We've about averaged that drop off the season after we've finished second, since 02-03 though. If I remember correctly, 16, 14, 23 and 21 points respectively.
Picked this up off reddit courtesy of
manateesherrif who compiled data from understat.
"At the end of last season, fans of Liverpool, Spurs, and Arsenal all had reason to feel good about their clubs. Liverpool had their best season ever in the Premiership, finishing with a record 97 points. Tottenham, meanwhile, had another solid season, solidifying a spot in the top four and making a run to the Champions League Final. And Arsenal had an encouraging first season under Unai Emery, fighting for a top four spot until the last day of the season.
And yet, some analysts would tell you that all three clubs were incredibly fortunate last season. Based on the chances they created and allowed, all three should have finished at least 10 points lower in the table. Could that data be right? Are all three clubs poised for regression this season? Let's take a look.
For those of you who don't know, expected goals (xG) is a way of measuring the chances that a team creates. By looking at xG and xG allowed, we can get a better idea of how well a team is actually performing. Sometimes, hot finishing or a streaky goalkeeper can make a team look better or worse than it actually is. Those hot streaks usually run out, but chance creation is much more consistent.
Last season, the data suggests, Liverpool should have finished with 13.55 fewer points than they actually did, Arsenal should have finished with 11 fewer, and Spurs should have finished with 9.56 fewer. But is this sort of data actually predictive? To find out, I looked at all the teams in the last five years who have outperformed their xPoints by 10 or more, and then looked at their performance the next season.
[View formatted table]
Year Club Real points xPoints Following season Change
14/15 Chelsea 87 75 50 -37
14/15 Swansea 56 43 47 -9
14/15 Tottenham 64 49 70 +6
15/16 Leicester 81 69 44 -37
15/16 West Ham 62 50 45 -17
16/17 Chelsea 93 76 70 -23
16/17 Spurs 86 75 77 -9
16/17 Arsenal 75 64 63 -12
17/18 Manchester United 81 62 66 -15
17/18 Burnley 54 41 40 -14
18/19 Liverpool 97 83
18/19 Tottenham 71 61
18/19 Arsenal 70 59
Look out! Teams that significantly outperform xG crash down HARD the following season. The good news for Liverpool is that even if they drop down to their "expected" 83 points, they would still be a very good team. Arsenal and Tottenham have more to worry about; xPoints has Arsenal in seventh, one point behind Wolves.
It's worth noting that there is one team in this list that did manage to avoid implosion the year after making this list. In 2015/16, Spurs were actually 6 points better than their "lucky" season the year before. So what did Tottenham do differently than all these other teams? Well, they had arguably their greatest transfer window ever. That summer, they picked up Dele Alli, Heung-Min Son, and Toby Alderweireld. They also swapped out the CM pairing of Ryan Mason and Nabil Bentaleb for the dominant duo of Mousa Dembele and Eric Dier. Their actual point total only increased by 6, but their xPoint total went up by 20. They were playing much better football.
That should be the lesson for the three "lucky" teams from last season. Don't look at your point total and think the team is good enough. If you don't move aggressively in the transfer window and address your weaknesses, your team is going to slip back to where they probably should have been all along. But if you scout well and act decisively, there is a chance to turn some of your "lucky" points into points that are well-earned."