The end of the season is the best time to get honest with yourself about your training. Not the "how did it feel" kind of honest, the "what does the data actually say" kind. I do this every year with my own power data, and every year it tells me at least one thing I would have gotten wrong from memory alone.
The good news is you no longer have to build these charts by hand. CoachCat already has every ride you have uploaded, your Level history, your power PRs, your FTP changes, and your OTS. You just have to ask it the right questions. Below are the five pieces of data I look at, the exact prompts I use to pull each one, and what my own numbers looked like when I ran them.
Key Takeaways
- See below for 6 Prompts you can copy and paste into CoachCat's chat to review your season.
- Zoom out in steps. Last month, then 3 months, then 6 months, then the full year. Trends that hide at one zoom level jump out at another.
- Translate your Level into weekly hours. That is how you actually plan training around life.
- Compare your power PRs to a fixed reference point like January, not to your best day ever.
- Copy the prompts below into CoachCat and it builds every one of these charts from your own files.
CoachCat Member Story
Can I just tell you that I followed your instructions/prompts for the season in review and I am blown away at the concise data review, insights and recommendations. I have never had such a thorough reflection of my training. Wish I had it 20 years ago!
Jim P.
CoachCat member since 2024
The 5 Pieces of Data to Look At
You are looking for improvement you can carry into 2027. Not a scorecard, not a grade. Five things, in this order:
- Weekly totals in hours and OTS, zoomed out 1 month, then 3 months, then 6 months, then a full year
- Your Level
- Your power PRs
- FTP tracking
- VO2 max power output tracking
Each one answers a different question. Hours and OTS tell you how much work you did. Level tells you what that work added up to. Power PRs, FTP, and VO2 max power tell you where the work actually landed, which is the part most riders guess at.
1. Weekly Totals in Hours and OTS
Start with how much you rode, because that is the most important part that everything else depends on. Look at your weekly hours and your OTS at four zoom levels: the last month, then 3 months, then 6 months, then the full year. Same data, four different stories.
At one month you see your average weekly hours. At three months you see whether your last training block was actually a block or just a collection of rides. At six months you see the shape of your season. At a full year you see the thing nobody notices from inside it, which is how many weeks you lost and where they went.
What you are looking for is the gap between your biggest weeks and your smallest ones. For most riders the big weeks are already big enough. The damage is done by the small ones, the 5 or 6 weeks a year that fall apart because of travel, work, illness, or weather. Those weeks drag your season average down further than one extra big week can pull it up. So the fix is almost never a bigger peak week. It is fewer collapsed ones.
Open the Progress tab in the bottom navigation, then tap 1M, 3M, 6M and 1Y. Same chart, four zoom levels.
CoachCat
Aug 16, 8:48 PM
Yes, the big picture from your weekly hours + OTS is:
- You improve best from repeatable load, not monster spikes. Your strongest gains came when you stacked several weeks of solid riding instead of doing one huge week and then backing off hard.
- Your sweet spot is probably around 8 to 11 hours a week most of the year. That's enough to drive real fitness gains for you without blowing up recovery or life.
- 12+ hour weeks should be used as short blocks, not a default. You can handle them, but they look more like peak building weeks than your everyday training floor.
What I'd do differently for 2027
- Raise your weekly floor. Try not to let normal weeks drop too low. For you, consistent 7.5 to 8.5 hour weeks will likely beat a pattern of 5 hour weeks plus occasional 14 hour weeks.
- Use a 3 up, 1 down rhythm. Three progressive weeks, then one lighter week. That matches how your Level and power have responded best.
- Protect the key sessions, trim the filler. Your best load comes from keeping the quality rides and the long ride intact, not from adding extra random volume.
- Keep strength in, but controlled. Your lifting has been helpful, but it should stay a support piece, not compete with bike load. In 2027, I'd keep it as maintenance except during early base.
Bottom line
For 2027, I'd build you around steady 8 to 11 hour weeks with controlled OTS, and only use big load weeks in deliberate blocks. That's the pattern that should keep your Level climbing without creating recovery crashes.
That is my CoachCat, reading my rides. Run the same prompt against your files and you get your numbers, not mine.
Look at the first thing on its list. A few paragraphs above, I had already written that the small weeks cost you more than the big weeks gain you. Then I ran the prompt, and CoachCat said the same thing back to me about my own rides, without me hinting at it. That is what makes this worth half an hour: it is not a general opinion about training, it is my own season telling me where the time went. Hold onto the 8 to 11 hour range too, because it turns up again in the next section from a completely different direction.
2. Your Level
Your Level is the single number that says how much fitness you are carrying. Hours tell you what you put in. Level tells you what came out the other side.
By itself, though, a Level number is hard to plan around. What makes it useful is translating it back into weekly hours, because hours are what you actually negotiate with your job, your family, and your sleep. So ask CoachCat for both.
Here is what came back for me. Your bands will be different, but the structure of the answer is the same:
How I would use that for 2027 planning
- If life is busy, hold 8 to 9 hours a week consistently. That is a strong minimum effective dose.
- If you want real growth in 2027, build toward 9 to 11 hours a week for your main build blocks.
- Use 11 to 13 hours a week only when your schedule, sleep, and recovery are genuinely good.
- Your Level rises when you stack several weeks in that 8 to 11 hour range. It does not rise from one monster weekend.
That last point is the one worth tattooing somewhere. The big Saturday feels like the training. The eight ordinary weeks around it are the training.
Do this before you plan a single week of 2027
A season review takes about 30 minutes and it changes what you build next. Pull your hours, your Level, your power PRs, your FTP, and your VO2 max power, and let the numbers tell you where the improvement actually came from. Then build the 2027 plan around that instead of around what you assume happened.
Review your season in CoachCat3. Your Power PRs
This is the fun one, and it is also where most riders get the wrong answer by looking at the wrong comparison. Do not compare today to your best day ever. Compare today to a fixed reference point, like January, so you are measuring your improvement from the beginning of the season rather than your all time bests.
Ask for the full range of durations, not just the famous ones. Five seconds through three hours, with dates attached. The dates matter as much as the watts, because they tell you when your form actually showed up.
Mine came back like this, January best against my best since January:
Two things jump out of that table, and neither one would have occurred to me from memory.
First, look at 20 minutes. Up 6 watts, while 30 minutes went up 20, an hour went up 39, and 90 minutes through 3 hours all went up more than 40. That is not a fitness problem, that is a testing artifact. My January 20 minute number came from an actual field test, a maximal, dedicated effort. My August 20 minute number came from a group ride. I never went back and tested properly, so the 20 minute number is the only one on the list that is not a fair fight. Lesson for 2027: put real tests on the calendar.
Second, look at the dates. Almost every long duration power PR landed on the same day, August 15. That is not a coincidence, that is what peak form looks like on a chart. It also tells me my form arrived in August, which is useful information if my A race was in June.
And the sprint numbers, 5 seconds up 160 watts and 1 minute up 172 watts, came from a season with real group rides in it. Neuromuscular power responds fast to being used, and it disappears fast when it is not.
One more thing this screen gives away. I looked at it set to 1Y, and then at 6M and 3M, and the five numbers never changed. Every personal record I hold was set in the last three months. Widening the window to a full year adds nothing, which is the clearest sign I have that the season worked.
Power PRs and the power curve on the Progress tab, set to 1Y. Every one of these is flagged NEW.
4. FTP Tracking
FTP is the number everyone quotes, so it deserves a chart of its own. What you want is not just where it is now, but the whole path it took to get there, with the dips included. The dips are the interesting part.
FTP chart
| Date | FTP | Change | Notes |
|---|---|---|---|
| Jan baseline | 260 W | Threshold at the start of the year | |
| Jan 7 | 264 W | +4 W | First bump of the year |
| Jan 10 | 284 W | +20 W | Big early season jump |
| Mar 18 | 280 W | -4 W | Small pullback, not unusual |
| Jun 16 | 284 W | +4 W | Rebound back to the prior high |
| Today | 295 W | +11 W | Current profile threshold |
From 260 W to 295 W, a gain of 35 watts since January, about 13.5%. The pattern is fast early gains, one small dip, then a steady rebuild to a new high.
Thirty five watts over the season, but notice how it arrived. Twenty four of those watts showed up in the first ten days of January, then nothing for two months, then a small step back, then the rest came in the second half of the year. If someone asked me in March how my season was going I would have said it had stalled. It had not. It was consolidating.
The March pullback is worth sitting with. A 4 watt drop is noise, not a crisis, and the temptation to react to it by training harder is exactly how riders wreck a spring. CoachCat treats your FTP as a tracked number that moves with your training rather than a score to defend.
5. VO2 Max Power Output Tracking
The fifth one is where you get to be curious. Pick a duration that matters to your riding and track it the same way you just tracked FTP. For me that is 5 minute power, which is the practical proxy for VO2 max. If you are a criterium racer, do it with 1 minute. If you ride long gravel, do it with 2 or 3 hours.
VO2 Max Power progression since January
| Date | 5 min power | Change vs prior best | Notes |
|---|---|---|---|
| Jan 7 | 304 W | First big marker of the year | |
| Jan 10 | 308 W | +4 W | Early jump |
| Jan 11 | 312 W | +4 W | Kept moving up fast |
| Mar 5 | 317 W | +5 W | Solid spring progress |
| Apr 14 | 323 W | +6 W | Nice step forward |
| Apr 28 | 338 W | +15 W | Big breakthrough |
| May 12 | 337 W | Basically held that level | |
| Jun 16 | 355 W | +17 W | Another big jump |
| Jul 22 | 339 W | Not a new peak, but still strong | |
| Aug 1 | 351 W | Back near the high end | |
| Aug 15 | 363 W | +8 W | New best VO2 marker |
5 minute power went from 304 W to 363 W, a gain of 59 watts, in three distinct jumps with flat stretches in between.
Fifty nine watts on 5 minute power, and it did not come in a straight line. It came in three distinct jumps, late April, mid June, and mid August, with flat stretches in between. Those flat stretches are not failures. They are what the body does while it is absorbing the last jump.
Compare this table to the FTP one and something useful appears. My 5 minute power gained 59 watts while my FTP gained 35. My aerobic ceiling moved more than my threshold did. That points straight at what to build in 2027: more sweet spot work to pull threshold up under a ceiling that clearly has room.
Turning the Review Into a 2027 Plan
Five prompts, plus a bonus one, and now you have to decide what to do with the answers. Here is the order I use.
Start with the floor, not the ceiling. Your hours chart shows you the weeks you lost. Fixing five bad weeks is worth more than adding one big one.
Pick a Level target and translate it to hours. If you want to sit at 61 to 65, that is 8 to 11 hours a week for me. Know the number before you build the calendar, because that is the number that has to survive contact with real life.
Find the duration that did not move. In my power PR table, 20 minutes barely budged, and it turned out to be a testing problem rather than a fitness problem. Yours might be a real gap. Either way, that is your 2027 focus.
Compare FTP growth to VO2 growth. If your ceiling moved more than your threshold, build threshold. If your threshold moved more than your ceiling, you need intensity. Mine says build threshold.
Then go to the 30 Week Weights + Sweet Spot Base plan and actually do it. The review is worthless if November looks like every other November.
Why CoachCat Knows This and ChatGPT Does Not
You could paste a spreadsheet into a general purpose chatbot and get a chart back. What you would not get is context. A general model is trained on the open internet, so what it hands you is the average of every opinion ever posted about cycling training, and plenty of those opinions are confident nonsense. It does not know your Level history, it does not know that your August 20 minute number came from a group ride while your January one came from a test, and it has never coached anyone.
The real difference is the dataset. ChatGPT has no proprietary training data to work from and no coach has ever taught it anything. CoachCat was trained on mine: 20 years of my own coaching, the workouts I have prescribed, the mistakes I have watched athletes make over and over, and the calls I have made when an athlete's numbers said one thing and their life said another. That knowledge was never posted anywhere for a model to scrape. It came from coaching real cyclists, one at a time, for two decades.
Then your own files go on top of it, so the advice is specific twice over. CoachCat knows how to coach, and it knows how you train. When it tells you to hold 8 to 9 hours a week, that number came from your training history run through 20 years of coaching judgment, not from a forum post. That is the difference between a chart and a coach.
Bonus Prompt: What Kind of Rider Are You?
One more worth running while you have your season in front of you. Your power curve says something about what kind of rider you are, and CoachCat will read it back to you.
Sprinter, all rounder, time trialist, climber. It comes from the shape of your curve, not from what you like to ride.
It is a useful gut check going into 2027. If the answer surprises you, that is worth knowing before you pick next season's goals, because the events that suit your curve are the ones you will do well at without fighting your own physiology the whole way.
Five prompts, one honest look at your year
Copy the five prompts above into CoachCat and you will have your whole season in front of you in about half an hour. Hours and OTS, Level, power PRs, FTP, and VO2 max power, all built from your own rides. Then take what you find into an off season plan that fixes the right thing.
See the off season plansReflecting on what worked and what did not is the part of training nobody posts about, and it is the part that compounds. Do the review, write down the two or three things the data actually told you, and build 2027 around those instead of around a feeling. Then go take some time off the bike. You earned it.
Frank Overton is the Founder and Head Coach of FasCat Coaching in Boulder, Colorado, and has been coaching cyclists with power data since 2002. To talk through your own season review and your 2027 plan, fill out a New Athlete Questionnaire to schedule a coaching consultation, or start your free week of CoachCat.
