Same match stats, but AI scores more?

[DeanMon]

22 minutes ago, santy001 said:

It's not my ME but you're trying to create a strawman of sorts because I never claimed the ME is perfect.

Also, if you mismanage your team... then yes there are going to be consequences to mismanagement. Making a mistake, or rather a series of mistakes, and getting an adverse result isn't a discrepancy. Bears no relevance to the topic at hand, but you might as well complain that sticking your hand in a fire burns. The AI too makes mistakes and it gets punished for them. 

---

There's only two ways in which it can happen, either the AI has gone rogue and is far more sophisticated than any other video game AI in existence or you're saying SI have counter-intuitively and counter-productively designed the game against all their aims as a business and are doing it to cheat you. 

Or human can't see AI's match preparation, but who knows if AI can. If I instruct my team to prepare defend ground and defend disengaged, then in match AI's Salah crosses to Mane to score by a header, which leads to a discrepant result even if all things are equal.

Edited July 15, 2020 by DeanMon

[santy001]

So who benefits from cheating you in this way @DeanMon?

What or where is the benefit?

[DeanMon]

Let me put it like this, SI does not intentionally cheat, but your imperfect ME (the black box whose operations are only revealed in anecdotes to the public) creates the favor to the AI.

AI makes all the PERFECT choices with the BEST effect: match preparation, TIs, pep talks, shouts,... 

While to human, for example, with the same pep talks, sometimes it works to full team, sometimes it does not work to anyone.

There might be many more which I can't think of right now... But all the little details accumulate to an advantage for AI.

What do you say?

[HUNT3R]

37 minutes ago, DeanMon said:

@HUNT3R

Yeah, such data speaks more than thousands words. May you post data of all teams?

There is one more thing is that you games must not be reloaded for any, because if they were, then the data is cherry-picked which will be skewed in your favoured. Hofenheim's data must be verify for that fact as well.

[HUNT3R]

2 minutes ago, DeanMon said:

AI makes all the PERFECT choices with the BEST effect: match preparation, TIs, pep talks, shouts,... 

You seem to start at the conclusion and then create a story to fit that conclusion. The above is just terribly incorrect.

If they made 'perfect' choices, their tactics would be perfect, which it isn't. That's apparent when you see some of the roles and duties used. You'd see players who are never nervous or complacent, which isn't the case. You'd never see "overwhelmed by the feedback", which isn't the case. All of these are easily seen when playing the game, so it's not even hidden knowledge or empty words.

 

[DeanMon]

Please let me replace PERFECT/BEST by MOST OPTIMAL. AI makes the MOST OPTIMAL choices with the MOST OPTIMAL effect.

With the same pep talks, sometimes it works and sometimes it does not work for human, but AI changes the pep talks choice to the one which works.

Yes, I saw AI's players are nervous complacent, that's why I wrote MOST OPTIMAL now. AI only needs its players nervous and complacent LESS often than human, then it is an edge for AI.

[yolixeya]

5 minutes ago, DeanMon said:

 but AI changes the pep talks choice to the one which works.

Earlier in this thread, you said to someone that he makes a lot of assumptions. This is what you do all the time with claims such as these. You have zero evidence of this and you just make this up to suit your story.

[Svenc]

1 hour ago, DeanMon said:

Let me put it like this, SI does not intentionally cheat, but your imperfect ME (the black box whose operations are only revealed in anecdotes to the public) creates the favor to the AI.

AI makes all the PERFECT choices with the BEST effect: match preparation, TIs, pep talks, shouts,... 

While to human, for example, with the same pep talks, sometimes it works to full team, sometimes it does not work to anyone.

There might be many more which I can't think of right now... But all the little details accumulate to an advantage for AI.

What do you say?

Let me rephrase this, and I'd probably agree.

The AI isn't as stupid as in older editions. 

It won't leave its back doors uncovered on like every other attacking corner (check my first post in this thread and the little Leicester story).

It may not make stupid stuff like this anymore as oftenly.

It may also time its decision a bit more competitively (say when losing a match not merely finally start attacking too late into the match).

Therefore, its performance may have increased some (though players such as @Rashidi can still outperform it massively -- and do).

That's not rigging naturally, that's coding more competitive AI for some added challenge, in parts based on the feedback received throughout the years. Even if SI could (and they can't as that would require an AI that would be able to "read" the second by second match play just as a human can), Si would never code the perfect AI anyway. Coding the "perfect" AI would a) would alienate players b) make the purist main coder go all mad again due to the increase in cheating accusations.
 

Edited July 15, 2020 by Svenc

[zyfon5]

4 hours ago, DeanMon said:

Hello, there might be a misunderstanding here. Let me clarify:

1) Many complain that why human does not often win despite dominating the match. Many "excuse" that human can't score because human's tactics are "too overwhelming" AI's tactics, which may be true.

You can overwhelm the AI and win at the same time as long as you know what you are doing.

4 hours ago, DeanMon said:

2) Now my argument is that: when AI's tactics is "too overwhelming" human's tactics (AI Liverpool), it seems to score just fine.

The poster said AI's tactics are not as overwhelming as human's tactics, that is why AI can score. I said it is wrong, please check the current AI Liverpool to see how overwhelming it is.

Your example Chelsea is the best team of Europe, but its tactics may not be the most overwhelming.

This match is before I adjust my tactics got absolutely battered. Clearly their tactics is very attacking.

This match took place around 2 months later in the same season with a similar Liverpool side. Notice how after I change my tactics even though the stats is similar the results are so much different.

Bonus on how I managed to FM the AI

My stats in the last 30 league matches. You can do your own calculation and will find out that I am more efficient than the AI.

Edited July 15, 2020 by zyfon5

[DeanMon]

1 hour ago, zyfon5 said:

You can overwhelm the AI and win at the same time as long as you know what you are doing.

This match is before I adjust my tactics got absolutely battered. Clearly their tactics is very attacking.

This match took place around 2 months later in the same season with a similar Liverpool side. Notice how after I change my tactics even though the stats is similar the results are so much different.

Bonus on how I managed to FM the AI

My stats in the last 30 league matches. You can do your own calculation and will find out that I am more efficient than the AI.

I said I still sometimes won matches luckily, and also won matches when dominating. I had all similar match examples like above, you don't need to provide.

My statement is that: given the same domination in match, AI (versus human) scores more than human (versus AI).

So, to nullify it, we just need the data to show that over a large enough sample (said 1 season to keep variable changes at minimum), AI and human convert chances at similar rate when dominating the other.

Edited July 15, 2020 by DeanMon

[Zemahh]

13 minutes ago, DeanMon said:

So, to nullify it, we just need the data to show that over a large enough sample (said 1 season to keep variable changes at minimum), AI and human convert chances at similar rate when dominating the other.

I mean, you've shut down every logical explanation you've been given so far and you don't seem to believe the developers themselves clearly stating the game isn't rigged and doesn't differentiate between the AI and human managers either, so maybe it should be YOU who provides the proof for your theory, not the other way around. Taking a season's worth of Goals Scored and Chances Created stats and running with it isn't it, because there's so many variables that could result in you having to create more chances than the AI, that looking at those stats in isolation really doesn't even begin to prove anything. Besides that, surely if the game was actually rigged, those stats would be rigged as well, or why would the SI leave themselves exposed to a few clicks around the stats section?

The reason those stats exist, is that you can figure out where things are going wrong or right in comparison to the rest of the league. In your case they paint a clear picture of inefficient chance creation, which should point you in the right direction.

Here's a good start:

 

On 04/07/2019 at 07:02, Rashidi said:

If you are not scoring so much you should be looking at the quality of the chances you get.  I did a video called SIBOT analysis, where I explain why I eschew normal stats in the game and focus specifically on identifying quality of shots. If a tactic can produce quality chances then its your strikers, perhaps then you focus on training. If you are not producing quality chances then its your tactic so then you make either changes to personnel or tactics.

 

[DeanMon]

1 hour ago, Zemahh said:

I mean, you've shut down every logical explanation you've been given so far and you don't seem to believe the developers themselves clearly stating the game isn't rigged and doesn't differentiate between the AI and human managers either, so maybe it should be YOU who provides the proof for your theory, not the other way around. Taking a season's worth of Goals Scored and Chances Created stats and running with it isn't it, because there's so many variables that could result in you having to create more chances than the AI, that looking at those stats in isolation really doesn't even begin to prove anything. Besides that, surely if the game was actually rigged, those stats would be rigged as well, or why would the SI leave themselves exposed to a few clicks around the stats section?

The reason those stats exist, is that you can figure out where things are going wrong or right in comparison to the rest of the league. In your case they paint a clear picture of inefficient chance creation, which should point you in the right direction.

Here's a good start:

 

 

Have you read the posts carefully?

1) Look at Hoffenheim screenshots on page 1. The poster has 1 season of data where he created much more chances than the average, but converted reasonably less.

2) You can't say in general terms that his chances are in low quality. A chance is a chance. For a match he plays attack, there must be other matches AI plays ultra attack (said, AI Bayern vs AI Dusseldorf). And because he played Hoffenheim, there must be matches where he played defensively (against AI Dortmund for example). Hoffenheim is quite an average team, the stats should be converging close to the average in the long run, not that far away.

Yeah, SI, maybe you forget to cover the your tracks in the stats page.

Edited July 15, 2020 by DeanMon

[themadsheep2001]

Been following this for a while now and the OP is intent on ignoring every rational answer given to him. Given that people are are putting real time and effort into answering, this feels very disingenuous. As a result going to close this. All the answers to the OPs post are already here.