#302 · Squad comparison, redesigned from scratch · round 6: History

Ross: "another view where you have all the history available to check past capture points and seasons if you want" - "current comparison right now based on everything against a comparison from last season". The page opens on it, with the centre backs.

1. At against Against. Round 4's two pickers come back, on History alone, and fold together as they did. At is the latest by default. Against is this point last season by default - Dec 25/26 against Dec 24/25 - with the end of last season one click away on its header, and any earlier capture in its chips. Move At and Against goes with it.
2. History's own switch, Attributes or Statistics & metrics, with the other tabs' views under it and each man's move since Against added: marks in the Grid, a tail on the Spread, a radar each on Shape with Against dashed, a ▲▼ on a figure, a trail on a chart. Only what somebody moved on stands in for Only where they differ.
3. Who has moved heads the attributes: each man's attributes summed, most gained first, by category, with his biggest moves. A keeper has 47 attributes to an outfielder's 36, so keepers are ranked apart. The Grid's category titles name the best at Against as well as at At, so folded they say whether the lead has changed hands.
4. The football, two rows a man: the season to date at At, and at Against under it, each as FM held it. The best of each season is bold - #364's who has had the better season, by its rules: everyone level at the top, a rate only among players over that capture's own minutes bar, nothing where there is no better end. Hover says the best's figure, and how far he is from it (#364).
5. What is set against what. This point last season is the same point of two seasons, so every total compares, and each figure carries its move. Against the end of last season, half a season's totals and a whole season's do not, so only Rating moves - and the metrics, which are rates. A rate moves only where he was over the bar at both ends, so a two-game August throws no arrow. The charts keep the same rule: a chart with a total on either axis draws trails for this point last season only.
6. A season split by a move, #364's open point: a row is his spell at that capture, as FM holds it, and the spell before it is named under the club - hover for its figures. The U21 midfielders, against the end of last season, show one.

Real: 25/26, July to December. Illustrative: 23/24 and 24/25 - #300's made-up seasons, which now carry football too, drawn around each man's own 25/26 and tagged made up wherever At or Against lands on them. Today's U21s played next to nothing for us then, as 25/26's own U21s at our club have (0 to 169 minutes by December). The made-up seasons know only today's players, so last season's minutes bar is set without anyone who has since left - the app will have them, as Player evolution does (#163).

Corrected with this round: the minutes bar's set is the card's own - every player on the team who played for us, whatever his squad. Rounds 3 to 5 took only the squads ticked on the rail, which the app does not. The First Team's figures are unchanged by it; the U21s' are not - see round 3's fourth finding.

Found building it:

1. A third of the First Team has no last season. … - the summer's signings, and February's. So not here by then is a row in three, not an edge case, and it says when he was first captured.
2. Early in a season the bar holds little. … So this point last season in August sets a game or two against a game or two - over the bar or not, the arrows then say more about two games than two seasons.
3. Nobody moved club inside the real half season: …. The moves come in January, which the save has not reached, so the split season here is drawn from the made-up loans alone.

Ross: "the attributes and stats views are latest capture point only, you cannot select previous ones or seasons. And then another view where you have all the history available to check past capture points and seasons if you want" - "current comparison right now based on everything against a comparison from last season".

1. Attributes and Statistics & metrics read everyone's latest capture, and nothing else: At and Change since have gone from both. The header bar says so - at their latest capture, Dec 25/26 - and Player comparison already works that way (#715), so the two comparison screens agree on what now means.
2. The metrics' Spread has gone - Ross: "i agree with the spread not being as useful here". Figures and Graphs stay, and the Graphs stay one at a time.
3. History is the third tab, and the next round: this season against last season, for everyone picked, with #364's who has had the better season. It marks its place for now.

Ross, on round 3: "The At and change since cards should be expandable/collapsible on attributes and stats views (linked together so both expanded or both collapsed. Stats should be collapsible as well. Rather than all the stats and metric graphs just being visible at all times a way to flick between them as you have a lot of scrolling otherwise".

1. At and Change since fold together, on both tabs: a click on either header folds both to a line each, or opens both, and the other tab keeps the same. Folded, each still says what is picked, back to the latest still works and Change since still switches. They start folded here, since the latest is the usual pick - say if they should start open.
2. The statistics fold to their title and one line: who has statistics, and who is under the minutes bar, since that is what greys the metrics below.
3. The Graphs show one chart at a time, drawn large. Every chart is listed on the left by section, each saying how many it plots or why it draws nothing; ‹ and ›, or the left and right arrow keys, flick to the one before or after. Large enough for the quadrant names to read, and with ten or fewer picked each dot carries his name.

Ross: "lets mock up the stats and metrics again not sure where to start with this. They need to be shown together because metrics mean nothing really without the stats" - and the charts set up in Settings, with their flag for whether each shows on Squad comparison, "as to what graphs we should show".

1. One tab, Statistics & metrics, beside Attributes under the same rail and header bar. At and Change since are the view's, so switching tabs keeps the capture. July is pre-season and the made-up seasons carry attributes only, so those say there is nothing to show.
2. The statistics always head the tab, whatever the metrics under them are showing: a row per player, where he played (✈ a loan club), then every statistic as FM held it, the season to date. The best of those compared is bold, by each column's direction, and none where it has no better end (Sub, Fouled). Click any heading, or any metric below, to order the players by it everywhere on the tab.
3. The minutes bar is the card's (#608): a third of the most minutes any of our players who played for us had by then, keepers apart. Under it a player is tagged, his rates are greyed, and he is never the best at one. Totals need no rule - most goals is most goals (#364).
4. The metrics under them, three ways:
   • A Figures - a card per category and a column per player, his minutes under his name; each figure coloured as the card colours it, a quarter each against our players over the bar; the best of those compared bold on a lighter cell. A blank is the card's rules (#414, #510, the denominators), and hover says which.
   • B Spread - each figure on its own line from our squad's lowest to its highest, the better end to the right, our squad's quarters along it and its middle ticked.
   • C Graphs - the charts Settings has on for Squad comparison, which as seeded is all 25: ten statistics pairings, then fifteen metric ones by category with their quadrants. Each player a dot, our squad faint behind. The quadrants split at our squad's middle, or at the middle of those picked - that switch is DesignNotes' open question about the dividers. Hide on a chart stands in for switching its flag off in Settings.
5. Change since: under each count, what he added since then; under his rating, his rating over those games alone. The metrics do not move with it - a rate between two captures is not a month's own figure, since FM writes per-90s to one place.
6. Goalkeeping first for keepers, as on Attributes - in the Figures, the Spread and the Graphs.

Real: 25/26's statistics and metrics, August to December, for all 85 players; the U18s have none captured. Illustrative: nothing new - the seasons before 25/26 are attributes only here.

Found building it - computed on this page from the December captures, the whole First Team picked:

1. Some charts say nothing for this squad. … And the two Goalkeeping charts are seeded for outfielders as well (AppliesToOutfield): Conceded vs xG prevented puts …, and Saves vs save % plots …. Worth making those two keepers only in Settings, and switching Red cards off here.
2. A keeper cannot be coloured. Of our keepers only … is over the bar, so a keeper's figures are measured against one man and read as no verdict - … - bar xG prevented, which has a scale of its own. The card has the same limit.
3. Blanks are common, and right: … of the First Team's metric figures have none - no minutes, a per 90 under ninety minutes, or a ratio over nothing attempted.
4. Loanees break the card's yardstick. The card measures a man against every player on the team who played for us, whatever his squad - so a U21 is measured against the First Team. At December …. A loanee's figures were earned in another league, so our squad is no yardstick for them - his own league's would be, or none at all - and that is a question to settle. (Corrected with round 6: until then the mockup took only the squads ticked on the rail, which the app does not.)

Ross, on round 1: "keep everything except the ranking, the 3 other ones are great".

1. The Ranking has gone. The Grid, the Spread and Shape are as they were, and ordering the Grid by an attribute still ranks the players on it.
2. Each category names its best in its title on the Grid - the best average of those picked, every man level on it included: Technical (14) · best Lisandro Martínez 11.2. Collapse all, at the top of the Grid, folds every category to its title, which leaves the best in every category on a line each. Ross: "so if we collapsed them all we would see the best player in all categories".
3. Goalkeeping comes first for keepers, on the Grid and the Spread - Ross: "that is the most likely one you are going to want to see". Try Goalkeepers - 5 below. With outfielders in the pick as well it stays last, where their empty column costs least - Keepers and outfielders shows that.

1. The rail is Player comparison's (#741, Multiple): the squad chips, the search, Filter by position, Hide players out on loan, the pinned selected list with Clear, Select all under each position, Add all, and the ⓘ beside every name for his card. No Former players. One difference, and it has to be one: Player comparison's rail folds itself once two players and a role are picked. There is no role here, so folding at two players would shut the rail on the man picking his third - it folds by hand (☰) only.
2. The header bar says who and when: Comparing 6 · Manchester United · First Team · at Dec 25/26, then the names.
3. At is the capture everyone is read at - the latest by default, newest season on top, the months in order. Each man is read at his latest capture on or before it; one with none by then - a later signing - is shown as not captured, and a chip with a * says who on hover.
4. Four ways to compare, at the same At and all computed:
   • A Grid - FM's figures in a card per category and a column per player, the best in each row bold on a lighter cell (every one of them where level). Each card opens on its category's average. Click an attribute to order the players by it; click a category's name to minimise it, as Player evolution's Grid does.
   • B Spread - a 1 to 20 line per attribute with everyone on it: where they sit and how far apart. Hover a name in the key to follow him down every line.
   • C Ranking - pick an attribute, or a category's average, on the left, and everyone picked is ranked by it on the right - dropped in round 2.
   • Shape - the radar, on the app's own axes: up to six drawn over each other in the Radar's colours, more than six a radar each against the middle of those picked. A keeper has his own axes, so a mixed pick draws two.
5. Change since is off by default. Switch it on and every view adds each man's change from an earlier capture - marks in the Grid, a tail on the Spread, and on Shape a radar each, From against At, as Player evolution draws it.
6. Only where they differ (Grid and Spread) hides what everyone has the same - mostly for a pair.

Real: 25/26, July to December, for all 85 players of the December save - the same data as #300's Player evolution mockup, with the rail's positions, loans and filed groups from #741's. Illustrative: every capture before 25/26 - #300's made-up seasons, generated backwards from each man's first real capture and amber wherever they appear. Joining dates are roughly real life's, so the summer's signings are not captured at an At in 24/25.

Found building it - computed on this page from the real December captures:

1. Level at the top is common. Among the First Team's six centre backs, … of the 36 outfield attributes have two or more of them level at the best, and across the whole outfield First Team …. So "the best" marks every man level at the top - a single winner per row would be wrong about a third of the time among six, and half the time across the squad.
2. A wide pick reads the positions. An attribute's range - best less worst - has a median of … points among the six centre backs, and … across the whole outfield First Team. Most of what the whole squad shows is who plays where, which is why Player comparison judges by role (#418) - and why Select all under a position is the pick that reads best here.
3. Within a season, Change since is about the young. Over the real half season, July to December, … of the 29 First Team players moved on any attribute, and the most any gained, summed over his attributes, is …. The U21s: …, the most …. Worth knowing before it is given room - and since round 5 it is History's.