|
Post by tachyon on Jan 13, 2018 13:39:42 GMT
Espanyol overall have been fairly solid defensively, both under QSF and his predecessors. You always get blips in LaLiga depending upon when you get served up Barca & R Madrid as opponents. So the big blip is just the big two's xG running through the system. Attachment Deleted
|
|
|
Post by tachyon on Jan 13, 2018 12:32:12 GMT
Here's Man City's xG trends. Frightening, even with their financial clout. Attachment Deleted
|
|
|
Post by tachyon on Jan 13, 2018 12:10:06 GMT
It's difficult not to draw that conclusion. Basic in house statistical analysis should have picked it up, but as you say, it was probably clear from the simple eye test. The xG trend is one of the worst I've seen for years, unfortunately.
|
|
|
Post by tachyon on Jan 13, 2018 11:53:26 GMT
I've taken a look at our underlying stats since the start of the 2016/17 season. Premier League teams increasingly evaluate themselves in terms of "expected goals". They rate the quality and the quantity of chances they create, rather than the actual number of goals they score and allow. The stat has been featured after games on Match of the Day this season. It tells you if a team has a sound process, rather than relying on just actual goals that are often streaky in nature and prone to randomness or "luck". If you have a good process, eventually you will tend to get the rewards you deserve, and if you have a poor process, the same is unfortunately true. Stoke's process has trended from that of an average mid table team at the halfway stage of the 2016/17 season to that of a relegation side, currently. Attachment Deleted Here's our plot from the start of 2016/17 to now. The orange line shows the amount and quality of chances we are allowing our opponents per game. It was around 1 expected goal per game at the start of 2016/17, slightly better than average and it is around 1.7 expected goals now. One of the worst in the league and continuing to climb. Chance creation isn't as seriously troubled as our defensive metrics, but that too has declined from 1.2 at the start of 2016/17 (broadly average for the league) to around 1.1 expected goals per game, now. If you look at the expected goals of all 20 sides and their current points/ goal difference totals and simualte the remainder of the season, Stoke go down in around 30 percent of these simulations. So we're in trouble, but not beyond being saved. particularly by a more solid defensive process.
|
|
|
Post by tachyon on Jan 6, 2017 21:34:23 GMT
This is Dan
|
|
|
Post by tachyon on Apr 27, 2016 19:06:51 GMT
I assumed that goals from set pieces would be the difference but I also like to see data. According to whoscored.com Stoke have scored 6 from set pieces (I didn't think it was that many) and only Swansea have less with 4. The median value is 12 and the highest is 19. So that appears to be only part of the problem. I am interested in stats on goals conceded from set pieces if anyone has that information. I don't know how reliable whoscored.com is and I'd be interested if anyone has any better sources. Attachment Deleted
|
|
|
Post by tachyon on Mar 14, 2016 9:58:47 GMT
Chances of WBA finishing above Stoke based on the relative abilities of the two teams, current position/goal difference and remaining fixtures. Attachment DeletedOverall, Stoke much more likely to finish ahead of WBA, but the 23% chance that we don't is far from insignificant. Most likely outcome is we maintain the current gap, even with their game in hand.
|
|
|
Post by tachyon on Mar 13, 2016 10:38:17 GMT
Most likely final points total for the 18th placed team at the end of the season. Attachment Deleted33 points currently the most likely based on remaining fixtures and the relative strengths of all 20 teams.
|
|
|
Post by tachyon on Feb 7, 2016 12:48:09 GMT
The expected goals from chances we're allowing with and without Ryan. Attachment DeletedBasically we're allowing one more attempt on target w/o Ryan compared to with him and that's costing us around 3 tenths of a goal per game when he's absent. We may be missing his aerial abilities most. Without Ryan ~ 18% of opponents on target attempts come from headers, resulting in five goals. With Ryan ~ only 13% of on target attempts from opponents come from headers, resulting in no goals.
|
|
|
Post by tachyon on Feb 7, 2016 9:41:37 GMT
Premier League teams are increasingly measuring their performance in term of expected goals rather than actual ones. They look at how many goals they might expect to score or concede based on the amount, location and type of attempts they are creating or allowing. Process rather than outcome. Arsene Wenger’s talked about Arsenal’s “expected goals”, but don’t let that put you off the idea. Here’s the expected goals profile for all 20 teams so far this season. (green is good, red is bad). Attachment DeletedOn this measure on a per game basis we are 16th. We’ve scored about as often as our chance creation merits, but we’ve actually let in about four fewer goals compared to the quantity, quality and shot type of chances we’ve allowed. So basically Jack’s saved our season from being one that could have turned into a proper relegation fight. The bad news is that extreme performances (like Jack’s) almost always regress towards more normal levels. He won’t stay “hot” forever and if I were looking at our core expected goal values, I’d be a bit worried.
|
|
|
Post by tachyon on Feb 6, 2016 22:15:40 GMT
One last graph. Most likely current final points tally is 48. Attachment DeletedAbout a 15% chance we get 54 or more. 99% chance we get the magical 40 or more for the pessimists.
|
|
|
Post by tachyon on Feb 6, 2016 20:58:31 GMT
[/quote] That's interesting, have you really done that graph? Impressive
13th sounds about right but after two top ten finishes, it would be underwhelming??
[/quote] I take each team's performance data after every game, adjust their defensive/attacking ratings, update the win/draw/loss odds for all remaining games, then simulate the remainder of the season 10,000 times using these odds. And finally add points "won" in future games to those already won in reality. Goal difference etc as tie breakers. The Graph summarises the results of the 10,000 simulations. I do it every week for every Premier League team, I just happen to be a Stoke fan. Agree 13th would be underwhelming as a fan. As a stats guy, there's a heck of a lot of statistical noise and randomness in a single season, so if we fail to match or better last season it would be totally unsurprising. There's more stats stuff on my blog in my signature. Many of the posts use Stoke/Pulis/Hughes to illustrate a particular point.
|
|
|
Post by tachyon on Feb 6, 2016 20:27:40 GMT
Latest projections for our final finishing position based on the remaining matches for all 20 teams and everyone team's performances so far. Attachment Deleted Our most likely finishing spot is now 13th and it's a 70% chance we finish outside the top ten & just 30% we end up in the top half. One chance in 500 we go down.
|
|
|
Post by tachyon on Feb 6, 2016 19:44:18 GMT
Based on the remaining fixture list and how each team has performed over the season so far it looks like being a typical relegation safety line. Attachment DeletedMost likely final points total for the final relegation spot currently stands at 35 points. So in that case 35 points could keep you up on goal difference. Only about a 1 in 50 chance that the 18th placed team gets 40 or more points.
|
|
|
Post by tachyon on Jan 31, 2016 22:35:35 GMT
Just a quick look at where we might have ended up after our first 23 games. The projections are based on the field position and type of attempt on goal taken by every team in every Premier League game played so far. Attachment DeletedThe first table shows our most likely current league position, again darker green = more likely, red colours less so and numbers are the percentage chance of each position being occupied by Stoke. So basically we're slightly more likely to be in the bottom half than the top half and more likely to be in the bottom three than the top five. Although neither extreme is that likely to have occurred. Second table looks at how many points we might currently expect to have. The range is from a high of around 37 to a low of about 24, with 31 points being the most likely. Sound about right?
|
|
|
Post by tachyon on Jan 31, 2016 19:08:06 GMT
Why is top more likely than second? [/quote] Not sure I follow. On the points chart, we've got a 1.6% chance of finishing with <42 points (so that's 41-40-39 down to our current 33 points) compared to a 0.8% of getting exactly 42, but we've got essentially a zero chance of finishing either in 1st or 2nd position.
|
|
|
Post by tachyon on Jan 31, 2016 17:25:09 GMT
I think you have to also factor in the continued trend in Hughes' managerial career that sees his teams perform better towards the end of the season. If that trend were to continue this season, then Europe is a real possibility. Hughes' first season at Stoke was the most extreme example of late season improvement. That could have been partly down to the side taking time to adjust to the change in style. If you include all his completed seasons as a Premier League manager his sides add a couple of tenths of a point per game over the final 15 games compared to the initial 23. Take out Stoke's transition season and the rate of increase more than halves. If the "late season Hughes" effect is real, it'll probably only add a point or two to the projections.
|
|
|
Post by tachyon on Jan 31, 2016 14:55:45 GMT
Lots of statistical predictive models doing the rounds based on goal attempts. Looking at Premier League and particularly Stoke with a dispassionate eye over the remaining matches we're most likely to get 50 points and finish in 10th spot. Attachment Deleted It's about a 2% chance we finish somewhere in the top 5 and 1 in 1,000 we get relegated. Darker the green, the more likely the outcome, darker the red the less likely.
|
|
|
Post by tachyon on Aug 10, 2015 8:41:38 GMT
|
|
|
Post by tachyon on Aug 8, 2014 11:19:00 GMT
|
|
|
Post by tachyon on Aug 10, 2013 20:22:25 GMT
|
|
|
Post by tachyon on Aug 10, 2013 20:17:25 GMT
|
|
|
Post by tachyon on Jul 11, 2013 8:24:28 GMT
It's a decent attempt to take a much more detailed look at Stoke's reliance on creating chances as close to goal as possible. The trouble was that once a side's methods become well documented, opponents then devise ways to combat even effective tactics, such as Rory's longthrow. So a side, even a relatively successful one, constantly needs to evolve to survive. The more times EPL regulars came up against Tony's Plan A, the better they performed against us. Here's a post about our record, sorted by the number of times our opponents had faced Stoke in the EPL linkThe parting of the ways or relegation was the only two likely outcomes.
|
|
|
Post by tachyon on May 14, 2013 17:42:25 GMT
Sean Haselgrave, John Ruggerio and John Lumsdon.....good spot!
|
|
|
Post by tachyon on May 14, 2013 15:23:40 GMT
Here's who I think they are from the start of the slideshow. Some I have no idea, some may be wrong. Dicko, Simmo,? Simmo,? Harry Burrows,?,? Harry Burrows. ?,Carruthers, Maskery, ? ? , Chris Iwelumo. Maskery, Big Chris, ? Lee Sandford, Griff, Clive Clarke,? LS,CC,Griff,? Dave Brammer,Carl Beeston,? Carl Saunders, Carl muggleton. Carl Saunders. ?, Brendan O'Callaghan, Huddy. ?, Steven Tweed, Dave Regis Denis Smith, Graham Shaw,? Regis, Smithy. Shaw. Doobs. Smithy, Regis. Mike Pejic. Nigel Gleghorn ? Pej. Biggins, Willie Stevenson, Andy Holmes. Stevenson, Holmes. Biggins. Ian Cranson, Kav. Kav Kav Phil Heath, ? Peter Thorne, ? Mike Sheron. Sheron. Chris Hemmings, Steino. Steino, Jackie Marsh. Marshy. Stein. Marshy, Hemming. Ade Akinbyi. Steve Waddington, Ade Gavin Ward, Wayne Thomas. Terry Conroy. Banks, George Eastham, Sir Geoff Hurst. Banks, Hurst. Ooooh Georgie Berry !!
|
|
|
Post by tachyon on May 14, 2013 7:12:05 GMT
|
|
|
Post by tachyon on Sept 29, 2012 19:42:15 GMT
|
|
|
Post by tachyon on May 13, 2012 21:55:17 GMT
|
|
|
Post by tachyon on Mar 20, 2012 9:20:41 GMT
How many goals have we actually scored from throw ins this season? I think that it's obvious that they are less effective now than they've ever been. Rory's throws were most effective in the promotion year and the first season in the Prem.He helped gets us up and then helped to keep us up. In the last two completed Premier league seasons goals scored from a Delap throw have halved in number and in value to the team.Although they have contributed greatly to our cup runs. I've quantified the importance of every Delap assisted Stoke league goal for the last five seasons in this blog post thepowerofgoals.blogspot.co.uk/2012/03/quantifying-delaps-long-throws.html
|
|
|
Post by tachyon on Jan 26, 2012 9:06:52 GMT
Last season the ball was actually in play for an average of 62 minutes per game in the Prem,one minute more than in Spain or Germany and 2 minutes less than Italy.
Stoke games had the ball in play for just 58 minutes,the lowest in the league.ManU had the longest time at 66 min.
|
|