#812 · Round 1: the three drawings nothing draws yet

PR 1 (#824) made every graph the app already drew follow its ticks. This is the rest of #812: three drawings that don't exist yet, each on the tab graph setup places it on. Pick a player on the left and every part redraws for him.

Real data: Man Utd's 25/26 to December and the December Premier League capture, the same export as the #810 inventory and #811. Every dot, middle and share is worked out live by the app's rules: the card's figure rules (FM's per-90s it never worked out, a ratio with nothing attempted), the minutes bar (a third of the most any of ours on his side had played at that capture), the evidence floor (#683) and the league peer groups. The boxes with a dashed amber edge are the mockup's, not the app's: what graph setup would have ticked, and the choices to compare.

Decisions, with my recommendation

  1. Player evolution draws his path as a third view, Charts, beside Figures and Graphs and under the same chips. Every chart placed there is drawn, category by category (the seed places 21, minus any for the other side), each as a dot per capture from the first of the season picked to the capture picked, joined in order and the latest ringed. It's Squad comparison's History, but for him alone.
  2. Our squad faint behind him, at the capture picked, with the corners split at their middle. Then it's the same chart as Squad comparison's and reads the same way, and its scale is the squad's spread rather than his own five dots. Scaled to him alone, a point of pass completion fills the chart and looks like a collapse. The switch shows him alone.
  3. A statistics chart shows the season to date, as FM holds it. A total only grows through a season, so the path always climbs, which is why Squad comparison's History draws no trail for a total (#786). But one player's steps are each that month's own football, so their direction is the reading: Goals against xG steepening is finishing above the chances. The switch shows each month's own figures instead. Counts and Rating can be taken apart honestly; a per-90 can't, so metric charts stay season to date.
  4. Months now, seasons later. Once there's more than one season, a Seasons switch draws a dot per season, each at the same point of its season. That's like for like, #786's own rule for totals.
  5. The card's comparison tabs: the charts placed there join the Scatter view, above each dial's own scatter, against the same crowd the rail picks and with the same evidence floor. They're drawn in their own units with their own corner labels. A dial's scatter turns every axis so good is high, but a stored chart's labels are written for the axis as it runs: Possession lost is lower-is-better, and turning it would put "Ambitious but costly" in the wrong corner.
  6. Those two tabs take metric charts only. The league capture holds no statistics, and our squad's crowd is built from metric captures. Graph setup would grey a statistics chart there and give that reason. Your #812 answer had the card taking any mix; this is the one exception the data forces. The Squad tab could take statistics if its crowd read them too.
  7. A donut sits with its parts: at the head of their category's group, as the save split heads Goalkeeping. Statistics have no group on that tab, so a donut of statistics, or one whose parts span categories, sits in a row above the groups. It's the same on the card and on Player evolution, where the figures are the season to date at the capture picked.

What building it found