I was recently asked for my opinion on implementing time standards as part of selection for National team programs. Those of you who know me can likely guess my response; I think they are almost useless. ”But,” the argument typically goes, “we need to have some form of objective criteria to help determine who is worthy of being on the team (i.e. worthy of being funded) and no one has ever come up with a better alternative”. I hate to be that guy who does nothing but complain while not offering a workable alternative, so here’s my shot at an alternative to time standards.
First, I feel like I should articulate the reasons that I think time standards are so useless. The idea of the time standard is that we choose a time that, if achieved by a Canadian athlete, gives us a reasoable expectation that the athlete is capable of a certain minumum performance. So we take some historical data, crunch the numbers and spit out a time that must be achieved. Assuming that the source data is reliable and that our Canadian athletes get to test themselves against this data under neutral conditions then the time standard holds up as a reasonable measure. Let’s look at those assumptions.
How do we know if the source data is reliable?
Because we all know that wind plays such a huge factor in times (it is no stretch to think C1 times can be off 10-20% over a 1000m race) we need to include as large a sample size as possible to come up with a time for ‘an average day’. However, when you include World Cups in the mix you run the risk of including races where none of the World’s best are competing, or including events where the best are not pushed to their best times. Either of those case would skew the results if included.
If we include only World and Olympic Championships where we know that all of the best are present and racing for their lives, we reduce our sample size to probably 4 or 5 events. Any more than that and you risk using times that are no longer as fast as they need to be due to the gradual improvement that occurs over time. The small sample size means the results can be skewed by weather conditions or other anomalies (think David Cal going 3:45 in Athens 2004).
Can we normalize the source data?
Theoretically we could take into account the effect of weather. Given enough data it becomes an exercise in statistics. If you went back over many years of data you could match certain weather events to changes in expected times – though of course you would need some way to determine the ‘expected time’. You would also need accurate weather data. How many regattas do you suppose would have collected weather data that would stand up to scientific scrutiny? If my involvement in Canoe ’09 is any example I would wager the number is vanishingly small (we did collect wind and temperature information, but I will not guarantee that it is accurate, or that it was actually taken during each and every race).
We also could theoreticaly calculate the effect of wind on the athlete based on physics. It is fairly trivial to account for an average athlete offering an average surface area facing a wind coming from a certain angle. But this misses some very important elements.
First, waves become a big factor at some point. In a strong tailwind the wind eventually stops assisting you because the waves are hindering your progress. In a strong head wind you have to overcome the force of the wind and the waves crashing over your boat. So wind can be compounded by waves.
Second, cross winds do not actually cause sideways motion but cause the athlete to apply additional energy just to keep the boat running straight. Not only that, but canoe athletes are affected in completeley different ways depending on whether they paddle on the windward or leeward side.
Third, there are other factors that have a compounding affect as well. Rain cold weather, course depth and water temperature would all need to be accounted for.
Based on all of this I say that normalizing times based on theoretical calculations is too complex. Even if we were able to come up with a formula to do this, we would need more reliable ways to take wind measurements over the length and width of the course since conditions can change dramtically from start to finish and from lane 1 to lane 9.
OK, but anyone with knowledge of canoeing could come up with a good guess at the winning time at the World’s this year.
Sure, I can tell you that in good conditions the winning time in 1000m C1 will almost certainly be between 3:48 and 3:52. We could use that as our time standard, even take off a percentage to give us some reasonable expectation of the 10th best time in perfect conditions. Then, if we get those great conditions at Trials we have a comparison to make.
If we don’t get those conditions we have three choices. Apply the standard absolutely. That is, regardless of the wind, if you do not achieve the standard you face the consequences. This is dangerous as it is completely possible that even our best athletes miss the standard on a bad day in Montreal.
Have a threshold at which you throw out the standard altogether (e.g. if the wind exceeds Xkm/h the time standard will not be applied).
Attempt to normalize times to the standard. Well, I’ve gone over that above but let me throw in this scenario. Pretend that we do have the ability to normalize times for wind, we have a left’s wind, a left winner and a right who was second by half of a boat. The normalized times changes the order of finish…now what?
So that is my longwinded summary of why I don’t like time standards, and why I don’t think they should have much weight.
None of that is news, it has all been said a thousand times by thousand voices. The problem is that we all accept that an objective measure is necessary and don’t offer an alternative. I predict that my alternative will not be very well received either, and will not make it into any actual selection criteria in the near term. That’s because my criteria in NOT objective, but rather almost wholly subjective.
Are you kidding me?
Now hold on, don’t get your knickers in a knot. I still believe that in the end if we are choosing the best Canadian crew for the World Championships or Olympic Games, that the crew comes from direct racing results. However, I see some different classes of selection.
Realistic medal chances are known at least a year in advance. For example, we know that our K1 1000m entry is going to be in a position to fight for a medal in London. So, we base selection on some kind of racing series where the winner goes to the Olympics and the loser stays home. We have no objective as a Nation other than making sure that the fastest person gets to go to the Olympics. No need for any type of Time Standard.
Outside chances at medals are also usually known in advance. These are events where we have a reasonable expectation of making a final and if all goes extremely well over the course of the summer we could even fight for a medal. We should treat these the same way we treat Realistic medal chances; a series of races in which the best overall result gets the entry. And again, no need for a time standard. The only issue is determining whether we have an outside chance at a medal or not. If our National Team coaches can’t make that determination based on the last couple years of races, we are paying the wrong guys.
Senior World’s bubble chances are events in which we did not make a final at the previous World Championship and realistically would need a superior performance to have a chance at the final in the coming year. This is where we must face the reality of the current funding model – perform or perish. We need to target the limited dollars at the best medal hopefuls and so, we need a way to determine which of the several ‘bubble’ entries deserves enough funding to attend the next competition. But we are not choosing events that we think we actually have a shot at this year. If these events are chosen at all it is because we believe there is a medal chance down the road. So, I would lump our bubble events in with…
…Development Opportunities. This can be a Junior World event, a Pan Am event, a U23 event a Sr. World Event that doesn’t have a hope for a medal this year, or a slew of athletes that should get National Team funding but aren’t yet good enough to get an entry at the Senior World level. The reason that I lump all of these together is that they all have the same objective – to provide opportunities for future potential Olympic medalists to compete at a high level and, by learning from that experience, move closer to being an Olympic medal threat. Note that this means that our objective at Jr. Worlds IS NOT to field the best possible slate of entries, but to field the entries that are most likely to develop into Olympians in 6-8 years. This is where a reliable time standard would make everything fairly easy. If you could set a normalized time for the very best in the world, it is easy enough to scale that time back to suit a U23 athlete, a junior athlete or an athlete hoping to make a B final at the Sr Worlds. You then look at all the different events and take the athletes who are closest to where you think they should be relative to the time standard. But as I have tried to show above, time standards are useless pieces of fiction. So, I think that the national team coaches should get to decide.
GAAAH! The Lawsuits!!!
I know the objections. The coaches are biased, they can’t be trusted to make the right choices, they can’t possibly recognize Canada’s future elite athletes. I call bullshit on all of that. Our National Team coaches have been professional coaches for a long time. We should be hiring them precisely because they have the skill to recognize great athletes and turn them into great canoe athletes. I also recognize the need for something to base decisions on. Here are my selection criteria. These can be weighted as you wish, but the basics are:
- Is this athlete competitive against Canadian athletes who are older and more experienced (i.e. a Midget who is winning at Junior)?
- Does this athlete’s technique conform to the “Technique Block” we are trying to encourage (more on that some other day), or is their success based mostly on physical development?
- Has this athlete shown the ability to adjust their technique based on coaching, that is, are they coachable?
- Does this athlete consistently train at a high physical level?
- Has this athlete achieved a time standard?
- Is this athlete currently the best in their event?
Athletes that score highest in these categories are the ones we want to mold into future Olympians.
This would obviously require the National Team coaches to spend a lot of time with potential members of our National Team. There are components here that require evaluation throughout the year, not just on the day of the trials. It would be a lot more work for the coaches. This would also require a tremendous amount of trust between the National Coach, personal coaches and athletes.
I am curious to hear what others have to say on this matter. Please comment if you have an opinion. I review all the comments before they appear, so I apologize if you see some delay.