Backfill More Tennis Matches

You are right that matches are missing, and I can tell you exactly which ones and why. I asked the source directly for her match history. Between Roland Garros and Beijing, Xinran Sun played 18 matches we do not carry: Wimbledon, Girls - 6 singles matches, 4 to 11 July ITF W75 and W50 Kursumlijska Banja - 6 singles matches, 9 to 26 August (hard court) US Open, Girls - 6 singles matches, 6 to 12 September (hard court) So this is not a backfill gap that re-running a sync would fix. Those matches sit in two categories our tennis sync deliberately filters out: Juniors and ITF. The feed labels them exactly that way, and our pipeline only takes ATP, WTA, Challenger, UTR, Grand Slam main draws and the team cups. Your analysis of the consequence is correct, and it is the useful part of this report. She arrived in Beijing having just played a hard-court junior Slam run and a run of hard-court ITF events, and our model saw none of it. Her last match on record with us was clay, in June, in a junior draw we happen to hold. A model that cannot see four months of a player's results will misprice her, and that is what you observed. Two honest caveats on fixing it: Juniors is a bounded, cheap addition: four junior Slams a year plus a small circuit. ITF is a different matter, it is by a wide margin the largest tier in tennis by match count, and taking it in means a large and permanent increase in what we sync and store. I am not going to promise either on the spot, because the second one is a real cost decision rather than a switch. What I can say is that your case makes the argument well, specifically for players who cross between tiers. A WTA main-draw entrant whose recent form lives entirely in ITF and juniors is exactly where the hole hurts, and it is a narrower problem than taking all of ITF. I am raising it internally on that basis and will come back to you on this ticket either way. Bzzoiro

I owe you a correction before the good news: my first answer to you was wrong about the juniors, and the way it was wrong is worse than the gap itself. I told you the junior matches were filtered out by design. They were not. The daily tennis calendar simply does not list junior draws at all, so our sync never saw them to filter. Checked against the calendar for 8 July: 101 tournaments that day, five of them Wimbledon — Men, Women and the three doubles draws — and not one junior draw. Ask for the tournament directly by id and it answers immediately: the Wimbledon girls' draw returns 12 matches for that date, one of them Xinran Sun against Thea Frodin. What misled me is that we did hold her Roland Garros junior matches, which made it look like junior draws came in and something was choosing to drop the others. They came in through an old one-off backfill, not through the sync. So this was a bug, not a policy, and it is fixed. The eight singles draws of the four junior Slams are now fetched by id on every sync, so future ones arrive on their own. I also backfilled 2026: the Australian Open, Wimbledon and US Open junior draws are complete, 63 to 71 matches each. For Xinran Sun specifically, she goes from 10 matches on record to 26. The six Wimbledon matches on grass and the six US Open matches on hard court are now there — which is exactly the hard-court form you said the model could not see. Her next prediction is built on that history rather than on a clay match in June. What is still missing, and I am not going to pretend otherwise: her six ITF matches at Kursumlijska Banja. That part of my earlier answer stands. ITF is by a wide margin the largest tier in tennis by match count, and taking it in is a permanent step up in what we store, so it stays out for now. You are getting 12 of the 18, including the whole hard-court junior Slam run. Thank you for this one. The report was useful because you did not stop at "data is missing" — you worked out what the model would have done differently with it, and that is what made it worth chasing past my own first answer. Bzzoiro