Dodge

You went up, so here is the new one. Built backwards from your own 309 starts rather than from theory: what actually paid, what the last plan got wrong in your favour, and 61 division-2 races with a projected placing against the field that really turns up. If you only open one of these, open this one, ya mug.

The plan page looks forward; this one keeps score. Every race you have started this season, what the model reckoned you would do, what you actually did, and whether your own entry band said to be there at all. The green rows paid, the red ones were starts you did not need to make.

Straight to the actions. Three of them are free and one of them is just a signup habit that cost you 212 starts for 180 points last season.

Thinking of falling down the divisions to grab some silverware? This works out the best day to stop scoring, from your own points: which trophies you land on, when each signup opens, what it costs you in division-2 points, and how close each signup is. Rebuilt every update, so the dates keep up with your season.

Last season, kept for the record — the audit that got you promoted. Good, bad and downright ugly, three moves that cost nothing, and all 43 division-3 races with a captain, an effort setting and a predicted placing next to each one. The review below marks it.

We told you what to do; this is us grading both of us on it. The plan gets an A-, your following of it a B. Sixty per cent of your season came out of seventeen races that page named by title — but you skipped nine more, and three you entered without the rider we named, all three for nothing. Also the bit where we were wrong.

Phone-sized scoreboard. Points, form, who is actually pulling their weight. Refreshed twice a day.

Straight to the race-by-race list, ordered from today forward. Bookmark it and stop guessing which race to enter.

The table that drives the whole plan: every entry you made last season, graded against what the model predicted. Time trials convert at 59% when it says top nine and at nothing from 26th back; road races pay from the 10-25 band instead, because the model cannot see a captain with a team behind him.

On the old page, and still the clearest table of it: every time-trial rider in the game ranked on the proper composite (tt+tq+fr) within each division pool, so you can see exactly where yours sit before you enter one.

Which hill races are actually the same race wearing a different hat, and who cleans up in each. Section 4 is a training plan written for Arjan Woering, who rides for this very mob.

Straight to the plan: where his remaining training points should go, and where they would be flat-out wasted.

The full saga. Sixty-odd seasons, a bloke who raced into his nineties, and a team that never sacked anyone. Get a tissue.

The trading card. Print it, laminate it, stick it on the fridge.

Season points, plotted. Line goes up, line goes down, such is life.

Zolnyanszi Tamas fronted up to a cobbles race with the weakest lead-out train in the whole field and came 8th of 108. Four charts on why that was no fluke, why 108 finishers is really only 12, and the nine races on the calendar that suit him better than the one he actually rode.

Every rider's eight stats on a radar, one on top of the other. Tap the names to compare, or use the presets to pull up a captain and the blokes who ride for him. Best way to see at a glance who is a sprinter and who is just along for the ride.

Last season's radar, kept for comparison. Old seasons are never deleted, so this is what the mob looked like before the training and the transfers.

Where the mob actually comes from, pinned on a world map. Four Aussies and a United Nations of imports. Snapshot from season 76, so a couple have since moved on.