Out of Our Hands: Match Projections for Week 37

Tottenham Hotspur were "in control of their destiny" leading into Wednesday's match against Chelsea. A hard-fought draw demonstrated a simple truth--it doesn't matter much whether you're in control of your destiny. It matters whether you have (a) a lead in the table, (b) an easier schedule and (c) a better team. Arsenal didn't have (c), but they had more points and they didn't have to play at Stamford Bridge. The projections had Arsenal well ahead of Tottenham going into Wednesday's game, and the outcome kept Arsenal in good position.

At the same time, now Arsenal are in control of their destiny. Which matters right up until the point they don't win a game. Obviously Arsenal haven't been dropping points to lesser competition for a while now, but they also haven't been dominating opposition in a way that suggests they're better than the projections say. We'll have to wait until Tuesday to see them face a possibly exhausted Wigan Athletic side.

The State of the Race

Spurs stand three points behind Chelsea and one point behind Arsenal. It would take a miracle for Spurs to pass their London rivals in goal difference. This means Spurs need either four more points than Chelsea or two more points than Arsenal. If Chelsea win once, they're safe. If Arsenal win twice, they're safe. The other winning scenario for the Gunners and the Blues involves Tottenham dropping more points, which is likewise entirely possible.

There are a number of ways of visualising the race. Here's one. These are the chances of Tottenham finishing top four depending on how many points they take from their final two games.

Points	6	4	3	2	<2
Probability	36%	26%	25%	4%	9%
Top 4%	71%	24%	11%	4%	0%

If Spurs can take the maximum six points from this and next week's games, we'll project as reasonable favorites (71%) to finish in the top four. The problem is that Tottenham is far from certain to win both games. We'll be favorites, but I have six points from six as only about a one-in-three chance. As fans, we should not be counting on six points from six. Clubs drop points all the time.

Of course, the same is true of Chelsea and Arsenal. For Spurs to make the Champions League, we need to win our games, and we need either Arsenal to drop at least two points or Chelsea to drop at least four points. I have the chance of Arsenal dropping two points or more as 64%, and the chance of Chelsea dropping four points or more as 18%. So I'm telling you there's a chance.

Match Projections

Of the top four contenders, Spurs have clearly the toughest match this week. Playing at the Britannia is never easy. One mildly positive note is that Stoke City have effectively clinched safety. Stoke are six points and 13 goals difference ahead of Wigan. They're not going down. Of course, they haven't entirely, mathematically clinched, and it's unknowable to what degree their near-certain safety will affect their play.

Chelsea face an Aston Villa side who are very nearly clear of the bottom three. Not quite. They've also been playing some excellent football recently. Of course, they have about 1/5 the talent and 1/50 the payroll of Chelsea football club. Arsenal face Wigan, who are big favorites to go down but not dead yet.

Club	G	GA	W%	D%	L%
Arsenal (v WIG)	2.7	1.2	69%	16%	15%
Chelsea (@ AVL)	2.1	1.3	56%	21%	23%
Tottenham (@ STO)	1.7	1.0	49%	25%	26%

I also have the projected chances of finishing top four based on this week's outcomes. To this week's table I've also added Wigan--it lists for them not the chances of finishing top four depending on the outcome of their match at the Emirates (that would be 0% 0% 0%) but their chances of remaining in the Premier League.

Outcome	W	D	L
Arsenal (v WIG)	85%	46%	39%
Chelsea (@ AVL)	100%	92%	82%
Tottenham (@ STO)	57%	17%	8%
Wigan (@ ARS)	78%	37%	20%

Arsenal, Spurs and Wigan all need wins pretty desperately. Chelsea can probably eat a loss and still survive.

Juking the Stats

One thing I noticed when I was running the power rankings is that the two clubs with the widest divergence between their overall quality on the season and their post-January performance are Wigan and Aston Villa. Both Sunderland and Stoke, by contrast, have been pretty terrible in the second half. I can make the numbers look a little better for Spurs, without completely casting good faith out the window, if I use the form numbers.

So I re-ran all the numbers, using post-January form as the basis for the projections instead of full-season team quality. The differences aren't stark, but they are notable. First the table for Spurs' chances of finishing top four based on how many points they take from the next two matches:

Points	6	4	3	2	<2
Probability	47%	25%	19%	3%	6%
Top 4%	72%	25%	13%	3%	0%

The chances of a six point sweep increase significantly. That's good! How about the expected game outcomes?

Club	G	GA	W%	D%	L%
Arsenal (v WIG)	2.4	1.4	60%	19%	21%
Chelsea (@ AVL)	2.1	1.7	48%	21%	31%
Tottenham (@ STO)	1.9	0.6	61%	22%	17%

Stoke City have been just astoundingly poor in attack since the new year. The chances of good outcomes for Spurs increase.

The probabilities of clubs making top four based on different outcomes don't change much. But what about the overall projected table? Can we make that look a little bit better? At least I can try. One last table for today, the projected table based on the form ratings. In this table, the "delta" column doesn't list change from last week, but change from the projected table based on full-season team ratings.

Club	W	D	L	Pts	%Top 4	ΔTop 4	%Rel	ΔRel
Chelsea	21	10	8	73	93%	-1	0%	0
Arsenal	20	11	7	71	65%	-7	0%	0
Tottenham Hotspur	20	10	8	71	43%	+9	0%	0
Mid-table	-	-	-	-	-	-	-	-
Sunderland	10	11	17	40	0%	0	5%	+1
Newcastle United	11	8	19	40	0%	0	11%	+1
Norwich City	8	15	15	39	0%	0	20%	+6
Wigan Athletic	10	8	20	38	0%	0	63%	-6

I don't know if shifts of five or ten percentage points matter much for anyone else's mental health. This exercise gave me just a touch more optimism, even though I know I'm cherry-picking an analytical method based on whether I think it will project slightly better numbers for my team. I did, and then it did. Win win.