Thoughts on Iron Butt Rally routing

The Iron Butt Rally has been described as a scavenger hunt with the only prize a three digit IBA number.The Iron Butt Rally is really better described as go where ever the rally masters decide. The original Iron Butt Rally was basically ride a pile of miles and make it back to the starting line. More recent incarnations have been ride to all of the state capitals or at least states. Previous versions were better characterized as time distance sensitive scavenger hunts. When routing one of these scavenger hunts you have to think about character of the roads, events and activities along those roads, and the inherent number of points you can pile up reasonably. As an example if you 24 hours to travel from check point to check point and there are bonus points 1700 miles away you’d have to exceed the speed limit by triple (round trip of at least 3400 miles) and though some bikes can do that not many riders would be willing. I’m definitely not willing. So, you set up some tools and look at reasonable expectations.

I route based on a few basic characteristics. I plot all of the points and bin them into 8 various size categories so I can visualize the number of points. For an out and back ride (like the 2013 Iron Butt Rally start to check point #1) I figure a radius that to me is a reasonable distance for the total time to accomplish and the number of stops I need to make. In a 24 hour rally that is fairly easy to figure. Each stop whether for gas or a bonus is usually going to be a 10 minute penalty on my overall time. In more than a 24 hour rally I need to build in stops for rest and food.

Each green circle is denoting by size the relative number of points. The light lined circle represents a time distance calculation giving me an approximation of total distance out that I can travel.  It would be possible to do a route with this map that went directly west, then cut back to the south eastern point, and then back to the start/finish. However, the radius of the circle would be nearly half and more points would drop out then you might expect. This kind of balance is what the scavenger version of Iron Butt Rallying is all about. It is an extremely large puzzle that has some interesting side effects.

Consider that the current Iron Butt Rally for 2013 looks something like this.

Leg		Hours	Avg MPH	Distance	Base Route	Base Hours	Difference
Pit-Pit	Leg 1	78	45	3510
Pit-Sac	Leg 2	60	45	2700	2466	54.8	234
Sac-Pit	Leg 3	98	45	4410	2466	54.8	1944

I’m using a nice easy set of numbers here just for the purpose of explanation. To complete a SaddleSore in the United States it requires no more than a 45 mile per hour average. Consider then that I just completed a 9 day trip where my average speed over five days was nearly 20 miles per hour faster and I was not speeding. In fact on several occasions as the traffic stream warranted I was doing less than the speed limit. Full disclosure I am not counting the three days I sat on my butt in the middle of that ride in a conference room as part of the averages. What I did do while riding was manage my time very closely and make sure that I was always moving and not stopped by rush hour traffic or similar time wasting activities. It is hard for people in the denser population areas to understand your average speed is fairly high when you go several 300+ mile increments with a speed limit of 75 miles per hour. As you go further south you have 80 mile per hour speed limits on wide open highways.

I try and build into my calculations some amount of slop time. The grid shows 60 hours. The actual time might be 64, 68 or 70 but I self penalize in what I call the, “Leave late come back early” version of rallying. Stuff goes wrong. I like to be “safe”. I am very risk averse in that I ride motorcycles kind of averse way. I like to be prepared and plan with that in mind. Not seen in the grid is the opportunity to bonus point tax myself. For each bonus point I subtract 10 to 15 minutes based on the bonus point for pictures, scoring, and recording information. Some bonus points have lengthy walks or hikes where the rally master just wants to get you off the bike. Some bonus points have mandatory rest periods which also need to be deducted from the total time available. Some bonus points it is plant your flag, take a picture, record the information and keep going. Just knowing that a rest bonus is likely in the Pit-Sac leg in 2013 tells me that they really want us to do nothing but haul butt (figuratively speaking) to Sacramento.

The first picture showing the Jacksonville rally represents a known amount of time and a limited geography. The Iron Butt Rally Leg 1 bonus listing could be comprised of basically the entire east coast of the United States and Canada and most of the midwest. The Leg 3 circle is most of North America. The Leg 3 route though is not an out back so figuring out a radius just isn’t needed. The circle and radius method is of limited utility for most rallies. Few rallies are out and back but when they are this is one method of finding the route budget.

This problem is known as the traveling salesman problem. Given a set of nodes that you need to touch in as quickly as possible and weights on each of the connectors to the first node you can start to figure out a best path from start to finish. In very simple graphs this is not hard to do. In really large and complex graphs it is nearly impossible. When you start imposing other costs on the graph concept like sleeping, eating and gassing up the bike the theory goes out the window. There are also other constraints such as time windows for bonuses, and daylight only bonuses. How do you know the window for daylight four time zones away? Where would you look to find out? A salesman directed graph in a very simple form starts out as the following picture dictates. Costs for each node are values that can be calculated. The other difference between this and the traveling salesman problem is that you are not returning necessarily to the start. In some rallies that is true, but not in all rallies.

The following image depicts the node on the left and the cost to get to another node in the center. The goal is to get to the furthest node. The target node is on the right. In the following diagram you can follow how each is supporting getting from A to the last node in the cheapest path possible. What we haven’t done at this point is determine and end point or total budget for costs. Cheapest is the metric most computer scientists learn as a primary metric problem. The issue is that theory is great, but each node not only has a cost, but it has a significant set of secondary constraints. So you have to prioritize and evaluate also against a total cost and highest level of efficiency. To be sure there is just a certain amount of luck involved in routing. The dynamic nature of the traffic stream insures that you have to take into account things like events, rush hour, closed highways, brush fires, weather and other events that might not be costed in the original plan.

What happens in a rally is the cost to get to every bonus is such that you have to select a path through them that has the lowest cost and highest reward with the reward meeting or achieving an arbitrary value. The cost or total budget to accomplish all of the bonus points is also an arbitrary value. Efficient routing is a non-trivial task. In routing you have to find the path through the “bonus” locations and end up at a finish or check point within a particular window of time. This is most assuredly not a race and is most assuredly a thinking persons game. When people talk about rallying I don’t think they truly understand how much of this is just thinking and enjoying the riding. It is a non-trivial set of tasks but it is not fastest/bestest type riding.

Adding an additional component to the concept of traveling salesman problem we include a finish point. This breaks the original problem and actually makes it easier for us poor humans to figure out. Since we don’t have an infinite budget of time the constraints actually give us metrics to bounce decisions off.

It is easy to set up a single dimension but you  have to think about the costs of each segment within a certain total budget. So if you only have a limited budget you can determine items to leave off the plan, and those to add to the plan. In the following graphic option 1 and option 2 show two different paths. I’ve left off the cost from F to finish as that was not in the original node graph. You can go back and look at how the directed graph gave costs and see where the items in boxes came from. This is a made up problem and so don’t look at actual distances between nodes. Two items on a map can be very close together but the distance or cost to get there can be more than something three times the physical distance away. As an artifact of explanation to a lay audience this should work.

In this example of two options if the budget is greater than 9 it makes sense to go with the first option. Option 2 does not provide nearly as many points as option 1 for the same cost.  I’m a very risk averse person so I like to leave a little extra in the budget so I can feel less stress, angst, and generally just enjoy myself more. I might skip a bonus location just to manage my cost budget better. That makes me less likely to gain the most points possible but it does mean I am more likely to finish. As they say your mileage may vary.

The base route from Pittsburgh to Sacramento is a highway slog that is pretty simple run down Interstate 80. That would give your likely minimum cost for finishing the leg. What you would have to do is figure out the north and south max points for the triangle and create an elliptical curve for bonus points that are within your range and budget you have set. In other words how far can you diverge from the least cost (minimum budget) path and still finish within the window given. Those are elements that are personal to each rider.

What are some of the options that are possible for the third leg of the 2013 Iron Butt Rally? It is pretty obvious that a lot of points on the map are possible within a fairly conservative budget. As an example there a few places everybody discusses up front. There is the the always bad but somebody will do it Sacramento to Keywest Florida return trip during leg 3 of the rally. Some feel that Keywest is possible during leg 1 of the rally too.

Perhaps Sacramento to Pittsburgh via Port Nelson?

If Port Nelson is possible then Hyder Alaska is possible. Pushing the ponies harder some other really neat places are possible. The issue for some and glee for others is we have no idea on the 2013 Iron Butt Rally where people might go. The riders will be first to find out, but there will be years of arm chair routing done by people who didn’t have to look at the weather, events, previous legs of the rally, and all of that in real time. I think it is funny when I see somebody with the excess of time of years since a previous rally say how with nearly infinite time they would have run a better route than some previous winner. A previous rally rider might have had 2 or 3 hours to make their route within the constraints of that rally at that time.

In closing this is my way of looking at routing over all and how I set up my rides in general. I don’t do a lot of rallying and I don’t ride with a competitive mind set. I just like the chess match and thinking strategies necessary to accomplish what I can do with my capabilities in a fairly risk averse manner. Motorcycling is an activity that for many is a lifestyle and to others just a mode of transportation. I enjoy being in the environment partaking of the heat, bugs, wind and rain and not shielded from the world. Rally routing and planning as a form of efficiency makes me no money, gives me  no accolades, but it does give my overactive brain something to churn out even if I’m more wrong than right. This is the basics of what I call fun.