AI managers

Why do the AI managers make such godawful transfers and signings a majority of times?

I was in Denmark for 10 seasons and allowed my contract to run out after a good stay in their premier league. I was offered the job at the recently relegated Peterborough who were in the EPL and were dropped down to the Championship.

I look into the squad and see no doubt why. One of the main reasons why they were dropped was because the stupid manager had bought a striker for 8 million and paid him top player wages of 85k a week (Peterborough were only in the Premier League for 1 year). The previous 3 seasons on the strikers history sheet showed 93 appearances and only 8 goals. The year he was bought he made 33 appearances and had 1 goal and 1 assist and a rating of 6.56 total.

Why in the world would some stupid manager spend and waste so much money on such a crap player. His one positive is 18 accel and 18 pace with only 9 finishing.

This extremely lack of AI signing intelligence is horrendous IMO.

To be honest, a player that fast should be getting goals anyway. What were the rest of his stats?

I guess all AI managers prioritise certain stats based on the leagues they are in and they squad they already have, probably he coveted quick players and that guy fitted the bill. To be fair good human like AI is just about the hardest thing to get right in programming terms, but I agree sometimes they are very odd.

In my game Barcelona just bought some Australian 19 year old from the Championship for about 13 million, who wouldnt even make it into my reserve team (at Ajax) they then proceeded to loan him out where he has sat not playing a single game. Where as Lukaku and Sanogo are still at their old clubs waiting to be signed... weird

It's a pretty major issue, well known. The AI is not sophisticated enough as it stands, which is why Ajax never play Eriksen, and big teams make such awful signings more often than not.

Why in the world would some stupid manager spend and waste so much money on such a crap player. His one positive is 18 accel and 18 pace with only 9 finishing.

Afonso Alves was £15m. Mistakes are made.

Why do the AI managers make such godawful transfers and signings a majority of times?

I was in Denmark for 10 seasons and allowed my contract to run out after a good stay in their premier league. I was offered the job at the recently relegated Peterborough who were in the EPL and were dropped down to the Championship.

I look into the squad and see no doubt why. One of the main reasons why they were dropped was because the stupid manager had bought a striker for 8 million and paid him top player wages of 85k a week (Peterborough were only in the Premier League for 1 year). The previous 3 seasons on the strikers history sheet showed 93 appearances and only 8 goals. The year he was bought he made 33 appearances and had 1 goal and 1 assist and a rating of 6.56 total.

Why in the world would some stupid manager spend and waste so much money on such a crap player. His one positive is 18 accel and 18 pace with only 9 finishing.

This extremely lack of AI signing intelligence is horrendous IMO.

I have 2 words for you:

Andriy Shevchenko

To be honest, a player that fast should be getting goals anyway. What were the rest of his stats?

Well after I got rid of him (he was bought for peanuts by a French team in their L2) I went back and scouted him again the report said he would make a leading L1 striker. If he would only make a leading L1 striker, why was he starting and playing most of the games for a EPL squad?

Well to be fair the manager was sacked............. Peterborough would have just spent almost their entire transfer budget on one player, and when he arrived he was crap, but they couldn't go out and sign anybody else. As per the posts above it is not as if this has not ever happened in real life.

The Alfonso Alves example is spot on, and imagine what would have happened if a team like Blackpool had bought Bebe instead of Man Utd - instead of playing in the Reserves team all season they would have almost certainly had played him, even if his performances were below standard - because they wouldn't have had an alternative.

Johan Elmander is a another good example - came across to Bolton with lots of hype and expectation and had an appalling first season

There needs to be more long-term squad building programmed into the behaviour of managers. And to do so, SI needs to implement player types as filter options for scouting, player types that are much more comprehensive than they are now. The AI managers needs to use these filter options as well, so that Stoke signs big, strong players and bullet-throw guys while Barcelona sign players with high technique, creativity and passing.

If you want a target man, you send your scout to find suitable target men. Poachers - they go to find players of that type. Flamboyant winger - a winger with high flair and good technique. Tenacious ball-winning midfielder - an athletic and hard-working aggressive midfielder. A leader box-to-box midfielder - a high-stamina and off the ball, high influence midfielder with a professional and determined personality.

The variables are there, they just need some work. I believe this would take FM to the next level as a gaming experience.

To those who mentioned Alfonso Alves and Shevchenko, there is a bit of a difference. Both of those had fantastic goalscoring records at their previous clubs. The OP's example did not. I think it's a bit silly for an AI manager to buy a striker who has scored 8 goals in 3 seasons, particularly if pace is their main attribute - the only thing that could justify that poor a return is if the player was a deep-lying grafter.

The AI for managers relies way to much on CA and reputation when buying players. There is nothing I can see in place to make it looks at the quality of the players performances over the last season or that makes sure it signs players suitable to its tactics.

The striker in the OP's example probably has a half decent CA mostly spent on pace and a highish Reputation because he has been playing regular football in a decent league and thats enough for the AI to sign him I don't think it looks at his Attributes or performances at all.

Seems pretty realistic to me. The AI would be unstoppable if it looked at stats, and we'd all be complaining about how it cheats.

James Beattie- Everton. Mistakes will be made.

The bigger issue with AI managers and crap sigings is when your in non league. With a large database there are thousands of players out of contract. There seems to be no system in which players AI managers sign. Players will go to top BSN/BSS club with players worth about 20,000, and they may not have an attribute above 10.

On mine there haven't really been many terrible signings, though I'm just 4 and a bit seasons in. Though Dalglish has just signed Scott Dann for £13.25m. Which if he'd had 4 top seasons my be justifiable, but he's only had one season rated over 7, and that was his 2nd season.

Sunderland have signed Demba Ba for 10 million, which in real life might be reasonable. In game he has 11 goals in 82 games. No reason to sign him whatsoever.

What I've found looking through the transfer histories of top division teams is that they've actually been quite astute with their signings, I checked out every player that cost more than 10 million, and those are the only two bizarre signings. The rest are understandable.

I find there are too many excellent players on the transfer list that the AI are not interested in. Gonzalo Rodriguez (26) for €2.6m, Cristian Rodriguez (25) for €2.5m, Matias Fernandez (25) for €950k. No one else was interested in these players. Even if they play average I should still make a big profit on all of them. Their wages weren't high either. Both of the Rodriguezes have performed very well too.