Quick Preview of Chelsea

You may have noticed, but Chelsea are a little different this season. As I wrote here, they've had a very impressive recent streak which has lifted them to the top of the table. The results have obviously been a big change after a season of struggling in midtable, but the way they've gotten them has been different too. Conte has installed a 3-4-3 formation, and exiled from the starting eleven many key contributors from their championship team. Below are their usage rates for the season so far:

Name	Successful Passes	Total Passes	Unsuccessful Passes	Key Passes	Assists	Goals	Shots	Take-Ons	Take-On Success %	Unsuccessful Take-Ons	Minutes	Possessions Used	Usage Rate	Positive Outcomes	Negative Outcomes	% Positive
Costa	275	372	97	18	4	10	41	25	35	46	1144	206	15.62%	63	143	30.58%
Hazard	534	632	98	26	1	7	36	53	72	21	1077	182	14.66%	63	119	34.62%
Alonso	258	341	83	10	1	1	13	2	33	4	666	111	14.46%	24	87	21.62%
Oscar	199	239	40	14	1	0	12	1	11	8	462	75	14.08%	27	48	36.00%
Moses	257	321	64	7	1	3	16	20	61	13	669	101	13.09%	24	77	23.76%
Willian	225	253	28	24	0	2	14	17	65	9	511	75	12.73%	38	37	50.67%
Pedro	175	210	35	6	5	3	14	14	61	9	577	69	10.37%	25	44	36.23%
Ivanovic	256	307	51	3	0	0	7	6	67	3	563	64	9.86%	10	54	15.63%
Matic	683	773	90	8	6	0	6	9	43	12	1135	122	9.32%	20	102	16.39%
Kante	767	857	90	9	0	1	15	16	67	8	1170	122	9.04%	24	98	19.67%
Azpilicueta	714	805	91	5	0	0	5	6	86	1	1170	102	7.56%	10	92	9.80%
Luiz	320	376	56	3	0	0	10	4	100	0	810	69	7.39%	13	56	18.84%
Terry	234	256	22	2	0	0	4	0	0	0	370	28	6.56%	6	22	21.43%
Cahill	580	657	77	2	0	1	6	1	33	2	1163	87	6.49%	8	79	9.20%
Courtois	109	164	55	0	0	0	0	0	0	0	1170	55	4.08%	0	55	0.00%

As you can see, the wing-backs are very involved in this formation. Moses and Alonso both have usage rates that are usually reserved for true wingers and use their possessions pretty efficiently to boot. Their late-breaking runs to open space on the far side after an overload on the near side are truly dangerous, as Spurs found out last weekend. That will be something City need to keep an eye on. I would definitely think a back four is the right play here, as I would not want to rely on some combination of Sterling/KDB/Sane/Navas to track back on them.

The possessions the wing-backs are using had to come from somewhere. Let's take a look at last year's usage rates for comparison:

Player	Successful Passes	Total Passes	Unsuccessful Passes	Key Passes	Assists	Goals	Shots	Take-Ons	Take-On Success %	Unsuccessful Take-Ons	Minutes	Possessions Used	Usage Rate	Positive Outcomes	Negative Outcomes	% Positive
Fabregas	2278	2700	422	59	7	5	45	23	55	19	2900	552	15.98%	111	441	20.12%
Oscar	837	1033	196	26	3	3	52	26	42	36	1774	313	14.81%	81	232	25.89%
Hazard	1204	1403	199	63	3	4	36	89	62	55	2196	356	13.60%	102	254	28.69%
Willian	1220	1473	253	75	6	5	60	84	67	41	2740	435	13.35%	141	294	32.39%
Costa	641	824	183	24	6	12	68	40	37	68	2374	349	12.35%	98	251	28.07%
Pedro	809	966	157	40	2	7	57	52	58	38	2036	294	12.11%	99	195	33.71%
Kennedy	248	288	40	6	0	1	13	26	62	16	560	75	11.24%	19	56	25.36%
Ramires	244	306	62	1	0	2	10	6	40	9	616	82	11.18%	11	71	13.41%
Azpilicueta	1438	1766	328	17	3	2	10	6	30	14	3202	372	9.76%	30	342	8.06%
Matic	1387	1577	190	17	2	2	34	38	72	15	2421	258	8.94%	53	205	20.56%
Rahman	427	513	86	2	1	0	6	8	73	3	1023	98	8.04%	9	89	9.19%
Ivanovic	1051	1272	221	11	2	2	28	10	67	5	2916	267	7.69%	41	226	15.36%
Zouma	578	720	142	0	1	1	17	4	57	3	1918	163	7.14%	18	145	11.04%
Courtois	138	303	165	0	0	0	0	0	0	0	2018	165	6.87%	0	165	0.00%
Mikel	1091	1218	127	4	1	0	7	20	87	3	1793	142	6.65%	12	130	8.45%
Cahill	624	751	127	2	1	2	13	3	60	2	1957	145	6.22%	16	129	11.03%
Begovic	96	184	88	0	0	0	0	0	0	0	1398	88	5.29%	0	88	0.00%
Terry	824	917	93	2	0	1	10	2	50	2	1987	107	4.52%	12	95	11.21%

Fabregas, Oscar, and Willian aren't featuring as much this year, and all of them had very high usage rates last season. It's Fabregas who is the most interesting case, as he had been a staple of the Chelsea team the past two years as a central midfield creator. This year it's primarily just Kante and Matic in the center behind Costa, and while Matic has contributed six (!) assists so far, I think that assist rate unlikely to continue. The usage rates match the eye test, which says they are functional distributors and defenders, but they leave chance and shot creation to others. Once again, assuming Conte sticks with the same 3-4-3 formation, this means City should probably concentrate on the wide areas when defending.

As for attacking, the 3-4-3 can leave gaps behind the wingbacks certainly. I think Sterling in particular will be important, as he will likely be attacking down Chelsea's left versus Cahill (arguably the weakest of Chelsea's center-backs) and Alonso (who doesn't have the speed to touch him). We'll also need KDB and Silva/Gundogan to be at their very best to not be swallowed whole by Kante and Matic, as that will be key to City establishing their attack. That said, Chelsea don't have a lot of holes on defense, allowing the second-fewest shots on target per Opponent's Possession.

This is by far the toughest Premier League test City have faced so far and I have to say I am not super confident of a City win here. Hopefully the full week of rest will help rejuvenate some players (particularly Fernandinho, who has still played every PL minute) and we'll be able to put out a good performance. No matter what happens though, it should be fascinating to see how the two teams match up.