A Race Needs Analysis is how Predictive Coaching answers the question every athlete asks before race day: what can I actually do here? It runs your race history, current fitness, and the course profile through a regression model, and returns a clear picture of what your goal takes, what you’re capable of today, and where the gap sits between the two.

A needs analysis isn’t a prediction. It’s a forecast. Predictions promise. Forecasts don’t.

Every race has its own persona. The course has hills or it doesn’t. The field has historical trends. And you show up with fitness that keeps evolving every workout. A forecast takes all of that, models what’s likely given the inputs, and gives you something real to train against instead of a gut feel. Same idea whether you’re chasing a 70.3, a full IRONMAN, a marathon, or a half. Different distances, same approach.

So what actually goes into the model?

Course Profile

No two courses race the same. A flat 70.3 bike leg and one with 3,000 feet of climbing are the same distance on paper. On race day they’re not even close. Elevation, terrain, whether the course finishes lower than it starts. All of it changes how you pace and what gear makes sense.

For runs, both numbers matter: total climbing and net change. Boston has a net drop of about 450 feet from start to finish. Chicago is flat. Those aren’t the same race, and the model doesn’t pretend they are. Swims work the same way. A calm lake is nothing like a choppy ocean point-to-point.

Where does the course data come from? The source, when we can get it. Actual GPX files from race organizers or your own recorded rides. Not marketing pace charts. And when a course changes from year to year, like World Champs moving venues or a reroute for construction, the model has a specific override for that year instead of overwriting the default.

Race-Field History

Your training log can tell you a lot. But it can’t tell you how your age group has raced this specific event over the last five years. That’s what field history is for. Are podium times getting faster? If so, why? Are they holding steady or slowing down? What’s the actual spread from top 10 to mid-pack at this race?

Running a 3:20 marathon is one thing. Running a 3:20 marathon at Boston, in the 45–49 AG, in a year when the AG median was 3:30, is a completely different thing. Same time, different story. That’s the context the forecast picks up.

Your Race History

Your own race history is one of the richest inputs we have. And not just finish times. Splits by discipline, across different courses, in different conditions. One race is a snapshot. A stack of them is a trend line. Trend lines beat snapshots every time.

The model also knows what your fitness looked like when you raced each of those races. The FTP you had going into that 70.3. The run threshold you carried into that marathon. And that matters. A 3:20 marathon off a 7:00 threshold is a totally different data point than a 3:20 off a 6:15 threshold. If you compare old races against today’s fitness, the math lies. Comparing them against the fitness you actually had? That’s honest.

Race history also shows how you respond to specific courses. Maybe you always lose time on climbs and win it back on the run. Maybe your open-water swims are way slower than your pool times would suggest. Patterns like those are coachable. But only if you can see them.

Current Fitness Thresholds

Where you’re at right now matters just as much as where you’ve been. FTP on the bike, Critical Swim Speed in the water, threshold pace on the run. Those are the anchors. They tell us what you can hold today and what’s realistic to shoot for.

But thresholds aren’t fixed. They move with training, fatigue, taper, and life stress. So the forecast uses your current values, not the ones from a test you did three months ago.

Forecasting Thresholds Forward

Here’s the piece I like the most. If your goal race is twelve or eighteen months out, using today’s threshold to plan for it assumes you’ll stop getting faster the day you sign up. That’s silly. It’s also inaccurate.

Say your run threshold has gone 6:45 to 6:30 to 6:15 over three years. The model doesn’t pin you at 6:15 on race day. It runs a linear regression across your threshold history, projects it out to the actual date of the race, and asks what that version of you can realistically do. Same for FTP and CSS. Confidence follows the data. A clean, steady trend forecasts with high confidence. A noisy trend forecasts more carefully. And if you don’t have enough history to project honestly, the model just uses today’s number.

That’s what I mean by forecast, not prediction. A prediction ignores time. A forecast respects it.

Thresholds are one signal. Split-level trends are another. Are your bike times on similar courses getting faster? Is your run off the bike getting faster? Or are you fading harder in the back half?

That’s where you can tell the difference between someone who’s actually getting fitter and someone who tests well but doesn’t race well. The gap between lab fitness and race fitness is real. Trends across actual races help you measure it.

Power-to-Weight Ratio

Weight is a sensitive thing to talk about. I get it. But on a hilly bike course or a marathon with real climbing, it matters. A lighter athlete needs fewer watts to hold the same pace uphill. Fewer seconds per kilometer, too. And the steeper the climb, the more that shows up.

The model works this in so your forecast reflects your body against this course, not an average athlete on an average day. It’s not about hitting a number on the scale. It’s about knowing how your body handles what the course is going to ask of you.

Putting It All Together

None of these inputs stand alone. A fast course doesn’t mean much if the field is stacked. A big FTP doesn’t help you on a run-heavy course if your run is lagging. And an AG that’s trending fast doesn’t matter if your thresholds are heading the wrong direction.

The forecast pulls all of it into one picture. It shows you what the data says is realistic, and where the gap between today and your goal actually sits. It won’t tell you what to eat on race morning. And it won’t promise a PR if the weather turns. What it does is trade the guesswork for something you can train against. Something honest. Something specific.

Ramin Pavlovic is a UESCA-Certified Triathlon Coach, Endurance Sports Nutritionist, and USAT Level 1 Coach based in the U.S. He coaches at Predictive Coaching, where the plan starts with your data, not a template.