Free Handy Tools

Race Split Planner

Detail level
Try an example
km

h:mm:ss

Average pace

4:59 /km

43 splits to the finish

First half1:45:00
Second half1:45:00
Halves differ by0:00
How this was worked out
  • Race distance in metres42.195 km × 1,000 = 42,195 m
  • Target time in seconds3:30:00 = 12,600 s
  • Average pace = time ÷ distance12,600 s ÷ 42.195 = 298.6 s = 4:59 /km
  • How the pace is spreadEven: every split is planned at the average pace, so both halves are 1:45:00
  • The part-split at the end42.195 km leaves 0.195 km after 42 whole splits, worth 0:58
Splits for 42.195 km in 3:30:00
AtSplitPaceElapsed
1 km4:594:594:59
2 km4:594:599:57
3 km4:594:5914:56
4 km4:594:5919:54
5 km4:594:5924:53
6 km4:594:5929:52
7 km4:594:5934:50
8 km4:594:5939:49
9 km4:594:5944:48
10 km4:594:5949:46
11 km4:594:5954:45
12 km4:594:5959:43
13 km4:594:591:04:42
14 km4:594:591:09:41
15 km4:594:591:14:39
16 km4:594:591:19:38
17 km4:594:591:24:36
18 km4:594:591:29:35
19 km4:594:591:34:34
20 km4:594:591:39:32
21 km4:594:591:44:31
22 km4:594:591:49:29
23 km4:594:591:54:28
24 km4:594:591:59:27
25 km4:594:592:04:25
26 km4:594:592:09:24
27 km4:594:592:14:23
28 km4:594:592:19:21
29 km4:594:592:24:20
30 km4:594:592:29:18
31 km4:594:592:34:17
32 km4:594:592:39:16
33 km4:594:592:44:14
34 km4:594:592:49:13
35 km4:594:592:54:11
36 km4:594:592:59:10
37 km4:594:593:04:09
38 km4:594:593:09:07
39 km4:594:593:14:06
40 km4:594:593:19:05
41 km4:594:593:24:03
42 km4:594:593:29:02
42.195 km0:584:593:30:00

A progressive plan ramps the pace smoothly with distance rather than stepping it at halfway, and the ramp is solved so the two halves land in exactly the ratio asked for. The splits always add back up to the target: the final row of a marathon planned in kilometres is the 0.195 km nobody budgets for, worth about 50 seconds at three-hour pace. Every row is rounded to the second for display, so a column can differ from the finish time by a second or two even though the underlying plan does not.

These are flat-course, still-air numbers. Hills, wind, heat, crowding at the start and the extra distance from running wide of the racing line all move the real split, and a course measured to a certified shortest possible route will read long on a GPS watch. Treat a row as a checkpoint, not an instruction.

The number you check the watch against

A goal time is not something you can run. What you can run is the next kilometre, and the only way a goal survives contact with a start line is if you know what each marker should read as you pass it. This planner turns a target into that list: one row per marker, with the split, the pace it implies and the elapsed time the clock should show.

Two strategies are offered. An even plan spreads the target across the distance and asks for the same number every time. A progressive plan asks for a slower first half and a faster second — the negative split — and works out how much slower and faster each segment has to be for the halves to land in the ratio you chose.

Why the default is even

Time banked early is borrowed at a punitive rate. Running a minute faster than target pace in the first third of a race spends fuel and legs that the last third then has to do without, and the minute usually comes back with several friends. Holding one number, boring as it looks on paper, is the plan that fails most gracefully when the day is harder than expected.

That is why the even option is the one the page opens on, and why the progressive option caps at a modest percentage. A large negative split is not a superior strategy, it is evidence the first half was run too slowly for the fitness available.

How the progressive plan is built

Stepping the pace down at halfway would produce a table nobody can execute: one segment two per cent slower than the next, with a cliff in the middle. Instead the pace here ramps smoothly with distance from the start value to the finish value, and the ramp is solved so that the total is exactly your target and the two half times come out in exactly the ratio requested.

Because the ramp is a straight line, each segment can be priced at its own midpoint pace rather than sampled, so nothing rounds away. Ask for a three-hour marathon with the second half four per cent quicker and the halves come back as 1:31:50 and 1:28:10, with every kilometre between them shading gradually from one to the other.

The last row is not a whole kilometre

A marathon is 42.195 km, so a plan in kilometres ends with a segment of 195 metres and a plan in miles ends with 385 yards. Both are marked out separately rather than folded into the row before them, because a runner reading a 58-second split against a 4:59 target in the last row of the table should see immediately that it is a short segment, not a sprint finish.

Those fractions are also where a mental plan built on round numbers loses time. Forty-two kilometres at target pace is not a marathon; it is a marathon minus about fifty seconds, which is a comfortable margin turned into a photo finish.

What a table cannot know

Every row assumes flat ground, still air and a clean line through the course. A hill moves the split without moving the effort, a crowded first kilometre costs time nobody can get back at the front of the field, and a course certified along its shortest possible route will always measure longer on a wrist. Compare rows to what the body reports, and let the body win.

How large a negative split is realistic

How large a negative split is realistic?

One to three per cent covers most well-run races. Beyond that the arithmetic still works, but a first half that slow usually means the target was conservative rather than that the finish was strong, so the tool caps the figure well short of anything heroic.

Should I follow the splits on a hilly course?

Not row by row. On rolling ground the honest approach is to hold effort and let each split drift with the gradient, then use the elapsed column at a few landmarks to check whether the race as a whole is still on schedule.

Do the splits really add up to the target time?

Exactly, including the partial segment at the end and including any progressive ramp. That is asserted by the test suite rather than assumed, because a table that quietly totals four seconds short is a plan for a different race than the one you entered.