Most athletes pick a goal time in the winter and then train for six months hoping it shows up in July. The goal sits on a sticky note. The training happens somewhere else, built from whatever plan or group ride is closest at hand, and nobody checks whether the two have anything to do with each other until race morning.
There’s a better order of operations. Start with the course and the time you want, figure out what it physically takes to ride that time on that course, then build the season to close the gap between who you are now and who you need to be in race week. Best Bike Split (BBS) was built to answer the race day question, but every piece you need for the season-long version is already in Best Bike Split.
Here’s how I’d use it as a coach, from the first model in the off-season to the plan on your top tube.
A goal time means nothing until you know what it asks of you. Five hours on a flat gravel 100 and five hours on a 100-miler with 10,000 feet of climbing are wildly different requests, and the only way to know what yours costs is to model it on the real course.
Add your race course first. Search the Best Bike Split database by event name, upload a GPX or FIT file from the event website, import it from Ride with GPS, or draw it yourself with Course Builder if the organizers haven’t published a route yet. For gravel and mountain bike events, turn on Advanced Surface Conditions so pavement, gravel, and dirt each get their own rolling resistance instead of one guess for the whole day. For weather, pull historical conditions for your race date. You won’t have a forecast in January, but you’ll know what the course usually throws at people.
The Goal Time model works backward from the clock. Enter the time you want and it finds the power target and pacing needed to get close to it. It’s also the quickest way to find the minimum power needed to make a bike cutoff, which for a lot of first-time IRONMAN and long gravel athletes is the real goal. For off-road and hilly events, open Advanced Settings and set a minimum Variability Index so the plan reflects how the course actually rides. Best Bike Split suggests starting around 1.07 for hilly road races, 1.15 for a course like Unbound, and 1.2 for Leadville.
The Normalized Power model answers the opposite question. Instead of asking what a time costs, it asks what a given effort buys you. Set your Intensity Factor, add the same minimum VI in Advanced Settings, and see where the finish lands. This is the better starting point when you know what you can sustain but have no idea what a realistic time looks like on a new course.
The Training Stress Score model is the one to use if you have a race you executed well and want to replicate it. If a TSS of 170 at Boulder 70.3 let you run well off the bike, you can apply that same TSS to a different 70.3 and let the model sort out what it means on a new course in new conditions.
Whichever model you use, the output is a race plan with specific power targets attached to your goal. Compare that cost to your current FTP and you have a number that matters for the season: the size of the gap.
Then dig into why the gap is there. Open the Time Analysis tool and start moving sliders. On the Gravel Worlds plan above, pushing the Power slider up 15 percent, from 149 to 171 watts average, takes almost 38 minutes off a 9:25 finish. Drop a little drag instead and compare. Take two kilos off the bike or the rider. On a fast, flat triathlon course you’ll often find that holding a better aero position buys you as much as a full training block of power gains. On a mountainous course, power and weight will dominate, and aero barely registers. That tells you where to spend the winter.
The Advanced Charts fill in the rest. The Gradients chart shows how much of your race happens at each slope, so you know whether you’re training for long 4 percent grades or short punchy walls. The Peak Power chart shows the mean maximal power the plan asks for across durations. Lay it next to your own power duration curve from TrainingPeaks or Strava and the weak spots tend to jump out. Maybe your 20-minute power is fine but the plan needs repeated 3-minute efforts you’ve never trained. Categorized Climbs breaks out each climb with its own power and duration target, which turns a vague goal like “climb better” into a specific effort you can rehearse.
Once you know what the race demands, you can train for it on purpose. The AI Workout Builder generates structured workouts two ways: directly from a race plan, or from a plain language prompt.
Choose “Use a race plan,” select your goal race, and set how long you have to train that day. Add a prompt if you want something specific, like a workout that rehearses the types of climbs you’ll face, or one that spends time in the same zones the race will. The race may be eight hours long. The workout can still be 90 minutes and target what matters most.
The prompt option is for everything else. Type in something like “short steep climbs followed by one long sustained climb, 60 minutes” and it builds a workout that simulates that pattern. Every workout can be updated or duplicated, so you can add or remove steps and adjust durations and targets until it’s exactly what you want, then push it to Garmin, TrainingPeaks, Wahoo, Zwift, or Intervals.icu, or download it as an ERG, ZWO, or MRC file.
For the most specific work, crop the course. Take your full race course, drag across the elevation profile to isolate a key climb and the lead-in to it, and save it as its own course. Build a plan on that section with the same power targets from your full-course plan and you have a realistic rehearsal of the hardest part of your day. Send it to your trainer, ride it in Trainer Studio (in beta) in your browser, or ride it with Indoor Sim in the iPhone app.
A word of caution here, because this tool makes it very easy to go race-specific too early. Develop the physiology most specific to your event closest to the event, and the least specific work furthest from it. In the winter, use what the race plan told you about your weaknesses, and train those. Save the full race simulations and the cropped-climb rehearsals for the final 8 to 12 weeks, when you’re sharpening the strengths you’ll actually race on. The workout builder will happily make you a Leadville simulation in December. Your coach, or your own common sense, should decide whether you need one.
A season plan built in January is a hypothesis. By April you need evidence, and this is where most athletes go wrong. They wait until race day to find out whether the plan worked.
Treat your tune-up races and key long rides as data. Upload the file to Best Bike Split Analytics 2.0 and you get a cycling power analysis of how you actually rode that goes well beyond average power.
Start with drag. Analytics 2.0 estimates your CdA (aerodynamic drag area) throughout the ride and groups it into clusters for race, climbing, and relaxed positions. If the numbers look accurate, you can apply them to your bike setup in one click, which makes every future race plan more honest. If your aero work over the winter paid off, it shows up here, and so does every time you sat up when the road tilted.
Then look at how you rode the course itself. The Cornering Summary groups every corner from gentle bends to hairpins and shows your entry, apex, and exit speeds, how hard you braked, and the power you put down on the exit. Ride Dynamics counts your braking, coasting, and acceleration events and the total time spent in each. On a technical gravel or mountain bike course, a lot of free time hides in those numbers. An athlete losing 30 seconds braking into every switchback would get more out of a skills session than another threshold block.
The Gradients chart shows your actual speed and power on each slope range, which can expose weak descending or climbing that falls apart on steeper pitches. The Peak Power chart compares your mean maximal power to equivalent Normalized Power across durations, and helps you set realistic power limits and minimum VI values for the next plan.
Then run Time Analysis on the ride you just analyzed. How much would 15 more watts have changed the result? What about a tighter position? This is the same tool you used in the off-season, only now it’s working on what actually happened instead of what you predicted.
Finally, open the Comparison Tool and put your actual ride next to the plan you built for it. Where you rode harder than planned, and what it cost you later, tells you more about your pacing discipline than any number on a training dashboard.
Take everything you learned back to Step 1. Update your FTP, your weight, and your CdA. Rerun the goal plan. The gap might be smaller than you thought. It might be bigger. Either way, you’re making the next training block decision based on what you can do now instead of what you hoped you could do in January. And if the ride turned up something specific, you can ask the BBS AI Assistant something like “What does this ride tell me about my cornering?”
By race week you’ve done the hard part. The final plan just pulls everything together.
Update your athlete profile and bike one last time so the plan reflects the athlete who’s actually showing up. Then update the race plan and pull in the latest forecast. Weather data gets more accurate as the event gets closer, and it’s specific to where you’ll be on the course and when you’ll be there. A headwind on the back half of the course changes your pacing strategy and where you should spend your watts.
For the final plan, rerun whichever model fits, using what your tune-up races told you rather than what you assumed in the winter. If you raced a similar event well this season, the TSS model lets you carry that effort straight into the new course.
Then get the plan out of the browser and onto your bike. Send it to Garmin as a Power Course or structured workout, sync it to Wahoo, or follow it on the Hammerhead Karoo app. Print the cheat sheets too. They break the course into power and speed targets by grade and wind, the same idea as the cue sheets pro time trialists used to tape to their top tubes, and they work when your head unit doesn’t.
Use the Notes tab for the things the model can’t hold: your fueling schedule, where the aid stations are, the descent you need to ride cautiously, what you’re going to do if the front group goes early. The BBS AI Assistant can help you build a nutrition plan around your expected time and the forecast conditions.
And then race. The plan is a guide built on good physics and your best data, but it’s still an estimate. Wind shifts, groups form, legs show up good or bad. Ride the plan when it’s right and trust what your body is telling you when it isn’t. The power numbers are there to keep you honest in the first hour, when everything feels easy and the expensive mistakes get made.
When it’s over, upload the file to Analytics 2.0, put it next to your plan in the Comparison Tool, and look hard at the first spot where the two lines split apart.
Start your season plan with a free Best Bike Split account. Build your first race plan today and see what your goal actually costs.
Dave Schell is the Founder and Head Coach of Kaizen Endurance coaching based out of Boulder, Colorado, where he works with off-road cyclists of all abilities. Dave is the former director of coach education at TrainingPeaks where he traveled the world teaching coaches how to better use training software with their athletes.
Read more about Dave Schell at Kaizen Endurance.
Enter your FTP, weight, bike specs, and aero data to generate precise race predictions. Don't know your CdA? Our system can estimate it from your position and equipment or from a past ride so you can start racing with confidence.
Choose from thousands of existing courses including most Ironman, 70.3 and road races, or upload your own GPX file. Our database includes detailed elevation profiles and road surfaces, plus historical or forecasted weather for accurate predictions.
Receive a detailed, segment-by-segment bike pacing strategy with variable power targets optimized for every section of your course. See your predicted bike split, IF, TSS and exactly how your pacing strategy balances speed with power.
Download your plan to Zwift, TrainerRoad, or any ERG-mode trainer for indoor training. Export to Garmin, Karoo or Wahoo for outdoor training rides and race-day execution. Practice makes perfect — run your exact race plan in training before the big day.
No credit card required. Upgrade anytime. Cancel anytime.
©2026 Caffeinated Catalyst. ALL RIGHTS RESERVED. |
Terms of Use | Privacy Policy
Normalized Power®, Training Stress Score® and Intensity Factor® are registered trademarks of TrainingPeaks, LLC.