Picking Zwift Routes That Match the Session You Meant to Ride

Every Zwift ride begins with the same decision, and I make it in the few seconds it takes to scroll the route list. I ride Zwift several times a week and I am training for Ironman Copenhagen, so a large part of my training week is settled in that scroll, and for a long time I got it wrong without noticing. The session I had written down and the session the route actually gave me were different things. I am a mechanical engineering student in Aalborg, not a coach and not a sports scientist, so nothing below rests on research I cannot show you. It is mechanism, arithmetic from this site's own climb model, and my own riding, each labelled as such.

Gradient decides whether your power turns into speed

On a climb you are lifting mass against gravity, and gravity does not care about the pack. On a flat road you are pushing air, and air is something other riders can move for you. That single difference is why the same interval works on one route and falls apart on another.

Here is the split for a 75 kg, 183 cm rider carrying 9 kg of bike and kit at 3.0 W/kg, taken from the physics model behind the climb time calculator rather than from anything measured. It is a solo estimate with no draft, which is exactly the point. The two columns do not add up to 100, because the remainder, between 4% and 10% here depending on the gradient, goes into rolling resistance.

Read the last column as a measure of how much the route hands your result to other people. At 8.5% almost nothing you do is aerodynamic, so your time is your sustained power and your weight, and a soft thirty seconds shows up immediately as lost speed. At 2.2% roughly a third of your effort is fighting air that a wheel in front would clear for you, so the pack, the surges and your timing decide the outcome.

The practical rule falls straight out of that. If the session is about holding a specific power, choose gradient, because gradient forces honesty. If the session is about racing, choose the shallow rolling route, because that is where positioning and the draft are the skills being tested. Length, gain and published average gradient for every timed segment in the game are listed on the Zwift climbs page, which is the quickest way to find the profile you need.

Duration decides which session the route can be

A climb is not long or short in kilometres. It is long or short in minutes, and the minutes depend on you. A 4 km segment is a five-minute effort for one rider and a twenty-minute one for another, and those are different training sessions with different physiology behind them. So before you commit, convert the route into time at the power you plan to hold.

The times below are model estimates for the same rider as above, 75 kg and 183 cm with 9 kg of bike and kit, riding solo at 3.0 W/kg, and they come from the same equation the calculator runs. They are arithmetic, not measurements, and your own numbers will differ. Put your weight and your power into the climb time calculator and you get your version of this table in a few seconds.

Notice how quickly the same category of thing becomes a different session. Leith Hill and The Grade are both "a climb". One is eight minutes and the other is nineteen, and no amount of good intent turns the first into a threshold session or the second into a set of short intervals.

Matching the route to the session you meant to ride

Once duration and gradient are the two axes, the framework writes itself. Decide what the session has to do, then look for a route that supplies it rather than one that merely permits it.

Endurance rides

What the route must supply is the absence of provocation. Flat or gently rolling, no timed segment long enough to tempt you, and preferably somewhere you have ridden before so there is no reason to push. The classic failure is picking a scenic route with a climb in the middle, riding the climb because it is there, and turning an aerobic ride into a middling tempo one that is too hard to recover from and too easy to adapt to.

Sustained threshold work

The route must supply a continuous climb at least as long as the interval you intend to hold. If the plan says twenty minutes, a segment that ends in twelve leaves you finishing the effort on a descent where you cannot hold the power at all. At 3.0 W/kg the model puts Keith Hill at about 15 minutes, The Grade at about 19 and Epic KOM Reverse at about 26, so a rider around that level has a genuine choice of venue. Check your own figure first, because a stronger rider needs a longer climb to get the same session.

Short, hard efforts

Here the route must supply steepness more than length, because you want the effort to end when your legs end it rather than when the gradient runs out. Radio Tower KOM is 1.16 km at 12.8%, which the model puts at about 6:46 for a 75 kg rider at 4 W/kg and about 5:27 at 5 W/kg. That is a real five to seven minute effort in a very small piece of road, and you can be back at the bottom quickly for the next one.

Racing and group riding

Now the shallow rolling route is the right answer, for the reason the first table gives: a large share of the effort is aerodynamic, so the draft, the positioning and the timing of a surge are what decide the result. Practising those on a 9% wall teaches you nothing about them. If that is new ground, the racing for beginners guide and the drafting article cover what the shallow routes are actually testing.

A worked example: is the Alpe the session you planned?

This site fitted a curve to roughly 500 verified Alpe du Zwift finishing times from ZwiftPower, each one a power figure paired with a completion time. The fit is time_seconds = 148.60*(W/kg)² - 1954.08*(W/kg) + 8329.87, and the methodology article explains how it was built and where it gets shaky. Put numbers through it and the route becomes a duration:

• 2.8 W/kg: about 1:07:03
• 3.0 W/kg: about 1:03:25
• 3.2 W/kg: about 59:58, which is where the hour falls
• 3.5 W/kg: about 55:11
• 4.0 W/kg: about 48:11

Now hold that against a plan that says "45 minutes at threshold". At 3.0 W/kg the Alpe is an hour and three minutes of continuous climbing, so it is not that session and cannot be made into it. You have two honest options. Ride it deliberately below threshold and accept a long steady day, which is a good session in its own right. Or treat it as a benchmark, ride it as hard as you can, and plan the day after around the fact that you will be tired. What you should not do is start at threshold pace and discover the mismatch at hairpin fourteen, which is the version I have ridden more than once.

The same arithmetic applies to Ven-Top. The timed Ventoux KOM segment is about 19 km and 1,481 m of gain, and the full Ven-Top route is about 20.9 km and 1,534 m, so be clear which one you are planning. The site's Ven-Top fit, built from the same kind of data and roughly 300 verified ZwiftPower times, is time_minutes = 3.21 + 253.38/(W/kg), which puts 3.0 W/kg at about an hour and twenty-eight minutes. That sits a few minutes under the generic physics estimate in the table above, which is what you should expect: one model is fitted to actual Ven-Top results and the other is physics applied to any climb. Either way the answer to "can I fit this in tonight" is the same. Set the target with the Alpe du Zwift calculator or the Ven-Top calculator, and compare the two efforts side by side on Alpe versus Ven-Top.

Why we choose badly anyway

The arithmetic is easy. Following it is not, because three pulls act on the route screen and none of them care what your plan says.

Familiarity is cheap, and that is the problem

A route you know well costs you nothing to plan. You know where it gets hard, roughly how long it takes, and where you can ease off, so you can set a pacing plan before you clip in. That is a genuine advantage, and it is why benchmark efforts should be ridden on routes you know. The cost is that you only ever train what those routes ask for. I have three defaults I fall back to when I am tired, and all three are rolling. Left alone, my week would contain no sustained climbing at all, which for someone with an Ironman in the calendar is the wrong hole to have.

The fix is not novelty for its own sake. It is to notice what your defaults have in common and pick deliberately against that. If everything you ride is under twenty minutes of climbing, the missing session is a long one.

The badge and the unlock

Zwift's route badges, level unlocks and drops are designed to make you keep riding, and they work. I own the Tron bike, which means I spent a long stretch of my Zwift life choosing routes purely by how much elevation they contained, with no reference to any training plan whatsoever. I do not regret it and I would not call it training either.

The honest position is that a reward you enjoy chasing is a reason to ride at all, which beats a perfectly structured session you skip. The problem starts when badge chasing quietly replaces the plan rather than sitting beside it. Give it a slot, ride the unlock session on a day where the effort roughly fits, and check the route's duration first so you know what you are signing up for. The achievements guide lists what is worth chasing.

The event calendar chooses for you

Group rides and races are the easiest way to end up on a route you did not select. That is not automatically bad, since other riders make an easy ride easier to finish and a hard ride harder to quit. It is bad when the event's route contradicts the session. A rolling race on a day scheduled for endurance is a hard day with an endurance label on it. Check the route before you sign up, not after the pen closes, and read the group ride dynamics piece for how the pack changes your effort whether you intended it or not.

Three ways this goes wrong mid-block, and the fix for each

Picking something you cannot finish

Abandoning a route halfway is discouraging out of proportion to the training actually lost, and I know what that does to me: I avoid anything that might do it again, and a month of too-easy riding goes by. The fix is arithmetic instead of courage. Take your longest recent sustained effort, add a few minutes, and find a climb whose estimated time at a power you can genuinely hold lands there. You then start knowing the number is achievable, which is a completely different mental state from hoping.

Letting the clock pick the route

Forty minutes free and a route that takes an hour is a real constraint, not an excuse. But the usual response, picking a short flat route and pedalling gently, wastes the slot. A short window is the right time for the steep short segments: Radio Tower or Leith Hill repeats give you a genuinely hard session inside half an hour of riding. Save the long routes for the days that can hold them, and check the estimate rather than guessing, because guessing is how a session gets abandoned at the 40-minute mark.

Avoiding the route where you had a bad day

There is probably a route you have quietly stopped selecting because of one grim time on it. I have two. That is exactly the route worth going back to, because a repeated effort on known terrain is the cleanest fitness comparison you have: same gradient, same length, same model prediction, only your power changed. A single slow attempt is a baseline, and a baseline is only useless if you never repeat it.

How I lay out a week of routes

This is my own template rather than a prescription, and it is not the output of any study. It exists to make sure that the sessions I would otherwise skip get a slot before the week fills up with whatever looks good on the day.

Most of the week goes to routes chosen to serve the plan, a smaller share to something harder than I am comfortable with, and a little to something I have never ridden. I do not defend those proportions with data, and there is none to defend them with. They exist because a week with no hard route stops producing progress and a week with nothing new stops being enjoyable, and both of those failures end training blocks.

Fitting that into an actual training block is a separate problem, and I wrote about it in training plans 101 and the 10-day training week. The rest of the training section covers periodisation.

What is worth tracking, and what this site's data cannot tell you

Four things about your own riding are cheap to check and actually inform the next choice:

• Segment time against the model. Ride a climb, compare your time to the estimate at the power you held. A large gap in either direction usually means your pacing or your power estimate is off, not the route.
• Repeat times on one climb. The same segment ridden three months apart is the least noisy fitness signal you can get without a lab.
• Time in zone per route. If a route you call endurance keeps producing threshold time, the route is choosing the session, not you.
• Rides you abandoned or changed halfway. The pattern in those tells you which routes you consistently overestimate yourself on.

What I cannot give you is a number for how much any of this improves adherence, and I want to be direct about why. The dataset behind this site is roughly 500 Alpe du Zwift and 300 Ven-Top verified ZwiftPower finishing times. Each record is a power figure paired with a completion time, and that is all. There are no timestamps, no training histories, no route-selection logs and no record of who kept riding and who stopped. Any claim on this page about motivation or consistency is therefore reasoning, offered as reasoning, and the methodology article shows exactly what the data does cover.

If you want the measurable part of route choice, it is all in the first two tables: duration and gradient. Those are published for every timed segment, they are the inputs the model uses, and they are enough to make the decision well.