Yeah, but getting the results right includes not having people in the results who didn't do the full course. At the Ironman races we run a script intermittently once the finishers start coming in. It lists everybody who misses any of the timing wires on the course, at any time, all day. We then go through their race segment by segment to figure out what happened. They either get DQ'd right there or if it looks legit they get marked as OK. That's the one time the timer will do the DQ's and not wait for a ref or RD to do it. At the end of the night that script has to come up blank - either people are DQ'd or marked OK. That's the only way somebody with a missing split isn't flagged by that script. After the race Jimmy Riccitello gets a list of all the DQ's and he follows up. And if anybody doubts it, he does follow up - even if it's a plodder at the back of the pack. Jimmy emails and calls me all the time to ask for detailed timing information on DQ'd athletes or just to hash it out back and forth to make sure that we get it right. We both want all the short cutters out and everybody who does the full course in.

Dev mentioned that Marc told him that 5-10 athletes will get DQ'd on average at a given race. Usually it's at the low end there, but I've DQ'd up to 30! Seriously. In a full-on Ironman 70.3. How many of those get reinstated later? Usually zero. It ain't rocket surgery. If they're missing split times and their time from the split before to the split after is out of whack - faster than you'd expect based on the rest of their race - they get nuked. I'd be happy to reinstate somebody if they can produce a GPS file that shows that they did the whole course but that doesn't tend to happen. I've seen more bogus screen shots of somebody else's Garmin activity (two occasions, including T3 of course) than actual tcx/gpx files or links to a real GPS activity proving that somebody did the whole course (zero occasions). The chips we use don't miss many reads and if, for whatever reason, there is a miss, the running/cycling speed from before and after the miss will tell the tale. Usually they'll go out at 9 minute pace, finish at 11 minute pace, disappear for the middle split and would need to be going 7 minute pace for the time they disappear. It's rarely hard to figure out. There might be 1 of those on average at a race, sometimes none, sometimes it's ridiculous (cough, Miami). The most common anomalies are people who finish the race after failing to do a lap. I've had people finish IMAZ after one lap of the three lap run. In a busy race I don't bother putting much thought into a situation like that. Finish after 1-lap? DQ. If they email after the race saying they just wanted to DNF and give their chip back and would rather be a DNF than a DQ, I'll go look at Finisherpix and look for a pic of them posing with a medal. I'll switch them to a DNF if there isn't one. If there is one, they stay a DQ and Jimmy gets informed that they tried to get a DQ overturned despite posing for a finish pic with a medal. Don't even think about it!

OK, Vancouver 2014. Why no DQ for the athlete in question? I asked the timer and he couldn't clearly remember but it looks like he didn't have a wireless connection to the one split point on the bike course, so there was a delay getting the times from there. In his defense, that Vancouver race is a tough day. Look at the list of events: Long Distance (70.3) @ 6:30 am; Aqua Bike @ 6:30 am; Olympic @ 7:30 am; Sprint @ 8:04 am. So there's a lot to do and by the time he downloads the split times from the bike he's probably already printing awards and putting out whatever other fires have popped up. He'd check for missing laps on the bike, but wouldn't have necessarily checked for fast laps. That race was using the older ChampionChip technology which is not conducive to multiple bike splits, so there was only the one, in a parking lot at the start of each lap. If there was one at the far end of each loop, the two shortcut laps would have had a missed read and a more obvious DQ situation. Lacking that, the reason that ugly text file has a ranking position for each split is so that you can easily scan up and down the full list of results looking for single digits where there shouldn't be single digits. Like those two anomalous 1st place bike splits which should have meant insta-DQ. So, no doubt, we should have caught that one. If there's a situation like a 2-lap in-water swim with no exit for lap timing, or if there's a crazy mickey-mouse-head-shaped bike or run course that doubles back on itself, it's possible that people can cheat and not be flagged by missing a wire. In that case we also have a script to list the fastest times for every intermediate segment. That's a bit harder to sleuth through but for example it'll show the time and pace/speed for the segment alongside the overall time and pace/speed for the discipline. So if somebody has the 30th fastest segment and the 400th fastest run overall it'll make you wonder. But then what? If they have a Kona spot and they go from 7:30 miles to 6:00 miles to 8:00 miles I would probably DQ them and wait for the appeal. If they're MOP or BOP, honestly, I'll probably let it slide as long as they don't miss a wire and aren't anywhere close to a rolldown slot. And if it's a big 70.3 I probably won't get the chance to analyze those splits at all. So there's some rudimentary technology at work but it relies on the organization and insight of the timer. If somebody wants to design software or a script that will take an import of every single split time recorded during the race, analyze it and spit out a list of athletes whose pace variance in a discipline is higher than an agreed upon standard deviation, OK. It's easy to spit out the raw data, but you need to get the timing companies and RD's and race mgmt corp's to agree to it.

---

Sylvan Smyth | http://www.sportstats.asia | sylvan@sportstats.asia | Starvas