Jump to content

Svenc

Members+
  • Content Count

    5,069
  • Joined

  • Last visited

  • Days Won

    1

Svenc last won the day on November 4 2016

Svenc had the most liked content!

Community Reputation

1,231 "Here's Johnny!"

4 Followers

About Svenc

  • Rank
    Youth Team

Biography

  • Biography
    "At the end of the day, all we can do as humans is create a tactic which dominates possession, creates clear cut chances and gets shots on target." -- perceived football wisdoms of an eternally to be frustrated Football Manager.

About Me

  • About Me
    Germoney

Currently Managing

  • Currently Managing
    Dishonored II

Recent Profile Visitors

10,120 profile views
  1. Hahah, true. Probably a legacy expectation, I'm afraid. In the glory days of CM, you presumably could totally smash the game 38-0-0. In parts because of weaker AI managers making worse decisions, because of match engine related stuff, player motivation/morale/consistency stuff and some more. Find the magic sauce formula, and you were set for eternity. It's also a "football simulation" vs "game" expectation thing. Games can be "beaten" for life -- a sports as random as football evidently not. Does FM aim to to be the first or the latter, and what does that mean for players?
  2. Btw, you wouldn't need to script a thing to simulate point drops. All you'd need is an RNG (which the game has, and which you can't fully beat and shouldn't). Additionally, even the biggest chances in a match within that RNG should have a significant chance of being misses. A penalty after all is a ~76% chance of being a goal, average, and chances higher than 50% are rare. Therefore, point drops despite a de facto domination of a match should be happening. In football they are legion. Think the the dudes on Fallout / Baldur's Gate 3 wondering how they could have missed a couple times
  3. 999 out of 1000 FM'ers were absolutely terrible at doing just that. (And sorry for the repeat, but they had ample role models all over the media).
  4. Technically perhaps one for feature suggestions, but graphics like these would be also a nice addition as to future iterations, me thinks. ErpUe4CWMAE7spc (2048×946) (twimg.com) #performanceanalysis
  5. Not much. If you look around say understat.com, there's teams not scoring with an xG 1.5-1.8xish on almost every match day in every major league. (The highest recorded without scoring was an xG of ~5 and no goal -- that's unfortunate). Arguably this only shows this needs a tutorial as to stats, and what the biggest use of them actually is. Picking random matches is no use.
  6. He's right though. Barcelona have the highest amount of shots per game in the league (on target as well). To bring this to FM 2021's new feature, xG, they also may be sitting eight place, but according to Expected Points, should sit 5th. They are scoring less than expected as well as conceding more than expected based on the quality of chances (xG). In other words, at least according to this model, whilst still hugely disappointing, they are also a bit unlucky. Usually, this doesn't last very long. La liga xG Table and Scorers for the 2020/2021 season | Understat.com I think the point
  7. Expected Points is based on what they should have collected based on the xG data of their matches finished. According to xG, they had been the better team oftenly, but didn't take the points. This may suggest that there was some bad luck involved as well. Newcastle in 2011ish finished 5th place in the EPL (albeit with a suspicious Goal Difference of but +5), upon which they gave Pardew a lifetime contract. The next season they finished in the lower third again. Expected Goals/Points analysis conducted by Brentford FC had suggested their actual performance hadn't actually changed; but thei
  8. What where their Expected Points like? Just interested, because: Borussia Dortmund's crisis isn't a crisis at all, and stats prove they will rebound (espn.com)
  9. That depends on what the coders would define as "accurate". Assuming they are targeting for an average conversion over dozens of penalties for an average taker on their level of ~76% (which is pretty much the average penalty conversion). Then it's a case of defining what the lower end of the spectrum should convert and what the higher. In football over dozens of attempts (not merely 20-30, sample size and all), the elite kickers regularly tacking kicks eventually convert roughly 85% of their pens mostly. Would be an interesting long-term test, but what SI are targeting is wholly up to them.
  10. Pretty much in line with all my tests new and throughout the years. Plus what SI once said: That as soon as a player is loaded from the db, no matter his attributes, he will be a (semi)professional football player, not some pub type. Once I edited all players of a league to have very low pen ratings, and then edited them to have very high. Most of the penalties were simply converted on both edits. Didn't collect long-term numbers, but I'd be naturally concerned if there wasn't a difference between the above taker and a specialist over e.g. 100 of penalties though. There's also a reason wh
  11. Most one on ones are actually rated in the 0.3xGish range (individually up to xG 0.5+). They are considered 1 in 3 chances plain average. Coupled with that the conversion is now higher than even 50%, and the many misconceptions of FMers (they consider one on ones near penalties or even better), that explains why there are so many "pleased with one on ones" posts now though. One on ones are great chances, but plain not open goals (as in other, even better chances that leave the keeper no chance of anticipating).
  12. That's XCOM Baby. (Assuming the average chance of scoring was a very decent 85%, missing five in a row would be a 1 in 13.000 chance. Long odds, but not quite akin to winning the lottery either. Over thousands of players, it should occur quite regularly. Only if SI were to ever hardcode that on x misses, the next kick is always a goal, wouldn't this occur.)
  13. According to the overall stats seen, there seems nothing obviously off -- this was posted also by SI. Now hear me out, this will be reported every release by someone, even if it needn't be a bug (which it could still be). You know why? Big numbers. And things not artificially being scripted to stay within "boundaries". Until a couple days ago I was playing Baldur's Gate 3 Early Access, an RPG based on lots of numbers crunching also. Going online, it was apparent that (just as in any random number game..) people were convinced there was something wrong with the random number generator. Peo
  14. At least they did score. https://understat.com/match/9887
  15. https://understat.com/match/11662 https://understat.com/match/11669 https://understat.com/match/11673 https://understat.com/match/11700 https://understat.com/match/553 https://understat.com/match/563 https://understat.com/match/573 https://understat.com/match/599 https://understat.com/match/1029 https://understat.com/match/1036 https://understat.com/match/1040 https://understat.com/match/1051 https://understat.com/match/1063
×
×
  • Create New...