|
Am I right about the AI and breakaways?
|
|
| Ad Bot |
Posted on 23-11-2024 05:39
|
Bot Agent
Posts: Countless
Joined: 23.11.09
|
|
| IP: None |
|
|
| cactus-jack |
Posted on 17-07-2011 20:25
|
Classics Specialist
Posts: 3936
Joined: 31-07-2009
PCM$: 200.00
|
I think I know how the AI reacts to breakaways; if they should reel them in or let them go. There seems to be a huge mistake in the programing; the AI solely looks at the placement in the GC rather then who it is.
I just played a race where there are mostly flat stages and one very hilly one. Sorta the Alp d'Huez of hills. The first 3 stages were comletely flat, then came the hilly one and it ended with flat one. On the third stage I tried to send a guy in a breakaway to secure the KoM-jersey. He has rubbish climbing stats, but you don't need to be a climber to win the KoM jersey this year either (sadly). This guy just happend to be 7th (+0:58) in the GC due to bonus seconds, but he was no danger what so ever to the GC since there would be a climb the very next day.
I tried to send him in a breakaway, but everytime I tried the opposition did everything they could to get him back. This happend time after time. However, Yohann Ofredo (I think) suddenly went in a breakaway and no one bother pulling him in. In my DB he has hill-stats of about 76 and was only 1 minute behind. The difference was that he was currently in the 96th place in the GC.
To sum up: completely non-threatening rider who is 58 seconds behind most be reeled in no matter what! But a huge favourite who's 60 seconds behind is of no threat what so ever.
PS: This has nothing to do with KoM-points because none of the riders who had points to begin with bother going for them.
Has anyone else experienced this to? Because if it's the case then I feel like AI has a pretty big fault... |
| |
|
|
| flatime |
Posted on 17-07-2011 21:16
|
Amateur
Posts: 18
Joined: 11-07-2009
PCM$: 200.00
|
hum..interesting.
If is TRUE, that's a BIG bug,i will try that night ,and tell you later! |
| |
|
|
| KiwiZ |
Posted on 17-07-2011 23:40
|
Amateur
Posts: 7
Joined: 08-07-2011
PCM$: 200.00
|
I have experienced the exact same thing. Another problem that annoyed me is the AI was fine with letting S Sanchez into a breakaway (he was 4-5 mins down on the GC), and let the breakaways lead get out to 8 minutes before cutting it back to just above 6 minutes. At this point I realised that I needed to do someting but due to the hilly/mountain end could only cut it back to 4 mins. This was a stage race and he managed to get himself back into overall contention (19 seconds down).
Point is, surely it is not unreasonable that the AI should respond when someone like Sanchez, a race favourite in a stage race with many climbs etc, tries to get back into GC contention? |
| |
|
|
| Lachi |
Posted on 18-07-2011 00:57
|
Grand Tour Champion
Posts: 8516
Joined: 29-06-2007
PCM$: 200.00
|
But this happens in real cycling also. Why should the AI hunt down all breakaways, you also have a team, so chase them if they are a danger to your leader. |
| |
|
|
| pbenoit4 |
Posted on 18-07-2011 02:51
|
Free Agent
Posts: 147
Joined: 06-08-2009
PCM$: 200.00
|
I had a similar scenario, but from the opposite perspective. With about 70km to go, I noticed that the break still had about 8min. I worked to pull that gap down to about 45 seconds with 10 km to go and then let the men who were working to pull it back rest and started to set up my sprint train. The problem was, in that 2 or 3 km transition time, the AI never worked and the break pulled back out to 1m30s with 5km to go and I sprinted for 12th place.
While I'm basing this on only a few examples that I've witnessed, including the one above , it seems like the AI makes its mind up as soon as the break forms whether it will ever bother to chase the break. Even if you do all the work and logically it's going to come to a sprint finish, if you let up for a few km, they are resigned to let it go. In real life, even on the most dominant HTC days, at least one team is going to put one rider on the front at some point to keep some vested interest in the sprint finish if it ever gets that close at the end of the stage. It seems as if there is some sort of pre-programmed event that turns off AI interest in the chase regardless of any change in scenario. |
| |
|
|
| KiwiZ |
Posted on 18-07-2011 02:56
|
Amateur
Posts: 7
Joined: 08-07-2011
PCM$: 200.00
|
Lachi wrote:
But this happens in real cycling also. Why should the AI hunt down all breakaways, you also have a team, so chase them if they are a danger to your leader. Yes but would people like A Schleck and Basso really want Sanchez back in the mix? The fact they were doing no work whatsoever is unrealistic. |
| |
|
|
| Rasmusen |
Posted on 18-07-2011 09:11
|
Neo-Pro
Posts: 340
Joined: 25-06-2008
PCM$: 200.00
|
cactus-jack wrote:
I think I know how the AI reacts to breakaways; if they should reel them in or let them go. There seems to be a huge mistake in the programing; the AI solely looks at the placement in the GC rather then who it is.
I just played a race where there are mostly flat stages and one very hilly one. Sorta the Alp d'Huez of hills. The first 3 stages were comletely flat, then came the hilly one and it ended with flat one. On the third stage I tried to send a guy in a breakaway to secure the KoM-jersey. He has rubbish climbing stats, but you don't need to be a climber to win the KoM jersey this year either (sadly). This guy just happend to be 7th (+0:58) in the GC due to bonus seconds, but he was no danger what so ever to the GC since there would be a climb the very next day.
I tried to send him in a breakaway, but everytime I tried the opposition did everything they could to get him back. This happend time after time. However, Yohann Ofredo (I think) suddenly went in a breakaway and no one bother pulling him in. In my DB he has hill-stats of about 76 and was only 1 minute behind. The difference was that he was currently in the 96th place in the GC.
To sum up: completely non-threatening rider who is 58 seconds behind most be reeled in no matter what! But a huge favourite who's 60 seconds behind is of no threat what so ever.
PS: This has nothing to do with KoM-points because none of the riders who had points to begin with bother going for them.
Has anyone else experienced this to? Because if it's the case then I feel like AI has a pretty big fault...
What about sprint points? maybe 4 this they dontwant him to go |
| |
|
|
| MacC |
Posted on 18-07-2011 10:19
|
Sprinter
Posts: 1594
Joined: 15-07-2008
PCM$: 700.00
|
Rasmusen wrote:
cactus-jack wrote:
I think I know how the AI reacts to breakaways; if they should reel them in or let them go. There seems to be a huge mistake in the programing; the AI solely looks at the placement in the GC rather then who it is.
I just played a race where there are mostly flat stages and one very hilly one. Sorta the Alp d'Huez of hills. The first 3 stages were comletely flat, then came the hilly one and it ended with flat one. On the third stage I tried to send a guy in a breakaway to secure the KoM-jersey. He has rubbish climbing stats, but you don't need to be a climber to win the KoM jersey this year either (sadly). This guy just happend to be 7th (+0:58) in the GC due to bonus seconds, but he was no danger what so ever to the GC since there would be a climb the very next day.
I tried to send him in a breakaway, but everytime I tried the opposition did everything they could to get him back. This happend time after time. However, Yohann Ofredo (I think) suddenly went in a breakaway and no one bother pulling him in. In my DB he has hill-stats of about 76 and was only 1 minute behind. The difference was that he was currently in the 96th place in the GC.
To sum up: completely non-threatening rider who is 58 seconds behind most be reeled in no matter what! But a huge favourite who's 60 seconds behind is of no threat what so ever.
PS: This has nothing to do with KoM-points because none of the riders who had points to begin with bother going for them.
Has anyone else experienced this to? Because if it's the case then I feel like AI has a pretty big fault...
What about sprint points? maybe 4 this they dontwant him to go
Surely the more obvious answer is that they were protecting the yellow jersey |
| |
|
|
| ricotero_uy |
Posted on 18-07-2011 14:18
|
Domestique
Posts: 610
Joined: 21-03-2008
PCM$: 200.00
|
If you remember Tour 2006 with Pereiro and Voigt taking 30 minutes... giving Pereiro the yellow jersey. The other day Phonak broke the race... with Landis doped... but he could attack for 130kms or so.
I guess some riders should have special treatment.. but not always they are a threat for the pack. |
| |
|
|
| cactus-jack |
Posted on 18-07-2011 17:49
|
Classics Specialist
Posts: 3936
Joined: 31-07-2009
PCM$: 200.00
|
Lachi wrote:
But this happens in real cycling also. Why should the AI hunt down all breakaways, you also have a team, so chase them if they are a danger to your leader.
There's about as much logic in that as saying "Why should Leopard-Trek care if Cadel Evans went in a breakaway?" Even though the other teams doesn't have the leaders jersey, there's no reason why they should just let everyone who wants to get away.
By the way, when have you ever seen a team exhaust itself to reel in a completely non-threatening rider for no reason what so ever, but then lets a huge favourite get away scott free? |
| |
|
|
| cunego59 |
Posted on 18-07-2011 17:59
|
Team Manager
Posts: 6508
Joined: 14-09-2008
PCM$: 1090.00
|
I experienced different szenarios, can't say that what you described is always the case.
For example: Tour Down Under, I sent Lucas Haedo in the break, took both intermediate sprints, and therefore was third overall, 2 seconds down. But for the overall, Lucas was of course no threat.
On the next stage, I could bring him into the break again, without big problems.
|
| |
|
|
| Lachi |
Posted on 18-07-2011 18:00
|
Grand Tour Champion
Posts: 8516
Joined: 29-06-2007
PCM$: 200.00
|
I was not talking about Cadel Evans, but KiwiZ mentioned Sanchez and nobody did chase him in this years TdF as far as I have seen. He did not went with the early breakaway but he still managed to get some time back. |
| |
|
|
| cactus-jack |
Posted on 18-07-2011 18:06
|
Classics Specialist
Posts: 3936
Joined: 31-07-2009
PCM$: 200.00
|
But you can't use that example, it's completely different. In my case it would be more like Leopard-Trek doing everything they could to keep Ben Swift from going in a breakaway, but when Contador went with 80km to go they just let him lose and no one bother reeling him in. |
| |
|
|
| KiwiZ |
Posted on 19-07-2011 03:55
|
Amateur
Posts: 7
Joined: 08-07-2011
PCM$: 200.00
|
Lachi wrote:
I was not talking about Cadel Evans, but KiwiZ mentioned Sanchez and nobody did chase him in this years TdF as far as I have seen. He did not went with the early breakaway but he still managed to get some time back. I get what you are saying but in my case, they GC contenders still had teammates with them so could easily have chased, I used my one guy just to cut it back by they couldn't really care less. In the Tour it was at a point where if they did follow Sanchez, the leaders would be wasting their own energy and could be potentially passed by doing so. I get there are some cases but in PCM it is far too common to be realistic. |
| |
|
|
| 7king |
Posted on 07-08-2011 14:11
|
Under 23
Posts: 65
Joined: 10-08-2009
PCM$: 200.00
|
pbenoit4 wrote:
I had a similar scenario, but from the opposite perspective. With about 70km to go, I noticed that the break still had about 8min. I worked to pull that gap down to about 45 seconds with 10 km to go and then let the men who were working to pull it back rest and started to set up my sprint train. The problem was, in that 2 or 3 km transition time, the AI never worked and the break pulled back out to 1m30s with 5km to go and I sprinted for 12th place.
While I'm basing this on only a few examples that I've witnessed, including the one above , it seems like the AI makes its mind up as soon as the break forms whether it will ever bother to chase the break. Even if you do all the work and logically it's going to come to a sprint finish, if you let up for a few km, they are resigned to let it go. In real life, even on the most dominant HTC days, at least one team is going to put one rider on the front at some point to keep some vested interest in the sprint finish if it ever gets that close at the end of the stage. It seems as if there is some sort of pre-programmed event that turns off AI interest in the chase regardless of any change in scenario.
This happends to me in almost every race. In Tour Down Under, 5 stages ended with a sprint between breakaway riders, and the peloton sprinter giving everything for a 12th place, 20 seconds behind... |
| |
|
|
| kwkke10 |
Posted on 07-08-2011 14:29
|
Stagiare
Posts: 201
Joined: 27-04-2011
PCM$: 200.00
|
They should make a thing what could define if the peloton would chase the rider or not, all the big GC contenders would have 15 so that the peloton will chase them if they are within 15 min or something
|
| |
|
|
| wroblitz |
Posted on 09-08-2011 03:42
|
Neo-Pro
Posts: 289
Joined: 27-12-2009
PCM$: 200.00
|
For me it doesn't matter who you breakaway with, when you breakaway and what team you're playing - everytime you're involved in breakaway you can be sure it will be chased (which makes breakaways pointless). It's pretty annoying seeing peloton doing nothing to AI breakaways and then suddenly speeding up as soon as you join it / start your own. It effectively ruins the game however I use it to chase AI breakaways - instead of involving 3-4 riders to lead the peloton I simply send one for a breaky. |
| |
|
|
| gmart |
Posted on 09-08-2011 03:43
|
Amateur
Posts: 16
Joined: 12-07-2011
PCM$: 200.00
|
During my Garmin-Cervelo career, my 2011 tour saw a crash on stage 1 which lead to significant time loss for a lot of riders, most notably Roman Kreuziger. By the time we got to the Alps Kreuziger had lost up to 15 minutes. I had Tom Danielson placed in 4th within 1:30 of Cunego (Schleck right behind).
Well, Lampre had no problem letting Kreuziger go in the break and get a gap of 15 minutes, they didn't even bother! I had to get to the front to try and mitigate SOME damage because there's no way I wanted Danielson going against RK.
Strange tactics indeed, I've seen some odd things in simulate as well. Mainly sprinter teams allowing breakaways on flat stages to finish well ahead. I've seen as much as 7 minutes, on a flat stage! |
| |
|
|
| Wiley |
Posted on 09-08-2011 09:49
|
Domestique
Posts: 683
Joined: 24-02-2011
PCM$: 200.00
|
Its is anissue, but like wroblitz i use it to my advantage.
TDU 2011, i had kennaugh going fot KoM, but he was 7th @12" on the leader, so would be chased extremly hard.
So to stop anyone i didnt want in a break, i simply attacked with kennaugh and the ai would chase for me.
But in Malaysia, i caught the break with 10km to go ... so i stopped relaying and prepared my spirnt, then suddenly thr break was out to 30" as the ai wouldnt relay or set up a sprint train.
Ferrari-BA > Psuedo Man Game
|
| |
|
|
| Kentaurus |
Posted on 09-08-2011 09:59
|
Classics Specialist
Posts: 3999
Joined: 26-07-2011
PCM$: 200.00
|
I am with many now that use sending a rider into the break as a tool for the team. I almost always try to get a rider into the break now, and then sit back and let the other teams do all the work, I don't have my break rider take turns nor do I have my pelaton riders take turns. It also gives me a chance to get some KoM points on many stages.
In classics you can get the AI to do some dumb stuff, I got Taylor Phinney into the break in Milan - San Remo, and watched Gilbert take over pulling the pelaton to get the break back. Got the win with Boasson Hagen and I don't think Gilbert ended up in the top 15.
AZTECA - NBCSN  |
| |
|
