You found it. Will they remember it?

In one-to-one video meetings at an Australian Football League club, coaches gave players nearly thirty feedback messages. Straight afterwards, players recalled about half of the summarised messages. A week later, they recalled 6% of all the feedback.

That figure comes from Mason, Farrow and Hattie 2021, quoted in Pearson and colleagues' integrative review of video feedback at the University of Portsmouth. It isn't that players don't care. It's that thirty messages get delivered when a handful land. And that was one to one, with nobody else in the room.

So here we are at the end of the series. We've collected the data, worked out which numbers matter, and maybe even lined them up against the GPS. Monday night. Twenty players. Forty minutes. What actually changes on game-day?

Wright and colleagues surveyed elite football players at three Championship clubs in 2016. 54% preferred video sessions of eleven to twenty minutes. 58% wanted the video two days after the match, 35% the next day, and only 4% the same day, all of it from that one survey of footballers rather than from hockey. Analysts working in Olympic sports describe two patterns that hold up: under ten minutes within the hour, or 25 minutes or more the following day (Nicholls et al. 2019).

Format matters as much as timing. Coach-led, top-down debriefs "restricted the athletes' interactions" (Groom, Cushion and Nelson 2012), while players given a say over the video spoke more and took more initiative (Middlemas and Harwood 2018; van Maarseveen 2018, both through the Portsmouth review). One more detail from Wright: some players admitted they engage more with video after a win than after a loss.

Roland describes video as a layer that runs through the whole week rather than a Sunday meeting: a collective debrief at the start of the week, iPads on the pitch during training so a player can see what just happened within thirty seconds, individual clips during the week, and a short, deliberately simple briefing before the game. His own limit is explicit: "I never do a briefing in more than 15 minutes when I use the video" (Get the most from video analysis). Less video, not more, as the match gets closer.

Fifteen minutes on Tuesday, not Sunday night in the group chat while everyone is still annoyed. Three clips. Emily Calderon sets the clock by age group: "For me under 14, you should try and aim for a team meeting of 15 to 20 minutes. That's the time that you keep them focused and you can keep it interesting" (Video Analysis for youth coaches). Twenty to twenty-five minutes at under 16, half an hour at under 19, and if you need longer than that, she changes the format: work in lines or small groups, or give players a clip to analyse and present. And if the team just won, that's probably the week for the hard clip.

My reading, and it is a reading rather than a finding: the player who arrives at the answer is more likely to remember it than the player who is handed it. None of the studies above measured that directly. What the Portsmouth review finds is that how the feedback is delivered, coach-led in a room or led by the players themselves, changes what they learn, and that coach and player working on it together may boost engagement.

Roland has two rules for the group session. The first is about names. "So talk about positions, talk about roles, but don't talk about individuals, ... don't use a name or things like that." His reasoning is practical: players who don't see themselves on the clip stop listening, and "if they are not on the video, it can still concern them".

The second is about who does the talking. Show a clip, let them think, ask what they see, let them propose. "Sometimes we can spend 20 minutes on an eight second clip, because there are always things interesting who come from the players." He says it works at every age, under 14s included, and that from under 16 players are able to analyse themselves alone (Get the most from video analysis).

This one costs nothing and transfers whole. One clip, one question, wait. Robert Noall turns it into a game: "Sometimes during team video, we used to play a game where I would show a clip and I would pause it and I would ask two questions. What should we do? And what do we do?" (The video never lies). Somebody owns up, somebody else offers the better option, and nobody had to be told.

I run these sessions, I don't coach in them, so take the next bit as an observation from the host's chair. In the Roland masterclass a viewer asked how you deal with an attention span of about eight seconds, and the answer wasn't shorter clips, it was letting them argue about the clip. I've seen that same question arrive in the live chat more than once, usually from coaches who assumed the fix was technical.

There's less research here than you'd hope. The Mason figures at the top came from one-to-one meetings, so going private doesn't fix the volume problem on its own. What the practitioners agree on is the split: the group session for the team's patterns, the individual message for one player's decision.

Katrina Powell was asked whether she delivers video to the whole team, to position groups or individually. "Any way that we can get that information across to the athletes. If it's specific to someone, then it's obviously done individually" (AMA about Paris with Katrina Powell).

Roland keeps the Monday meeting collective: "we will never isolate an action of a player". During the week "all the players receive their individual clips", where "I can check if the players watch them", with five or six one-to-ones a week (Get the most from video analysis).

A message with a twenty-second clip and one line. "Look where you checked before you received here." Sent Tuesday. Noall types the comment into the clip as he codes, sends each player their videos, usually on a Monday, through Playerdash or, for his girls' team, Google Drive, and leaves it to them to watch. His argument for the private channel is hard to beat: "Would you rather that your coach told you you did something wrong by sending a personal clip to you or saying it in a group of 20 people?" (The video never lies). Calderon gets to the same place from the youth end, starting one to one at under 14, before anything is shown to the group.

In the Irish survey of 538 coaches from part one, coaches with analyst support used performance analysis more often to shape their training, and a Level 2 coaching qualification was a key dividing line (Martin et al. 2018). The Olympic-sport analysts from part two named time as the thing that most limited their feedback (Nicholls 2019).

Roland's loop from part two: don't give the players every result, "just strong trends", turn those into ideas for training, and with individual players check the numbers again three weeks later.

Same shape without the software, so one thing you noticed, one exercise built around it, and a look at your own tally sheet later in the month.

This is the one finding where the research is thin. The memory studies say our recall of a match we watched is partial. Nobody, as far as I can find, has measured how accurate club coaches' impressions of their own matches actually are. So treat it as a question rather than a finding: how often does what you felt survive contact with your own tally sheet?

By capturing the feeling while it is still live and then testing it. That is what the specials are for, the clips a coach flags live during the match rather than afterwards. Roland: "So it's clips I decide to select during the game, not afterwards." He has a feeling for how each quarter went, checks it on the video afterwards, "to be sure that My feeling was good", and those specials are "the base of the collective briefings" (Get the most from video analysis).

Ask whoever films to mark three moments with the minute, or keep a note on the bench yourself, the same minute column as in part three. Those three moments are Tuesday's clips. And part two's habit of writing the feeling down before you open the numbers still applies.

Today, AI mostly makes the clip exist. Cameras that follow play by themselves (Veo, Pixellot, Spiideo, Hudl Focus) mean a club can have footage without anyone on a tripod. Spiideo's AI Highlights, launched in May 2026, auto-cuts story clips in four sports, and field hockey is not one of them. Playerdash and platforms like it push the clip to the phone. Hudl AI Insights writes match summaries over Hudl Assist data in football, as part two described. None of that decides which eight seconds a coach should show.

For 2028, part two's expectations still stand, and they are expectations rather than promises: plain-language clip search and written summaries over the tags. The one I'd add for this part is automatic playlists per player. Useful, and a real saving for whoever is doing the clipping.

Not on any roadmap I could find: the selection, and the question that goes with it. The research in this part says that is the part deciding whether anything changes on Saturday, and it is a coaching decision. A voice note recorded after the match, next to the numbers, is the cheapest version of that loop, and it is what TheHockeySite.app is built to hold.

https://subscriber.thehockeysite.com/p/two-files-one-game-do-they-agree