# How does race timing GPS actually calculate the gap?

#### saunaking

So there are 2 groups on the road and the GPS says the gap is 1'30"

How is this calculated exactly?

Presumably the organizer has a moto with both groups.
Is it the motos that have the GPS on them or something on the rider's bikes?

Either way, we now know the position and speed of both groups.

Seems to me the best way to show a gap would be to (assuming it is known) display the time the 1st group took to arrive at their current position from the
current position of the trailing group.
The downside of this method is that if the chasers are fresher, they will naturally take less time to cover the distance than the leaders and thus the gap
will be (slightly) overstated. But its the best method i can think of.

Is this how it works?

Other options could be to take the speed of the chasers and the gap distance to make the calculation, but this would not be accurate if the terrain changes and would work
terribly if there is climbing or descending involved

Would love to know how these gaps get calculated.

#### spalco

Seems to me the best way to show a gap would be to (assuming it is known) display the time the 1st group took to arrive at their current position from the
current position of the trailing group.
The downside of this method is that if the chasers are fresher, they will naturally take less time to cover the distance than the leaders and thus the gap
will be (slightly) overstated. But its the best method i can think of.

Is this how it works?
I don't have an answer to your (interesting imo) question, but this would basically be the traditional method of handtiming gaps, no? You have a guy with a stopwatch on the side of the road who measures how long it takes for the chasing group to reach the point where the lead group passed through.

#### saunaking

I don't have an answer to your (interesting imo) question, but this would basically be the traditional method of handtiming gaps, no? You have a guy with a stopwatch on the side of the road who measures how long it takes for the chasing group to reach the point where the lead group passed through.
Yes very similar except if there is a larger gap, say 5' , then the gap is actually what the gap was 5' ago and not really 'realtime'

#### Logic-is-your-friend

BANNED
Don't know how they are, but how they should be calculated. simply start counting at any given point when and where the first group passes, until the second group passes. Just like when they cross the finishline.

It's a shame that there is still no decent solution for this, in this day and age, while the technology has been available for ages. Every rider has a transponder with a GPS as far as i know. If they would just put that info into software and keep track of where a rider is and passes by at which point (basically, no different from what an app like Strava does), this problem would have been solved ages ago. They could even give time gaps for any rider where ever in the race. Be it the leader or the red lantern.

#### Netserk

It's a shame that there is still no decent solution for this, in this day and age, while the technology has been available for ages. Every rider has a transponder with a GPS as far as i know. If they would just put that info into software and keep track of where a rider is and passes by at which point (basically, no different from what an app like Strava does), this problem would have been solved ages ago. They could even give time gaps for any rider where ever in the race. Be it the leader or the red lantern.
That has been the case for the Tour for years.

#### Logic-is-your-friend

BANNED
Yes very similar except if there is a larger gap, say 5' , then the gap is actually what the gap was 5' ago and not really 'realtime'
There is no "real time", considering the "gap" is the distance/time between both groups and thus variable, and in order to be accurate, you need them to have passed the same point. The software can not predict the future, so what you are expecting is simply not possible. Let's say there is a 3 minute gap, judging by the distance and speed at which both groups are riding, but then suddenly there is an attack in one of the groups, and the time goes up or down dramatically. Then the calculation of 5 seconds ago, will already prove inaccurate.

That has been the case for the Tour for years.
Thanks. So in the Tour it is basically confirmed that all the times are 100% accurate? I didn't know they calculated gaps like that already.

#### spalco

There is no "real time", considering the "gap" is the distance/time between both groups and thus variable, and in order to be accurate, you need them to have passed the same point. The software can not predict the future, so what you are expecting is simply not possible. Let's say there is a 3 minute gap, judging by the distance and speed at which both groups are riding, but then suddenly there is an attack in one of the groups, and the time goes up or down dramatically. Then the calculation of 5 seconds ago, will already prove inaccurate.
I disagree, a smartly programmed algorithm could certainly do that (to an extent), and extrapolate a more realistic time gap based on current/expected speed differences and gradients.

But I would question if we would actually learn any new information that way or if we are better off figuring out in our own minds if a gap's coming down or not.

#### Logic-is-your-friend

BANNED
I disagree, a smartly programmed algorithm could certainly do that (to an extent), and extrapolate a more realistic time gap based on current/expected speed differences and gradients.

But I would question if we would actually learn any new information that way or if we are better off figuring out in our own minds if a gap's coming down or not.
You can disagree, but i don't see much wiggle room here. A scenario:

Group 1 leads with 3 minutes at a timecheck. Group 2 obviously passes the timecheck 3 minutes later. Both are riding a somewhat constant pace. Software and a GPS could calculate that this gap is also 3 minutes, based on the gap (in distance) and speed at which both groups are traveling. Suddenly there is an attack and the attacker takes 5 seconds by count (timecheck) method. What does the software do? It now has to predict a trajectory based on the speed the attacker is currently riding, but who knows if that speed will still be the same 10 seconds later. Maybe the attacker blows himself up, maybe he sits up when he sees others react. Then the tempo suddenly drops, until there is a new attack. Software can not predict when an attack will be placed, how hard it will be, how long it will be sustained etc. So there is no real benefit to be had. In this scenario, it's possible the software at one point estimates the gap to be 2m30s, while in reality, the gap at the timecheck will be closer to 4 minutes.

Best possible way, imho, is to have every rider 's time and position to be recorded in software, at certain intervals, either in seconds or meters. Say every 5 seconds or every 10 meters the position of the rider is stored in memory. That way the software can compare how much time the chaser is behind at every point he passes where the leader already passed before him.

so nobody knows how it works exactly?

#### spalco

That was a decade ago, has nothing changed?

Are they even doing the thing with the blackboard still?

I don't think the blackboard has gone away and using the TV bikes for timing gaps is still a thing as far as I know,

#### bNator

You can disagree, but i don't see much wiggle room here. A scenario:

Group 1 leads with 3 minutes at a timecheck. Group 2 obviously passes the timecheck 3 minutes later. Both are riding a somewhat constant pace. Software and a GPS could calculate that this gap is also 3 minutes, based on the gap (in distance) and speed at which both groups are traveling. Suddenly there is an attack and the attacker takes 5 seconds by count (timecheck) method. What does the software do? It now has to predict a trajectory based on the speed the attacker is currently riding, but who knows if that speed will still be the same 10 seconds later. Maybe the attacker blows himself up, maybe he sits up when he sees others react. Then the tempo suddenly drops, until there is a new attack. Software can not predict when an attack will be placed, how hard it will be, how long it will be sustained etc. So there is no real benefit to be had. In this scenario, it's possible the software at one point estimates the gap to be 2m30s, while in reality, the gap at the timecheck will be closer to 4 minutes.
I agree with you. It makes sense to feed the only reliable information at disposal which is, time difference between two groups at certain checkpoint. Of course, checkpoints should be as close together as possible. But any extrapolation or prediction would be pure guesswork at the expense of the only exact information we have.

#### spalco

I agree with you. It makes sense to feed the only reliable information at disposal which is, time difference between two groups at certain checkpoint. Of course, checkpoints should be as close together as possible. But any extrapolation or prediction would be pure guesswork at the expense of the only exact information we have.
I wasn't advocating for something like that, I just stated that I believe it would be technologically possible.

#### bNator

I wasn't advocating for something like that, I just stated that I believe it would be technologically possible.
Sure it’s possible but you need to consider a couple of factors. I’m not sure what kind of equipment the organisers use but iPhone for instance will produce GPS locations of 5m accuracy at pretty much exact time period of 1 second and it could relay these data via LTE to backend server each second as well (though I’m not sure the battery would last the entire stage). This is actually better GPS location sampling rate than some cycling head units are capable of but these head units are of course not capable of LTE communication (they only talk to a phone using Bluetooth which consumes significantly less energy).

Whatever the organisers would use, it should be significantly smaller than iPhone and I’m not sure it’s a good idea to sample GPS each second and it’s especially bad idea to send this data via LTE to a server each second. This is the most energy consuming process.

These circumstances need to be considered by the developers of that system and getting GPS data to a central location in real time and frequently enough while keeping the units small is still a technological challenge...

#### Walter1234

I understood from the Belgian Dutch commentators at some point it is now calculated by a motorcycle with GPS info at each group.
That is also why sometime the timing is completely off when they put the wrong motorcycle against a certain group (the motorcycle has stopped or drove to an other group).
But how they calculate I have no idea.
I would expect they use the past location data and timing of the motorcycle at a leading group and constant calculate how much later a motorcycle at other groups passes at past location from the leading group.
And then I expect there could be errors depending om the position of the motorcycle relative to the group he follows.
Other automated calculations don't seem to make sense (exceptif you know the exact riders position).
In the examples above about attacking upcoming groups, I don't think that are the major challenge I. The speed is maybe 10-20 % difference.
When the leading group starts climbing , projected speeds could cause much bigger issues.

#### Finn84

With the peloton I suspect GPS can give different gaps depending who do you compare? It might be some 20 seconds between front and back.

#### Nick2413

Side question unrelated: In sports like swimming and track running where there is a WR line on the tv, are they using an averaged time throughout the event or a per-lap or actual time? For example does the WR line speed up knowing the athlete who set it probably went faster on the last lap?

#### Googolplex

Side question unrelated: In sports like swimming and track running where there is a WR line on the tv, are they using an averaged time throughout the event or a per-lap or actual time? For example does the WR line speed up knowing the athlete who set it probably went faster on the last lap?
I think in swimming it goes per lap. However on a lap, it goes at steady pace from end to end. However, in reality, swimmers go faster at the start of the lap, because they can kick speed from the end of the pool. Thus, it often seems that the swimmer is ahead of the line, but gets tired in the end.