Zwift TT Tests

In an attempt to understand Zwift's drag calculation I ran some test on Tempus Fugit.

The timed section was 2km starting at 0.5km from the inception point and finishing at the end of the sprint. I chose this because it made it easy to log in/out whilst changing weight and height. It also has the benefit of not being completely flat but rolling slightly.

The runs were all pretty much the same, ramp up watts then start timing at 0.5km. I used manual timing for the test but those below come from a Strava segment that I setup so should be accurate. 


I tested 2 setups. 

The first was a warmup with the best TTT bike (Venge + Super 9).  The "down" runs were towards the sprint and have more drafting benefit because more riders go that way. The "up" are the reverse.  No surprise here, more draft = faster speed.

The last Venge run was at 325W the target for the test. I chose this because it should be fast enough to highlight any trends and it's the ballpark target for a TTT..

The second setup was a Zwift TT frame with Classic wheels. This setup is slightly slower than the Venge can manage (with some drafting assistance)

I also did one run with the best possible TT bike just out of interest.

The final "polarised" run was just to see if going hard when the gradient was positive and easing back when negative made any difference. It did, it's worse.

Most weight/heights were extremes just to test but a couple are for me and Ian. (Zpower seems a bit bugged as it does not show height for some riders so could not do more)

Results below. There is not enough data to work out Zwifts algorithm but there is enough to see that weight clearly does have a bigger effect than I thought even on a flat course. 

There may not be enough to work out what's going on inside Zwift but there is IMO enough to set a clear target for the lead rider and a simple basic strategy.

That is, the lead rider should target holding a steady 5.0 w/kg. I

Shifts are done using the main clock which is ofc the same for all of us, doing transition in last 10 seconds of every minute.

• Lead rider does 1 minute at 5.0w/kg
• Second rider starts slowly accelerating to 5.0 w/kg when the clock passes mm:50  so that they are level with the lead rider by mm:58
• All others pick up pace but they will not need to up their pace by as much due to drafting.
• Lead rider drops to back.
• For all riders this should mean 1 minute at or above FTP but  nowhere near red-line and 4-7 minutes under depending on the number of riders. This should be enough recovery time to allow all or at least most to get to finish.
• It should also make it as easy as possible to ride in a pace-line rather than a blob which means more recovery for all those drafting.

It will also be easy to change on the fly. If things are going well then the target can just go up say to 5.2 W/kg, if not it can come down.  The Tick Tock course is especially good for this, at the end of of the short desert climbs there is time on the descent to up or reduce the pace.


This is not the fastest possible way to ride but it will be a good baseline and we can judge how to improve it once we have seen how it goes. 

Test run results (Venge in order they were done, TT in speed order, fastest first)

Run	Lap	Frame	Wheels	Kg	Cm	Watts	Time	Dir	W/kg	AVS
1	3	Venge	Best	68	170	159	03:29	Down	2.3	34.45
2	6	Venge	Best	68	170	168	03:32	Up	2.5	33.96
3	8	Venge	Best	68	170	210	03:14	Down	3.1	37.11
4	10	Venge	Best	68	170	210	03:17	Up	3.1	36.55
5	12	Venge	Best	68	170	327	02:38	Down	4.8	45.57
8	18	Zwift TT	Worst	50	150	328	02:29		6.6	48.32
9	20	Zwift TT	Worst	50	170	324	02:33		6.5	47.06
10	22	Zwift TT	Worst	50	200	330	02:38		6.6	45.57
Best TT	30	Felt	Best	68	170	330	02:38		4.9	45.57
7	16	Zwift TT	Worst	68	150	325	02:39		4.8	45.28
11	24	Zwift TT	Worst	60	190	329	02:42		5.5	44.44
Me	14	Zwift TT	Worst	68	170	326	02:43		4.8	44.17
12	26	Zwift TT	Worst	70	180	344	02:43		4.9	44.17
Ian	28	Zwift TT	Worst	68	177	333	02:43		4.9	44.17
Polarised	32	Zwift TT	Worst	68	170	337	02:43		5.0	44.17