Indianapolis 500

Ordered on the source’s position column and printed from position_text. Those two disagree on purpose: position is a sort key that is never null, and position_text is what happened to the car. 9 of 34 cars here were not classified (9 retired). Across the archive that is 9,612 of 26,048 result rows, so a page that printed the sort key would tell more than a third of every field it finished.

Movement is grid minus finishing position, and it is computed for 25 of the 34 cars entered. It is left blank for the other 9: a car that was not classified has no finishing position to subtract from, and no car here carries grid 0, the value the source uses for a pit lane start or for no recorded grid slot. Scoring a grid of 0 as a gain of every place on the grid is the arithmetic this rule exists to stop; 266 of 26,048 rows across the archive are grid 0, and 40 of those were classified.

Every figure here is one race, read from f1_results and f1_races. The finishing order is the source’s position column; the printed position is position_text, which is a letter on 9,612 of 26,048 rows across the archive (R retired 9,089, W withdrew 364, D disqualified 159). Grid 0 is a pit lane start and prints as PIT, not as a place ahead of pole; 266 of 26,048 rows are grid 0. Time is a total for the winner and a gap to the car ahead for everyone else, and it is present on 8,508 of 26,048 rows only, so most of the archive shows the status instead; where both exist and the status is not “Finished” the status is printed beside the time, because 8,335 of 16,436 classified rows finished lapped or otherwise not cleanly. The fastest lap column is empty before 2004: the source carries lap timing on 8,977 of 26,048 rows and a ranked fastest lap in 447 of 1,161 races, and speed on 8,238 of those, none after 2024. Points are printed as the source stores them and are never rounded to a whole number: 38 of 26,048 rows are fractional (0.5, 1.33, 3.14), which is half points for a shortened race and a split score for a shared car, and every one of them is real. Three races (1951 French, 1956 Argentine, 1957 British) print two cars as first, because two drivers shared one car and the record books credit both; in those three the source hangs the race points on other rows, which is why 5 of 1,164 winning rows across the archive print zero points. Constructors carry no logo in any f1_* table, so none is drawn and none is invented. The 1950 season is complete at 7 rounds.