Anything else that deserves a little mention go on mark it's

just as much as anything we're talking about metrics i just like always banging this drum um more broadly at the the season level and just talking about season state so we can think about metrics for you know each team but i think that when the teams or the players actually sort of have those events have those actions and record those metrics, I think is crucial to then acknowledge and sort of slice and dice rather than just at the season level, this team has done X, Y, Z, because it's so important to understand when they're playing and what they're playing for, I suppose. So I did a piece midway through the season to basically try and quantify whether Liverpool actually had the easiest schedule for the first half of the season. And of course, everyone has played everyone at that point. But thinking about the strength of the opponent at the time that they played them, I think is crucial. There's a few examples that I gave of Liverpool when they played Manchester City without Rodri. there were other examples of that that actually sometimes the circumstances that are out of control of the team can actually then influence the metrics that they post at the team level at the season level as well so you think about you know teams maybe at the bottom half of the table as well they might be more likely to play for a draw if they know that they've got a greater idea of what the league table is looking like and they might take a different approach. So just to think that every team goes into every match trying to sort of instill the same ideas and post the same sorts of metrics is obviously nonsensical. And I think that we sort of need to acknowledge that when we're looking through these metrics, because it's a broad brushstroke, but with good reason. If I

may, I wanted to give some flowers to mark for his match prediction model this season um we had a bit of a running joke that it seemed to hit every scoreline bang on um but it'd be a really cool opportunity if he's obviously prepared to sort of share how you sort of put metrics into um you know work out things like that or what the process is and actually this is a good example of how data is used in more of a predictive sense to say not just what has happened, but what might occur and what might be likely based on performance or more underlying stuff.

Yeah, well, it all, as most roads lead to XG. So it's using it as a historical look at every sort of per shot XG of every team to essentially create an index or a score from an offensive and defensive perspective for every team so you can sort of understand more broadly what the team strength is at that level and then essentially simulate over many many sort of iterations what if you were to pitch the strength of this team in the underlying numbers not necessarily what they've done you know in terms of the actual goals they've scored what those underlying numbers actually would be if you pitch them against each other from a defensive and offensive perspective then it gives a probability a likelihood of what the score will be and then as a consequence who will be most likely to to win or what the the outcome would be a win loss or draw i think what's interesting and elegant about the model which i have stolen so it's not necessarily me being super intelligent i'm just good at theft but um it's it's that it's there's a decay to it as well you can weight the the strength of those scores more strongly for games that are most recent and then slowly decay it as the the games go further back so that you're placing more importance on form essentially it accounts for home field advantage as well so just accounts for multiple factors to then come out with the overall score. And there were some good ones. I did have some hits, I must say. The best one was Manchester City-Brentford 2-2. I think a last-minute goal as well, which I was very pleased about. So, yeah, we're using it carefully. We're not just sort of throwing it out there. But those that I've used have fortunately been bang on.