Is BURGER SEEDING INDEX CALCULATED CORRECTLY

[orbeaorca]

My friend asked me if the Burger index was calculated correctly cause despite Improving on his engen time by 22 min his index only improved by less than 5. He has trained hard and completed his first DC.

 

                        WINNER                    HIS TIME                    INDEX

ENGEN                227                        257                             60.08

 

BURGER               223                        235                              64.8

 

It does not seem likely that it could be from the winner being 4min faster.

 

They must have adjusted the beta factor they use in a big way.

 

If i understand it they would compare your index you got in the burger to your base event then if you did too well then they would assume that the burger was to easy and penalize you by changing the beta factor so your index would increase.

 

I have a couple of questions.

 

1. Why would the beta be so different between two rides on the same course?

The faster winning time from the better field is a penalty in itself.

Is the field being punished because they are fitter at the burger

 

2. I assume they would have used the Argus 2009 as the base event.

 

     How can you use this as a base event considering the nature of the day when everyone was off the mark  and there will be a difference in indexes?

 

Lots people trained specifically for the Burger considering they did so poorly in the argus.

 

SO did you get penalized twice by using the argus?

 

3. would the beta factor have been different if you used last years burger was the base event? probably

 

4. Is this a question of garbage in garbage out?

 

5. Is this new model they are using far less objective and more subjective by trying to factor in for to many things.

 

 

 

 

 

 

 

 

 

 

[orbeaorca]

bump

[Bram_on_Madone]

Similar thing happened to me - I rode about the same time, but my index deteriorated substantially.  I asked PPA about it the other day and apparently its got a lot to do with how many people rode the race that also rode the Argus, as they use the top hundred previous Argus riders' times as some sort of base for the calculation.  So in short, although its not subjective, it certainly depends on a number of random factors that is unrelated to your individual performance. Feltrider2009-12-10 06:51:29

[Woofie]

Something is very wrong there

 

Look at my recent seeding results.

 

If it is vastly different, then you will need to take it up with PPA

 

 

Event Date	Event Name	Distance	Winner Time	Race Time	Ave Speed	Weighting	Ride Index
2009/10/31	Durbanville High School Funride & MTB	94.3	02:12:57	02:36:32	36.15	34	21.97703
2009/10/18	PPA One Tonner	158	03:54:24	05:07:12	30.86	34	21.78751
2009/10/25	Tiletoria West Coast Express League and Funride	80	01:56:12	02:19:42	34.36	32	24.30061
2009/11/29	Die Burger Fietstoer	97	02:22:57	02:53:35	33.53	0	25.63918
2009/09/20	Orbit Tour d' Worcester	105	02:24:49	02:49:11	37.24	0	28.65403
2009/07/05	Pick n Pay/Weekend Argus/Rotary Knysna Cycle Tour	100	02:33:44	03:28:09	28.83	0	30.87005
2009/04/04	Geard Pharmacy League Funride and MTB	78	01:50:37	02:05:19	37.35	0	26.2

 

[FanieFiets]

it is because it is a fairly flat race.  Thus the groups time will greatly influenced by the couple of stongest guys working.  The engen had small groups, thus greater variation between groups thus seeding could be done 'better'.  Burger, MASSIVE groups, thus weaker riders rode with much stronger riders for very long and then after the bump just latched on the next group.  Weak riders got relatively good times, thus time will be adjusted much more.

 

Makes sense?

 

[orbeaorca]

it is because it is a fairly flat race.  Thus the groups time will greatly influenced by the couple of stongest guys working.  The engen had small groups' date=' thus greater variation between groups thus seeding could be done 'better'.  Burger, MASSIVE groups, thus weaker riders rode with much stronger riders for very long and then after the bump just latched on the next group.  Weak riders got relatively good times, thus time will be adjusted much more.

Makes sense?
[/quote']

 

I think you missing the point. The real question is can you use Argus 2009 as the base event then adjust times accordingly because everyone did badly in it. We from the cape did not do the 94.7, so I indexes would be adjusted more negatively.

 

I think it would be more fair to use the Burger of last year as a base event and you would get a different index.

 

There are a lot of people who now will struggle to get into a decent group in the Argus.

 

Go have a look at the engen results and you will also how mixed up the results were with group g&f giving a-d a run for their money thus a few strong riders also had a marked impact.

 

@Woofie your results that this does not make sense   

[PeterM]

I have looked at the way the PPA calculates the seeding and i have spoken to the guy who does it.

It seems that the new approach is better and "cleaner". It takes the concept of % slower than the winner (or best possible winner). So if your index is 30 than you are 30% slower than the best possible winner, plus it makes provision for the nature of the race plus who attends. This is done mainly via a straight calculation plus an added linear regression.

 

But i really think that the way it is done in reality, is not consistent at all.

Looking at my own results and the general results within each group seems to confirms that.

Now after a few races, little meaningful adjustment / movement within the race starting groups have happened. Now we still have start groups, where the first 1/4 are faster than those 3-5 start groups in front of them, etc.

 

I while ago i was very unhappy about this. Now i just "let-it-be" and rather enjoy the cycling & racing and focus towards the Argus.

I think the PPA / Racetec need some more time and more confirming data to adjust their methodology and than start to listen LESS to their cycling advisors regarding the type and qty of linear regression they should use. And just stick to a consistent and confirmed mathematical approach.

 

It's not perfect but i don't want to let it all spoil my time on the bike!

 

PS @ Woofie - great race times!

 

[orbeaorca]

Hi PETER M wont you ask this guy if they are using Argus 2009 as the base event for all the seeding events like 94.7 etc. If he is doing this then it will be fair cause everyone will get a messed up index to seed them in the Argus 2010.

 

However if we in the Cape get the Burger compared with Argus 09 and the 94.7(2009) gets compared to 94.7 (2008) then that is unfair.

 

The obvious answer you will get is that the burger was too fast therefore increase everyones index and the  94.7 was consistent with the previous

year therefore index will remain.

 

[Bram_on_Madone]

I guess no system will ever be fair in the sense that everybody is happy.  As long as they are trying to improve the system and you can still climb the seeding with hard work we should be satisfied....

[niklasvm]

For the argus seeding do they only use the previous year's results or older ones as well? I've only done this year's argus with a bad time but then did reasonably well in Die Burger. I got an index of 23.4 for the burger. Last year, people who got that index were seeded quite well for the Argus (A - H). And i'm guessing that everyone wants to be there so they can aim for the sub 3 so I'm hoping I make the cut.

 

 

 

How can an accurate seeding index actually be calculated if not everyone did all the races in the alphapharm series. A standard linear regression would need everyone to have done every race. Alternatively maybe they just rank everyone based on their past ability and then sub-divide them into groups. It would be interesting to see how PPA groups A - J are seeded in the argus. As far as i know the rest of the PPA members are just seeded then in the P groups

[niklasvm]

They must have adjusted the beta factor they use in a big way.

 

 

 

The way that the beta is calculated is completely mathematical. Its the linear regression beta coefficient based on the top group of riders that did both the Argus and the race in question. It assumes that the riders should perform in the same way. The only thing i worry about is that the spread (variance) of the times changes for every race depending on its course. Also some riders do worse and others get better. I'm note sure how this would affect things either.

[Jakes_1]

 

 

My friend asked me if the Burger index was calculated correctly cause despite Improving on his engen time by 22 min his index only improved by less than 5. He has trained hard and completed his first DC.

 

                        WINNER                    HIS TIME                    INDEX

ENGEN                227                        257                             60.08

 

BURGER               223                        235                              64.8

 

 

 

There is something realy wrong with that index, have a look a mine.

Event Date Event Name Distance Winner Time Race Time Ave Speed Weighting Ride Index 2009/11/29 Die Burger Fietstoer 97 02:22:57 02:53:30 33.54 34 25.58333 2009/11/08 Medscheme Tour de Vino 117 02:52:28 03:53:59 30 34 28.95792 2009/10/10 Engen Dynamic Cycle Challenge 99 02:26:52 03:19:15 29.81 32 37.61041 2009/10/25 Tiletoria West Coast Express League and Funride 80 01:56:12 02:34:44 31.02 0 39.84927 2009/08/01 Die Burger Bergfietsuitdaging (voorheen Stellenbosch Bergfietsuitdaging) 30 01:07:05 01:36:48 18.6 0 147.7609

Jakes_12009-12-15 01:25:27