mate, do you have any idea of the kinds of variables that are required to model any system to any degree?
My god, do people follow doctors on Twitter and then start going round claiming they can treat any disease?
The analysis of statistics is a science. All scientific claims require proof.
A lad claimed that Harry was weak in analytics. If your going to try to use a science to argue your point, prepare to be challenged. So I was mearly inquiring into said lads qualifications is this field? How many years did they study, did they build their own models? is the weather included in the modelling
I'm trying to point out that the clubs models are far more in depth than the models that are available in the public domain, so when seeing someone using analytics to write off Harry it's probably best to see if that person has any insight or just another numpty following a real mathematician on Twitter
Wrong quote, right person though, sorry
Just a few points to clear up :
1) Statistics ARE an analysis. You analyse certain events or occurrences by counting how often they happen and recording information about the circumstances in which these occurrences happen and their relationships to other occurrences, then you are analysing occurrences or phenomena and using statistics to do it.
2) It is not the qualifications that determine whether or not someone's analysis is valid. There are plenty of weak analyses by people who are very qualified to make them, but they just make important errors (i.e. bias) that weaken their position. more important are the metrics they actually use, and how their were derived, and how they use them and interpret them.
3) I think the club probably look at analytics as a portion of whether a player is 'undervalued' in the current market. For instance, if on paper one player does exactly the same things in a game as another player, but that player costs 10% of the other player, then it would make us question whether or not we are getting 90% more value by paying 90% more. In other words - the club is looking at "bang for buck". If you get players for free, the opportunity for "bang for buck" is the highest (because we haven't really paid anything for them). Young players are generally on less money than established stars and so their is perhaps arguably more value there, and less risk also. Rather than blind luck, I would say all of the signings we have made are generally players that are significantly undervalued by the market and we saw that there was genuine quality there but nobody else seemed to have noticed (hence their value not being that high). Can you think of anyone we have really overpaid for?
I'm not sure what Wilson's metrics are, but in our situation, I'm not sure that it actually matters a great deal. We aren't evaluating whether to buy him or not. The most important thing is that we evaluate what he is worth accurately (and that will also change as his contract runs down). If an offer much greater than that comes in for him, then we should sell. If we don't get an offer more than what we value him at then we shouldn't sell. In the meantime, while he is still here, more meaningful than analytics will be for the boss to see him in action and give him an opportunity to stake a claim at the beginning of the season.