Football statistics - Expected goals - xG - Why Jeff Stelling is wrong!

[@elastronaute1198]

I can't believe he goes into so much depth with his analysis... this can't be normal... no wonder he is so successful. I sincerely hope you do not need to be like this to be a successful trader, or I may as well give up now. I don't have the dedication or obsessiveness to do this.

[@betangeltv]

No, you don't need to go into so much depth. I just like digging much deeper than anybody else.

[@andyleitch6191]

Geoff should read some Nassim Nicholas Taleb. Having done precisely no research on XG, my guess is that it has to be a function of the geometry of the shot angle, the number of shots, the competency of the shooter and the competency of the goalkeeper. It's interesting to try and quantify the competency of the shooter though. If he/ she is on a good run of goalscoring form does this create a "gambler's fallacy" effect when their chances of scoring in the next game are overestimated? If the shooter's confidence is high maybe they merit the higher rating in the XG calculation. If strikers sign from inferior leagues or are coming off a long injury lay off how does this affect the XG? Anyway, lots to look into for the inquisitive...

[@betangeltv]

Many years ago I stared with geometry of shot angles long before xG was a thing. But essentially it's an evolution of that I guess!

[@silviupopescu1088]

Xg was created to help football players is not for betiing purpose.I'l give you an example .. you as a striker ( main shooting foot your right one) ... you on the left side of the pitch at 16m mark what options do you have 1.shoot with the right leg (what's the propbility to score from there ) 2. Try to get a better possition (lost the oporrunity and time) 3. Pass the ball Can you use this dinamix to place a bet I can't What interesting for beting /tradi g are the goals.scored from set pieces ( corners , free kicks). (I just review.world cup final.2006 for fun ... Pirlo was a master of this kind moments ... and with two "stikers" both almost 2m height guess the result ) Sorry for my english.To make real money from trading footbal you need to watch the game .Everything else is already in the odss (like statistics) - football being such a dinamic sport played by 22 player s is hard to predict.Try tenis (4 surfaces 2 players)

[@MMAMedia-hh7sn]

I was wondering if you use Regression analysis as a tool when building your models and also how important you think it is, if at all?

[@betangeltv]

I model everything and cross compare to real life and actual results. So yes.

[@skintslots]

Fascinating stuff. Just a quick question,you mentioned that roughly 10 shots would equal around a third in goals,so over 2.5 with maybe small bets on 1-1,2-0 and 0-2 would seem appropriate but I am guessing those 10 shots mean a mixture of shots on target and shots off target added together rather than 10 shots on target?

[@betangeltv]

10 shots would usually end up with three on target and one of those a goal. Those are the rough ratios. But there is a lot of complexity behind that.

[@wilo3141]

hi peter, amazing video, just wondering which site you use for live stats (shots, etc.)

[@wilo3141]

sorry and the name of the video you mention on which statistics you look to determine how many goals likely to be scored, very appreciated

[@betangeltv]

I currently use info gol to get live data, but it's a radically developing market and more are likely to come around. I've just finished developing my own model so I can collect data real time and make my own refinements.

[@6madx]

@ 3.48 yes I did that in the Swansea v Bournemouth game today..SOT said a goal either way (as my stats at KO)... final score??? 0-0!! ;i-)...liked subscribed by the way and 100% agree re Geoff!!

[@douglaskercher9032]

Sorry for my ignorance but what is SOT?

[@thewealthkitchen]

shots on target... maybe?? I'm only guessing.

[@alexh2292]

This stuff is fascinating. 👍

[@elastronaute1198]

I still don't understand how knowing where a goal might be made from can help you trade or predict actual matches?

[@betangeltv]

You need to frame a trade, to see when price action could occur and how that will influence your payoff.

[@chrisstyles6803]

Expected goals is more of a fallacy. It is one of those things that sound so nice and revolutionary but is not realistic. So who calculates the expected goals? Is it the fans? Every game should have an officially released expected goal; we should start taking expecting goals seriously only when that happens. The expected goal will be next to the ball possession, cards, corners and offsides stats in every game

[@betangeltv]

I agree, it would nice to have an 'official' stat. Never use statistics like a drunk uses a lamp post, for support rather than illumination.

[@elastronaute1198]

what sites do you use for football data?

[@betangeltv]

I'll do a full video on it next.

[@elastronaute1198]

@@betangeltv had a good trade on the germany holland match tonight :D - 2 nill to germany half time, laid them at 1.16.

[@nastastic]

I was thinking that if you built a model using 19 variables with differing importance, and you create an algo/your recipe for prediction, how do you rate each variables importance and have you changed this model over time?

[@mediapc4747]

Knowing what things affect and outcome and predicting those things are very different. Peter mentions that the pass to the shooters position is a factor, true, but it's no something you can just drop into a formula and use. 'Building a model' doesn't mean building a prediction engine, it just means understanding which factors are significant.

[@grahamh2457]

Jeff is right. Football’s not a science.

[@betangeltv]

All major clubs have stats divisions now. So it's now a game of oneupmanship!

[@Britpopsoul]

Xg treats all footballers as equality ability therefore it’s fundamentally flawed

[@betangeltv]

Xg has many flaws

[@robertfield4001]

And if my aunt had bollocks, she would be my uncle