Switching formations

One of the aspects of the game that always gets a bit of disagreement is the issue of switching tactics.

What the rules say:

You are able to prepare your team for how your opponent will play. If you think your opponent will play 4-4-2 you will be able to choose to prepare your team for that by selecting that formation in a dropdown list on the tactics page. The tactical option is possible to save within the default tactics.

If your opponent uses that formation this will have a negative effect on your opponent’s ability to create chances. How much effect it will have depends on how accustomed your opponent is to playing with that formation. That is determined by the formations used by your opponent the last 3 matches (including friendlies). The more matches (out of the past 3 matches) the opponent has played with a certain formation, the better he will be at playing that formation.

In other words this means that if the opponent uses a formation that he has used before and you didn’t prepare for that formation then your opponent will get a boost in his ability to create chances.

If you choose to prepare your team for a formation that the opponent won't use, then your opponent will always get a small positive extra effect on their ability to create chances. If you choose not to prepare your team for any formation then your opponent won't get any extra positive effect other than the experience they have on their formation used in the match.

Pros and cons of varying formations

If you switch formation each time it's likely your opponent will guess the wrong formation and you will gain an advantage because of that. The downside in this respect is you will never get a 3/3 formation so you don't get the maximum advantage available.

You also have the advantage of adjusting your formation to

This strategy may have worked well in the past but one of the big problems nowadays is so few formations are effective. 5 man mids (352/451) are the dominant formations and while 4 man mid formations can work they aren't as strong as 5 man mid. Anything else is just plain rubbish so you have limited options of formations you can switch too.

Advantages to keeping the same formation

One of the big advantages of playing just one or 2 formations all the time is familiarity and experience. You get to know what works so much better than when you are constantly switching all the time.

Another advantage of a single formation is there is no advantage to the opponent guessing how you will play, and if for some reason they guess wrong (it happens) then you gain a significant advantage.

Overall

Personally I tend to stick to just 1 formation with each of my teams. Varying it opens up the risk of the opponent guessing a 1/3 formation and almost always gives them a 2/3 formation advantage and I would much rather remove this variable. I can build my squad around the formation and get familiar with what attacking styles work.

How FT & form work

People often wonder how the form can seem to twist and turn in an unfathomable way - here's an attempt at exposing why.

The first graph is a simple sine wave type graph. Imagine that is your underlying form, bouncing up and down in a reasonably ordered manner.

Player's form will be adjusted maybe 1-2 bars based on performance of the team section they are in and how they performed during games. This 2nd graph is a series of random changes meant to represent what happens when you play games. Mostly they are small changes between +-1.

If we overlay these 2 sets of data we get graph 3. The general trend is the same as graph 1 but at specific points it looks different. EG point 5-6 is flat and looks like the form has stopped going up, but this is caused by a bad performance. Points 8-10 show the same misrepresentation. Conversely point 11 shows a marked increase at exactly the point when form begins to drop. Also note that as the effect of performance is relatively small players form can still go up after a bad performance and vice versa.

What you see with the X11 VIP graph is something akin to graph 3. The secret to preemptive training is taking this graph, guessing the effect of the match (the data from graph 2) and unpicking them and enabling you to see the data in graph 1.

Goalkeepers and penalties

A while back I did some statistics on free kicks, penalties etc - and one area I looked into was keepers saving penalties. Cool is supposed to be the SQ for saving penalties so the first port of call is checking where the keeper has the cool SQ and comparing that to keepers without cool.


The following table shows penalties against keepers with and without the cool SQ (those without can have any other SQs). The first column is the difference in performance (not skill) of the player taking the penalty and the keeper. The bottom 2 lines are total for all penalties and total for situations where the keeper and penalty taker seem to be of similar skill (the greyed area)

	Cool Keeper			No Cool
Perf Diff	Scored	Missed		Scored	Missed
-10	0	0	0%	1	2	33%
-9	0	0	0%	1	1	50%
-8	0	1	0%	0	0	0%
-7	1	0	100%	2	8	20%
-6	0	0	0%	2	6	25%
-5	1	0	100%	8	7	53%
-4	1	1	50%	5	5	50%
-3	2	2	50%	9	10	47%
-2	2	1	67%	16	10	62%
-1	2	1	67%	24	9	73%
0	3	2	60%	34	6	85%
1	2	1	67%	45	4	92%
2	6	0	100%	56	13	81%
3	10	0	100%	53	12	82%
4	6	0	100%	54	4	93%
5	5	0	100%	55	1	98%
6	2	0	100%	48	1	98%
7	2	0	100%	56	2	97%
8	5	0	100%	39	0	100%
9	2	1	67%	22	0	100%
10	2	1	67%	15	0	100%
Total	54	11	83%	545	101	84%
Tot -3 to +4	33	7	82%	291	68	81%

Overall there are 65 penalties against keepers with cool , and 83% are scored. 646 against other keepers - and 84% score. The situation is similar for the players of similar skill - and the conclusion is cool doesn't seem to do anything to help keepers save penalties. 

I then checked other SQs and most of them (king of the air, breakaways, intelligent) looked similar to above. The other 2 SQs seemed very different - see the following table. The left hand side is for reflex SQ, right hand side for Positioning SQ.

	Reflex SQ			Position SQ
Perf Diff	Scored	Missed		Scored	Missed
-10	0	0	0	0	0	0
-9	0	0	0	0	0	0
-8	0	0	0	0	0	0
-7	0	0	0	1	0	50
-6	0	2	0	0	1	0
-5	1	0	50	2	1	50
-4	1	0	50	1	0	50
-3	1	3	20	1	0	50
-2	3	2	50	5	0	83
-1	3	1	60	5	0	83
0	2	0	66	5	1	71
1	2	1	50	10	0	90
2	5	1	71	10	0	90
3	7	4	58	12	1	85
4	6	1	75	15	2	83
5	9	0	90	4	0	80
6	2	0	66	8	0	88
7	12	0	92	6	1	75
8	7	0	87	7	0	87
9	4	0	80	3	0	75
10	3	0	75	2	0	66
Total	68	15	82%	97	7	93%
Tot -3 to +4	29	13	69%	63	4	94%

The overall result for reflexes seems skewed because there are a lot of penalties by relatively high skilled kickers but looking at players of similar skill only 69% are scored - compared to a typical 83%. SQs often work as opposites so positioning has the opposite bonuses and penalties to reflexes and that us very apparent in this table. Keepers with positioning almost never save penalties. 

Sample size & conclusion

I think a sample of around 700 penalties overall is reasonable to show the conversion rate is around 84%. For specific SQs the sample sizes are small but I would say in all instances they are probably adequate enough to show what is actually happening. I had planned to collect more data but - never happened.

It seems very much like the cool SQ does nothing to help saving penalties and that reflexes is the SQ to have in this situation - and positioning is the kiss of death. The fact that positioning has the opposite effect is logical from how X11 SQs work but does beg the question why are the rules wrong?

Change report - actual data

The following charts show actual Development Value data I've collected from players in my teams along with the amount they changed by at the change report. For each DV value in the main age bands I've identified how may players gained the corresponding amount of skill at the change report. There is also the average change for each given DV value within the skillband.

Ages: 16-20:

	CR		0.9	1	1.1	1.2	1.3	1.4	1.5	1.6	1.7		Ave
DV
15				3									1.00
16			1	3	1	3	1						1.10
17					2	5	14	3					1.28
18							8	17	11	7			1.44
19								3	2	5	3		1.56

Ages: 21-22:

	CR		0.8	0.9	1	1.1	1.2	1.3	1.4	1.5		Ave
DV
14			1									0.80
15			3	1								0.83
16				1	3							0.98
17					2	3	5	4				1.18
18						1	5	3	2	1		1.28
19									1	2		1.47

Ages: 23-25:

	CR		0.6	0.7	0.8	0.9	1	1.1	1.2	1.3		Ave
DV
15			1	3	1							0.70
16					1		4					0.96
17						2	8	10	4			1.07
18								3	2	1		1.17

Substitutions

In an earlier blog I wrote how minutes played affected Player DV gained and how using substitutions can be used to increase DV over the season.

Here's a tip for substitutions that may prove useful in a number of situations.

Double substitutions

Sometimes you want to do a double substitution - bring one player on then bring the same player off later. Here's one of my teams and I have 3 young midfielders.

I want to start the game with Luxumburgo, bring on Sampaio after 45 minutes then swap him for Daish after 70 minutes. The system won't let me do this because Sampaio isn't a started so I need to get sneaky.

I set the team up normally and go to substitute Sampaio on for Luxemburgo after 45 minutes - that's all normal.

I then go and alter my lineup to put Sampaio in the team. The substitutions page will now let me substitute Sampaio out so I put my 2nd substitution in. I then go back to lineup and set it back to the original settings. Voila

Youth Academy Musings

Following on from the previous article giving a series of facts about the youth academy I'd like to present a load of speculation. Some of this may be correct - I expect a lot of it is, but some of it will be wrong. You have to choose what to believe and if you use it as the basis for any decisions.

Skill adjustment by age

We  do know that older players come into the game at higher skill levels. The vast majority of 16yo player are 2-4 skill. At 20 that range is 4-6 skill. It's not unreasonable to guess that players gain .5 bar per year after 16 but always bound between 2-6 skill. From a coding point of view it's simple too.

Looking at the 17yo graph I did originally wonder that it would be nice to have a similar graph for other ages - but does it really matter? In the original YA information was also the following graph:

This shows the distribution by skill according to various investment levels. Even for a 300k academy there are only about 4% of players that are 6 skill, and these will mostly be 19/20yo. If you look at 5 skills it's a decent proportion but 20/5 - worthless, 19/5 - rarely worth much, 18/5s are usually worth similar or less than 17/4s. Essentially if a player isn't 16/17 then they have limited value.

Musings

Looking at the skill distribution for 17yo players it looks very much like a parabolic/bell curve BUT as the data is rounded to integers and the range only covers 4 meaningful points then the curve is very much distorted.

The way the datapoints settle is a bit bizarre. Averaging them it shows each increment in the YA increases average skill levels of players by 0.1 or 0.2 bars but it's not a straight add. If you look at the values for 150k investment for example there are around 11% of players at 4 skill. 4 skill covers the values from 3.5 to 4.5 and assuming roughly equal numbers of players at each skill level you would expect that a 0.1 bar increase across the board (upgrading to 200k academy) would convert 1/11 of these players to 5 skill but that doesn't happen.

What does happen at all levels of the academy is 3s become 4s, but very little in the way of 4s becoming 5s. Higher levels of the YA seemingly just get rid of the dross but don't offer a lot in the way of primo players.

Conclusion

Looking at it as a whole you have to conclude the YA is about 17/4s for the most part - and here's the bottom line of the matter: Player age/skill is rarely what is going to make a player valuable. It's about what SQs a player comes with. A vanilla 17/4 has a an estimated value of under 500k - add 2-3 SQs, and/or hard trainer and he can be approaching 10m. What the purpose of the YA seems to be is to ensure that when a youth with nice SQs does come along he is a 17/4 and worth good money rather than being a 17/3.

In summary I think the YA is a bit of a swizz. The idea that was pitched was that you get more "super" players, and the data doesn't seem to support it. What you seem to get is less dross, more average players, and the big wins are likely to be on the 17/4s with good SQs rather than 16/4 path players. Great if you want the YA to fuel a team, but not good if you are hoping for those superstars to support other team building approaches.