Ched

Right, to give this thread the background it requires i will first outline what i consider to be the main controls on finishing (this will become relevant later on, those of you who are familiar with my posying style will know this could go on...);

1 - Player ability - obviously a league 1 striker will score fewer than a premiership striker against opponents of the same ability.

For the purposes of this experiment i assume that the chances are being created for the striker, and as such, the most relevant stats should be, finishing, composure and decisions.

2 - Tempo - if i play a high tempo, i would expect chances to be rushed, if i play a low tempo, i would expect chances to be taken when the best opportunity presents itself - granted this is linked to decisions.

3 - Time wasting - some users suggest that this controls how long a player dwells on the ball as well as things like slow free kicks, hences, high time wasting should link to a low tempo, and provide more clinical finishing.

4 - Mentality - an attacking mentality should be focused on getting the ball forward - this should not create the best chances, just more of them. A more defensive mentality should wait for chances to appear, rather than forcing the game.

Right, that was just to outline my train of thought when comparing finishing rates in the following, and to point out the basic sort of tactics employed.

Llamas look away now

In a nutshell, the tactics involve low tempo, short passing, marginally above norm mentality, and mid-high time wasting.

Llamas can come in again

The point of this? I've seen several threads complaining that the AI "cracks" their tactics, and the less educated threads along the lines of "the AI cheats". Due to the fact i usually employ a 3 formation rotation i did not notice this greatly, but there were uncharecteristic slumps that i had no way of explaining. So i decided to investigate. Before an SI employee comes on here telling me i should just play the game, i would do except it's not much fun right now.

So, i started a league as chelsea (holidayed one season to remove the interferance of the african cup of nations - would mess up results spectacularly) and used only the one tactic.

Until Jan 30th - results were - played 23, won 23, scored 55, conceded 5.

Average Human Shots per game: 27

Average Human Shots on target p/g: 16.6

Average Human Goals p/g: 2.4


Average AI Shots p/g: 8.5

Average AI SOT p/g: 5.2

Average AI Goals p/g: 0.22

Average possesion was 62% in favour of the user.

Compare and contrast the above results with what happened between february and march (it all seemed to even out again in april after the league had closed up a bit).

Average Human Shots per game: 21 - 23% less than pre feb

Average Human Shots on target p/g: 11.8 - 29% less than pre-feb

Average Human Goals p/g: 0.85 - 65% less than pre-feb


Average AI Shots p/g: 7 - 18% less than pre-feb

Average AI SOT p/g: 3 - 43% less than pre-feb

Average AI Goals p/g: 0.6 - 172% more than pre-feb

Average possesion was again 62% in favour of the user.

So what can be deduced from these results?

Firstly, it must be said that the quality of chances was broadly similar throughtout the season (due to the fact the AI scored so infrequently it was easy to review each goal).
Secondly the pattern continued in cup competitions, however, due to the vastly fluctuating quality of teams faced, i chose to omit cup results.

So in essence, when the AI *cracked* my tactics, the following happened;

A) The user had vastly more possession than the AI.

B) The number of chances for both teams decreased.

C) The users finishing decreased markedly.

D) The AIs finishing incresed by 170%

E) The user still had SUBSTANTIALLY more SOTs than the AI, from positions of similar quality.

Now, points B and C are what i would WANT to see form the AI as it countered a succesful tactic. In principal, the AI should try and close games down and make them as tight as possible.
My main issues lie with the fact that despite finishing ability of the strikers remaining constant (and morale remaining high in the case of user and often low in case of AI) the % of chances converted was phenomenally increased for the AI.

Possible explanations:

1) - Complacency. It could be argued that having won so many games in a row the users players lost the desire to win. I pick ONLY players with high work rate and a determined personality to specifically avoid this.

2) - Weather. The reduction in chances created could be due to the poorer weather during feb-march, however this does not explain the changes in finishing.

I am all for the AI figuring out how to beat tactics - it's realistic above all - HOWEVER, just for once, i would like to be outplayed, rather than the finishing ability of agbonlahor, aliadiere et al suddenly rocket.

If anybody wants to retry this experiment i can upload my tactics (or maybe if anybody just wants a devastatingly effective tactic for 6 months lol).

Appologies if it seems that all my threads are monas, it's just that i would like to see the game progress, and i feel that constructive criticism is the only way to achieve this.


Again, a well done to those that read this.
All useful comments would be appreciated.

Cheers

soundian

1) It's not unknown for teams to have a mid-season slump.
2)Strikers can go off the boil.
3) Maybe the more direct approach by lesser teams is better in wintery conditions.
4) One season<>statistical proof.
5)Some of your initial assumptions may be wrong.
6) My perceived slump normally happens in January and I often waltz through Feb and March

deadeyedunc

I completely agree that unexplainable slumps happen and have experienced the same things as your results suggest. Tactics that have been working well give the same stats (possesion, shots, etc) but give very poor results, how are we supposed to overcome that?

heathxxx

A constructive and well presented post Ched.

I have to say, I have not taken an equally detailed exploration into this as you have, though I have noticed some of the elements you mention.

Perhaps though I'm fortunate, or possibly even ignorant of these sort of issues, as on the whole it does'nt affect the overall outcomes and enjoyment of the seasons I play.

Constructive criticism/feedback is a good thing, as it will undoubtably be noticed by SI and possibly help them with development for future patches and releases. Without such input, some problem areas might never be seen by the developers, who in fairness are never going to personally explore the fullness of the game in testing, as we the players do.

I only wish that more discussions regarding existing or potential gameplay issues were as thoughtfully posted as your own. :thup:

hyperion479

Well thought out post and analysis, but I'd suggest that you're slump is due to the combination of the weather, English fixture congestion around that time of year, and complacency.

Bad weather greatly hampers a more skillful tactic because crisp passes get blown off course or get bogged down in a slushy pitch. You might be able to get around this by changing to a more direct attack.

But the bad weather also requires your players to exert more energy to get the same level of play, hence tiring them out faster. Combine this with the huge fixture congestion in the Dec to Feb months in the English schedule, your players begin to wear down.

Combine this with the physical exertion needed to compete on four fronts, your players begin to struggle with the rigors and stresses of all the competitions. I would imagine that given a decent lead in the league, your players are willing to let a few games slip due to a less than complete physical/mental effort rather than wear themselves out beyond the point of exhaustion.

Katarian

I don't think I have had a February slump of form in almost 13 seasons. My slump usually happens long before February starts and the team is usually improving once they hit February.

November and December are my worst months, the Christmas/New Year schedule in the English leagues is always a nightmare for me. Too many players get injured due to the number of games. In my current season I had to play 3 games without any strikers as all my forwards at the club were injured, luckily one of my wingers stepped up and scored 6 goals in 4 matches .

Jimbokav1971

Ched.

1. What about fatigue? Isn't that another reason why there might be a slump in form?

2. Injuries linked a little to fatigue might play their part. For all their World Class players, Chelsea do not exactly have a deep squad in certain areas compared to the likes of the other big 3.

3. In stage 2 you won all 23 games but only scored 55 goals. That's not too many for 23 out of 23. Is there anything to be read into that?

4. The only stat that you posted that would concern me is ..... Average AI Goals p/g: 0.6 - 172% more than pre-feb All the others make senst to me, but this one just seems wrong.

Sorry I have more questions than answers. I must admit that this isn't something that I've come across.

scousebr

I didn't play this game enough to actually debate about it, but the high percentage here has to do with the number being small (0.22 to 0.6). Since the number you'll be dividing by is quite low, it'll inflate the percentage. Not saying it doesn't happen or that it isn't an issue, but it just looks way worse than it is.

Jimbokav1971

That's a REALLY good point.

Thanks:thup:.

So ignoring the % increase it's an increase from 0.22 goals per game conceded to 0.60 goals per game conceded.

If I look at these 2 stats in isolation I would say that the first figure is the one that is wrong.

As a result of that, I would possibly suggest that rather than the AI *crack* your tactics as seems to be the popular claim in GQ, isn't it more accurate that it's actually sorting it's own tactics out?

Ched. Have you got any other titbits of information that you have held back, such as AI formation and generalpattern of play?

Jimbokav1971

When I say wrong, I don't really mean wrong, I mean surprising or at least not what I would have expected.