Forum: Race Officers
Cox-Sprague Scoring System
Peter Clapp
Nationality: United States
Certifications:
- Race Officer In Training
- Judge In Training
0
Our RC model racing club is planning to implement the Cox-Sprague Scoring System for our seasonal race series. I am hoping that this group of experts can help shed some light on a couple of details concerning 'after calculation' adjustments.
For those unfamiliar with RRS Appendix E and how RC racing is conducted, our club runs weekly races which typically consist of 8-12 heats (with a minimum of 4 heats to constitute a race) and then a minimum of 3 but no more than 12 races to constitute a series. As you can see there is a lot of scoring to do. Our races are scored using RRS Appendix A (Low-Point Scoring System) with all of its criteria for drops and tiebreakers changed only in the SI to afford more drops at 5 heat intervals.
We intend to convert our series scoring to the Cox-Sprague system to take advantage of its benefits. Those being, its ability to reward winners in a large fleet on race day but not penalize competitors that struggle to meet the 50% participation threshold by too much.
The table of scores based on the number of boats racing on any particular day is pretty straight forward. The ambiguities that arise are with regard to drops and tiebreakers after computing all the scores. The sources of information about this system of scoring are silent on the topic of dropping races. It is only in the scoring software that we find a default to a dropped race(s) scheme that resembles the Appendix A declaration with an accommodation for additional drops at subsequent thresholds. As far as tiebreakers are concerned, two of the sources describe two different criteria for breaking ties. One has only a 'sail off' as its tiebreaker and the other describes a two tiered process using 'who beat who more' as its #1 and then the ranking of # of 1st, # of 2nd etc.. We are all familiar with this latter method. The scoring software, I believe, uses some iteration of the two tier method but does not describe it specifically in its documentation.
Questions we have:
1. Is there a definitive location for the definitions of, criteria used and execution process for the Cox-Sprague Scoring System?
2. Has anyone had a good/bad experience using this method of scoring?
3. Is dropping races a redundancy because of the inherent benefits of this scoring system?
4. In the extremely unlikely circumstance of having a tie to break, is it necessary to have more than one tiebreaker?
5. Are we just better off defining all these issues in our SI to suit our own needs?
Thanks for any feedback.
Resources:
https://www.racescore.org/home#h.lw7sb5phqec4
https://www.ussailing.org/wp-content/uploads/2018/01/Cox-Sprague_Scoring_Sys.pdf
https://www.sailingscuttlebutt.com/wp-content/uploads/2018/12/Cox-Sprague-Scoring-System.pdf
For those unfamiliar with RRS Appendix E and how RC racing is conducted, our club runs weekly races which typically consist of 8-12 heats (with a minimum of 4 heats to constitute a race) and then a minimum of 3 but no more than 12 races to constitute a series. As you can see there is a lot of scoring to do. Our races are scored using RRS Appendix A (Low-Point Scoring System) with all of its criteria for drops and tiebreakers changed only in the SI to afford more drops at 5 heat intervals.
We intend to convert our series scoring to the Cox-Sprague system to take advantage of its benefits. Those being, its ability to reward winners in a large fleet on race day but not penalize competitors that struggle to meet the 50% participation threshold by too much.
The table of scores based on the number of boats racing on any particular day is pretty straight forward. The ambiguities that arise are with regard to drops and tiebreakers after computing all the scores. The sources of information about this system of scoring are silent on the topic of dropping races. It is only in the scoring software that we find a default to a dropped race(s) scheme that resembles the Appendix A declaration with an accommodation for additional drops at subsequent thresholds. As far as tiebreakers are concerned, two of the sources describe two different criteria for breaking ties. One has only a 'sail off' as its tiebreaker and the other describes a two tiered process using 'who beat who more' as its #1 and then the ranking of # of 1st, # of 2nd etc.. We are all familiar with this latter method. The scoring software, I believe, uses some iteration of the two tier method but does not describe it specifically in its documentation.
Questions we have:
1. Is there a definitive location for the definitions of, criteria used and execution process for the Cox-Sprague Scoring System?
2. Has anyone had a good/bad experience using this method of scoring?
3. Is dropping races a redundancy because of the inherent benefits of this scoring system?
4. In the extremely unlikely circumstance of having a tie to break, is it necessary to have more than one tiebreaker?
5. Are we just better off defining all these issues in our SI to suit our own needs?
Thanks for any feedback.
Resources:
https://www.racescore.org/home#h.lw7sb5phqec4
https://www.ussailing.org/wp-content/uploads/2018/01/Cox-Sprague_Scoring_Sys.pdf
https://www.sailingscuttlebutt.com/wp-content/uploads/2018/12/Cox-Sprague-Scoring-System.pdf
Comments
[You must be signed in to add a comment]
https://www.sailingscuttlebutt.com/2022/05/19/deep-dive-into-scoring-systems/
Don't get too hung up on the drops thing, a high-point system (which CS is) with a percentage of participation effectively figures in the drops. SailWave is shareware that has both CS and CHIPS in it, I think it's an add-in. Many long term series are scored this way (I used it for Frostbite racing in my misspent youth), there should be tons of experience out there.
To enable these simply from the menu select - Setup - User Interface and then check the High point scoring you will then get the following options for the scoring system
There is also a Sailwave User Group which you may find handy
Jon
I sent a private email to the developer of the racescore app (and author of the scuttlebutt article referenced above) that we intend to use for scoring our series. Included was a link to this thread. Here are his responses to the question I put out there.
I have developed software (racescore.org) that scores races using Cox-Sprague. Based on that experience I have answered your questions below. Please don't hesitate to reach out if you have further questions - Lee Morrison lm06820@gmail.com
1. Is there a definitive location for the definitions of, criteria used and execution process for the Cox-Sprague Scoring System?
"Definitive" is a little over arching. Cox-Sprague is a system based on a table of scores that references finish position and size of fleet. All of the tables I've seen are consistent but as you have described there are nuances in scoring (not specific to Cox-Sprague) to meet the needs of different fleets.
2. Has anyone had a good/bad experience using this method of scoring?I've done a deep dive on Cox-Sprague vs High Point scoring that outlines the minor differences. In a long running series, where participants aren't expected to sail in all races, both Cox-Sprague and High Point average point scoring systems will do a better job of ranking the "best" sailors (the ones that consistently finish at the top of the fleet - when they sail) over sailors that participate in the most number of races - when compared to scoring with Low Point.
3. Is dropping races a redundancy because of the inherent benefits of this scoring system?There are 2 thoughts here - Average point scoring systems already account for sailors missing races in a series. An OCS, breakdown, or just a poor finish is part of the sport - get over it, is one point of view. The other is that throwouts, that are earned at defined participation intervals, encourages participation by giving sailors an incentive to sail in more races. e.g. for every 4 races (after qualifying) in a series, racers get an additional throwout.
4. In the extremely unlikely circumstance of having a tie to break, is it necessary to have more than one tiebreaker?We've seen scoring circumstances in small, not deeply competitive fleets, where either of 2 competitors typically win every race. There still may be a tie after using average points forcing a determination of who beat who in head to head races.
5. Are we just better off defining all these issues in our SI to suit our own needs?If it will make a material difference to the members of your fleet, then yes. But, as you have probably discovered, the scoring itself can be time consuming and prone to error when done manually. So, then it comes down to the software you use. If the person who does the scoring finds an easy to use product and your fleet can live with the features built into the software, then great. racescore has many features/scoring options built into it - if there is something specific that your fleet needs and its not in racescore, let me know
Rewriting this as a custom formula as needed for direct use in Sailwave (http://www.sailwave.com/), and designating s as the number of starters (replacing n) it becomes:
Or in perl from my own implementation:
I'd be happy to provide it in other languages like C or Swift.