most consecutive weeks ranked number one > on the ATP computer. > > http://www.atptennis.com/5/en/rankings/entrysystem/ > > Federer has now been ranked number one in the world for 162 weeks. > He bettered the previous record set by Jimmy Connors in 1974-77. > > Women's tennis has a similar ranking system. > > http://www.sonyericssonwtatour.com/2...es_numeric.asp > > Several other sports operate ranking systems along similar lines. For > example: > > Golf http://www.pgatour.com/r/stats/current/186.html > Soccer > http://www.fifa.com/en/mens/statisti...b-2007,00.html > Cricket http://www.icc-cricket.com/icc/rankings/lg.html > Snooker http://www.snooker.org/rnk/ > > Formula 1 doesn't have (or need) a ranking system, but it might be > interesting > to create one. > > Actually there is such a system available on the FORIX website - it's > called > the Wildsoft F1 Ranking - but, you have to subscribe to FORIX to be able > to > see it. > excellent idea I look forward to seeing the full ranking. It would be great if you make a website out of it as well.