Race Seeding

[Dgas]

Does anybody where the seeding in races comes from? ASG seem to do their own seeding, but where do the other organizers get a seeding from? For example, where is the seeding for the Fast One come from? Is there anywhere that you can see your seeding?

 

My seeding on ASG is now 9.33 which I guess should get me a A group seeding in ASG races but is this shared with the others? For example I got a good seeding in the Sondela Classic (enough has been said about that race) but this seeding is only on the ASG site, it has not been put on Racetec, there are other results there so that should get me a seeding.

 

The guys in the Cape have no idea how lucky they are to have the PPA. I lived and raced there for 5 years, what a pleasure compared to the mess that is Gauteng.

[Patchelicious]

ASG said that they will be sharing the results with Racetec from Sondela if I am not mistaken.

 

Most races would get their seedlings from your PPA seeding index?

[Dgas]

ASG said that they will be sharing the results with Racetec from Sondela if I am not mistaken.

 

Most races would get their seedlings from your PPA seeding index?

 

The Sondela Race is not on Racetec yet.

 

Do you know if PPA only redo their seeding every few months? On the ASG page they have all my races (ASG events only) up to date, the latest being Sondela, and have given me a seeding of 9.33. PPA have my seeding at 15.28 but only go up to the Jock.

 

My ASG seeding for the Jock is 9.95 and my PPA seeding for the same event is 15.28. How is that possible, I am not arguing about the seeding just about the inconsistency.

[Baaisikilist]

Figuring out the dark art of seeding is very much like figuring out women...

[DJR]

.................The guys in the Cape have no idea how lucky they are to have the PPA..................

..............some of us do get it, but others just moan and bitch!

[ASG]

Hi Everyone,

 

As there have been a number of questions regarding seeding recently, we thought we would share some ideas, information and facts, to help clear up the “dark art of seeding” (as someone appropriately referred to it! J).

 

As you all know, seeding is a useful tool to assist race organisers place riders in the most appropriate starting batch of a race, thereby reducing (not eliminating) the chances of congestion (eg, faster riders getting stuck behind slower, or less technically able riders), and/or ensuring the riders are placed in a bunch of similar paced competitors (especially in a road event, where bunch-riding is quite critical). A seeding “index” is calculated based on previous results of all competitors, and is an arbitrary value (usually, but not always, a percentage) designed to be comparable between other athletes that have been seeded using the same system.

 

There are a number of seeding “engines” in South Africa currently. Combining them all into one national system would have merits (and ASG would support such a move), but that is a debate for another day (and is fraught with politics and complications). Most of the systems have an excellent history of results, and therefore are robust and produce consistent seeding indices. However, although they are all aimed at producing a relative index per rider based on historical performance (and thus, by inference, current rider ability), they all apply different rules and algorithms. For example, rules (or combinations thereof) are applied that weight results based on the age of the result (how long ago it took place – obviously the older it is, the less one can infer CURRENT ability therefrom), distance of the race (where a marathon counts a bit more when comparing relative performance to riders who did a half-marathon), difficulty/grading of the event (which is extremely complex, and even subjective, to quantify obviously), number of riders competing (which is an often-debated topic), etc. The fact that different rules are applied, and approaches may differ, means that no two seeding indices will be the same. An index of 10.5, say, in one system may well be very comparable to an index of 19.7 in another. What is important is whether the relative seeding position of one rider versus another, using the SAME system, are comparable and consistent. So, if one rider is “top 25%” in one system, they should ROUGHLY be “top 25%” using another system. Comparing the actual indices makes no sense, and should be avoided. In fact, in some systems a low index is “good”, and a high index is “bad”, whereas in others the converse is true. So, a difference in indices between systems is not necessarily “inconsistent” (as one Hubber asked), so long as over time your relative position/seeding in the “bunch” (or amongst your fellow riders) remains the same over time using the same seeding system (but, again, this assumes your performance levels, and of those competing with you, remain pretty consistent over time, which is not always the case.). Furthermore, there is not a single place one can go to see all your seeding indices, or compare across seeding systems (see previous comments regarding (i) the dangers of doing this comparison, and (ii) the scope for an integrated, national seeding system…). 

 

Needless to say, the more historic results you have for a rider, the more “robust” the seeding calculations become. Similarly, not having any results for a rider unfortunately means one cannot seed them, until such time as you have built up some history of their performance. ASG took the (difficult) decision to develop our own seeding engine, for a number of reasons including (but not limited to) (i) the rich set of results data we had for ASG events, (ii) the belief that we could apply more considered “rules” to the datasets, which should result in improved seeding, (iii) the costs associated with buying seeding information for each and every event (with such costs impacting on the entry fee for riders), and (iv) the ability for us to handle seeding queries and changes in a faster, more efficient manner. We have had a few teething problems (almost exclusively related to not having sufficient historical results, particularly on the road side, as we previously relied on our timing and seeding partners to store such results), but the “coverage” of our results has increased dramatically in recent months, enabling us to improve our seeding as each event takes place. 

 

We take all the feedback we have received regarding seeding very seriously, and are looking at new, innovative ways we can make our seeding more robust, and even allow individual riders to interact with the engine, and manage their own data a bit closer. Again, this ability to be flexible and apply a more innovative approach is one of the reasons we took the decision to implement our own seeding system.

 

In conclusion then, we are sure most of you will agree that seeding is NOT an exact science, and no two systems will “agree” with one another. However, by being diligent in the collection and storing of historical results data, and the application of simple but meaningful rules and weightings to the results, a system will be produced that allows organisers to batch riders appropriately, leading to a more rewarding and enjoyable experience for riders during a race. We at ASG are certainly committed to continue investing in such a robust seeding system for our events.

 

ASG Events