Why flat-pace math fails in ultras
Multiply your marathon pace by 50 miles and you get a number that will embarrass you on race day. Flat-pace math ignores everything that defines an ultra: thousands of feet of climbing that turn running into hiking, technical footing that caps your speed regardless of fitness, and the simple reality that hour eight does not feel like hour two. Runners who plan off road pace routinely miss their estimate by 20 to 40 percent, which wrecks crew logistics, drop bag planning, and pacing decisions in the first half when it still feels easy. The fix is not a better guess. It is a structured estimate where each source of slowdown gets its own explicit allowance.
How this estimate is built
The model is deliberately simple: flat pace times distance, plus a climbing allowance of 6 to 10 minutes per 1,000 feet of gain, all inflated by a fatigue factor for the late-race slowdown. Each knob is stated so you can see exactly what the estimate assumes. Your comfortable flat pace is the pace you could hold on a road for the first hours of a long day, not your tempo pace. The climbing cost covers the time vertical gain adds on top of the horizontal distance: 6 minutes per 1,000 feet for smooth, runnable grades, 8 for typical trail climbs, 10 for steep or technical terrain where you are hiking with purpose. The fatigue factor then scales the whole thing up: roughly 10 percent for a 50K, 15 percent for 50 miles to 100K, and 20 percent for 100 miles. If the result looks wrong, you can point at the exact assumption that is off and change it, which is precisely what a black-box predictor never lets you do.
Worked example: a 50 miler with 8,000 feet of gain
Take a 50 miler with 8,000 feet of gain, and say your comfortable flat pace is 10:00 per mile. Fifty miles at 10:00 pace is 8 hours 20 minutes of flat running. The typical 8 minute climbing cost adds 64 minutes for the 8,000 feet of gain, bringing the base to 9 hours 24 minutes. A 15 percent fatigue factor inflates that to about 10:48:36 of moving time, an average of roughly 12:58 per mile even though you never ran a slow mile on purpose. Then add stops: with eight aid stations at 4 minutes each, call it 11:20. That last step matters. Tell your crew 10:48 and they will be pouring coffee while you are still two ridgelines out.
The error bars, honestly
This is a planning envelope, not a prediction. Terrain technicality, heat, altitude, night running, and how your stomach holds up swing real results by an hour or more at 50 miles and several hours at 100. No formula sees the storm that turns the last descent into a mudslide, or the 20 minutes you spend at mile 70 deciding whether you can keep food down. Use the estimate for what it is good at: sanity-checking cutoffs, telling your crew a realistic arrival window instead of a fantasy, and deciding whether your goal requires running the climbs or allows hiking them. Then update mid-race. If you hit the halfway aid station 40 minutes behind the model, the model was wrong about something, and the second half plan should be built on what the day is actually giving you.
Related tools and guides
To sharpen the climbing side of the estimate, the grade adjusted pace calculator shows what hills really cost per mile, and the hike-run pace calculator blends running and hiking paces for the strategy most ultras actually demand. Fuel the hours you just estimated with the ultra fueling planner. If your race has no fixed distance at all, the backyard ultra calculator handles yard pacing and rest math, and training for a backyard ultra with a job and kids covers fitting the volume into real life.
Frequently asked questions
How do I know my climbing cost?
Look at training data from terrain similar to your race. Take a hilly long run, subtract what the flat miles should have taken at your easy pace, and divide the leftover time by the thousands of feet you climbed. Most trail runners land between 6 and 10 minutes per 1,000 feet. If you have no data, use the Typical (8 min) setting for normal trail climbs, Runnable (6 min) for smooth fire-road grades you can jog, and Steep/technical (10 min) for hands-on-knees hiking terrain.
What fatigue factor should a first 50K use?
Start with the 50K-ish (10%) setting, and be honest about your training. Ten percent assumes your long runs have prepared you for most of the distance and the wheels stay on. If your longest run is well short of 20 miles, or the race is hotter or hillier than anything you have trained on, bump to 15 percent. First-timers who fade usually fade because they picked the optimistic number.
Does downhill running save time?
Less than people hope. Descents return only a fraction of what the climbs cost, because braking on steep or technical downhill is slow, and hammering the runnable descents costs quads you will need later. That quad damage is a big part of the late-race slowdown the fatigue factor covers. The calculator deliberately does not credit descent time; treat any downhill gains as a small bonus, not part of the plan.
How do cutoffs relate to this estimate?
Remember the estimate is moving time. Add 3 to 5 minutes per aid station, then compare that total against each cutoff, not just the finish. If your projected arrival at a mid-race cutoff is inside 30 minutes of the closing time, you have no buffer for a low patch, and low patches are guaranteed in an ultra. Either the goal pace needs to firm up or the aid station stops need to get shorter.
Why do race time predictors fail for ultras?
Standard predictors extrapolate from road times using formulas built on flat, continuous racing. Ultras break every assumption in that model: vertical gain, technical footing, hiking, aid station stops, eating on the move, heat, altitude, and sometimes a night on the trail. A predictor might say you can run 50 miles at 8:45 pace; the trail says otherwise. Building the estimate from flat pace plus explicit climbing and fatigue allowances, the way this tool does, keeps every assumption visible so you can correct the one that is wrong.