Anyone have any science or personal experience in optimal pacing strategy for a 3 minute (+/- 15 seconds) PR attempt on a hilly segment (avg 5%)?
Being so short, I was wondering if there might be a better pacing strategy than slight negative split which is what best bike split suggests. And taking into consideration gradient changes.
How long can you hold 120%, 130% ,140%, 150% of your ftp? This should give you some idea of pacing strategies.
Good call! I’ll take a look.
Gradient changes, well go harder when its steep and ease off a tiny when less steep. Other than that, look at your best 3 min power and use that as the baseline to float around.
What is this “pacing” thing you’re talking about?
That video is super helpful. Good to keep something in the tank for the final 30 seconds.
I went out too hard when setting a baseline and went about 20% lower power on second half vs first. The opposite of negative split :). Always too ambitious.
The grade on my hill is only 5% so speed is actually pretty high (17/18 mph or 28/29kph) so probably shouldn’t get out of the saddle at all until maybe the last kick.
Aiming for 3 min flat but did it in 3:09 so have some work to do!
Good tips! I should have looked at my power earlier on but the beginning is quick with a couple turns so didn’t realize how over my target I was until I was blowing up at the top. I also did it pretty cold (5 min ride over) so less perception of effort.
This is a use case in wich the concept of critical power by Monod & Sherrer is really useful. You can get a really close estimate of what power you are able to hold for a certain amount of time. A slightly less precise alternative is to use your power/duration curve to find your target power.
The most important thing to consider is to not overcook in the first minute although a slightly higher power than your target average power would probably be ok. In my experience it is usually possible to dig a little deeper the last 10-15 seconds, depleting the remaining anaerobic capacity, if necessary.
It also depends on the topography of the hill. One ~4-minute hill I used to do hill repeats on would have two small dips, so you would not want to pace evenly.
I’m going to try this hill segment again to try and average 500 watts for 3 min and did some ChatGPT research on the lead-in and saw this that was def counter to what I thought. In my head I figured if I wanted to average 500 watts for 3 min I would need to be efficient about the effort pre-start (maximize speed into the segment but keep effort low) but this is a very different take. Anybody done something like this with a harder lead-in? It suggests a 30 seconds 400-450 lead-in will result in better overall performance than just a 500 watt start and hold for 3 min. Is this a hallucination or legit?
Pre-Segment Pacing Plan (30s Roll-In)
t-30s to t-20s
Target: 380–400 W (~105–110% FTP)
Purpose: Start ramping VOâ‚‚ demand without spiking lactate.
Cadence: Smoothly build toward your target (90–95 rpm).
t-20s to t-10s
Target: 400–415 W
Purpose: Hold steady, keep breathing controlled, stay smooth.
t-10s to Start Line
Target: 430–450 W (~120–125% FTP)
Purpose: Lock into “VO₂ gear” so you hit the line already primed.
Focus: Relax upper body, steady spin, don’t sprint — this isn’t about burning a match.
Shockingly ChatGPT has no idea what it’s on about. Those “purpose” and “focus” bits make absolutely zero sense.
This is a good example why LLMs should not be taken at face value, you need to know enough to identify bs.
When you write segment, it leads me to believe you will ride there. Plenty of time to get warm. You don’t need to do anything then, provided the ride to the segment is long enough. I’d also make sure to go above threshold for a bit, but nothing extreme is necessary. It is just to open the taps.
Ok, yes, I realize I should have just thought about it for a minute. I’ve been using TR plus an LLM for recent training goals and it has really helped from a day to day basis but still prone to errors. If the above were correct we could just do all of our interval training with an added VO2 max lead-in and be just as successful ;).
Pacing is maybe not needed on a 3 minute hill. I recall seeing Dan Lloyd’s GCN video prior to the TdF where he took a KOM on the course, ahead of the race.
He worked on a positive split strategy, which is just a fancy way of saying, go all out and hang on for as long as possible.
Either way, it worked for him and his average power was impressively high for the effort.
I often think that there is more in the tank than you really realise. I was out testing a (segments) race course today and managed to take a KOM on a fairly long, gnarly gravel climb by a single second. It took 6 minutes 30 at 507w and this is what I looked like afterwards. I can confidently say that I didn’t leave much in the tank.
That is impressive power! Very nicely done! One has to endure some real pain for a 6.5 min KOM. My 3 minute effort will seem like no big deal.
Everyone has forgotten the most important part: fast friends to paceline you for a final all out 30 second effort 
That’s what sucks about that LLM crap. If you don’t know there is an error, it leads you down the wrong path. And it always has errors. If you know there is an error, you didn’t really need it to begin with.
One other thing. When in the last 75-45 seconds or whatever, if you feel like it’s all turning to shit, your body seems like its all over the bike, pedalling squares etc, don’t worry because your power and relative speed will still be around the mark. So don’t give up on the effort, or tell yourself it’s not gonna be worth it just because it feels like you’re falling apart. Its just a new level of unfamiliar sensations.
Or perhaps none of this tends to be factor for you.
3m? go all out and hang on. Unless its an irregular hill. For example on the Cauberg (3m03) i started a little under and went all out on the steep part and then just emptied the tank on the false flat bit on the top
