It seems like a decent improvement with PCM 20. Obviously we'd need way more testing to know for sure, but we don't have much to lose at this point in my opinion. PCM 18 is already pretty terrible for us. PCM 20 could be better. Might as well try it out.
RIP Exxon Duke, David Veilleux, Double Feature, and Monster Energy
I still think the game is not worse than PCM15 to be completely honest. Well rounded riders score better (as they should), breakaways are A LOT more realistic, it's a lot less predictable,....
Croatia14 wrote:
I still think the game is not worse than PCM15 to be completely honest. Well rounded riders score better (as they should), breakaways are A LOT more realistic, it's a lot less predictable,....
That can be debated, but after 2 seasons I'm pretty confident that it is worse for our specific purposes.
RIP Exxon Duke, David Veilleux, Double Feature, and Monster Energy
Croatia14 wrote:
I still think the game is not worse than PCM15 to be completely honest. Well rounded riders score better (as they should), breakaways are A LOT more realistic, it's a lot less predictable,....
That can be debated, but after 2 seasons I'm pretty confident that it is worse for our specific purposes.
Indeed, your specific purposes. Even other people who relegated are against going back, yet you still remain bitter at the switch. 18 is without a doubt better than 15 overall
Croatia14 wrote:
I still think the game is not worse than PCM15 to be completely honest. Well rounded riders score better (as they should), breakaways are A LOT more realistic, it's a lot less predictable,....
That can be debated, but after 2 seasons I'm pretty confident that it is worse for our specific purposes.
Indeed, your specific purposes. Even other people who relegated are against going back, yet you still remain bitter at the switch. 18 is without a doubt better than 15 overall
15, 18, 20, our biggest and acutaly only issue, is inflation. With a normal PCM DB all three games work fine. The MG DB is broken and is the main cause of our problems, not the specific game.
Croatia14 wrote:
I still think the game is not worse than PCM15 to be completely honest. Well rounded riders score better (as they should), breakaways are A LOT more realistic, it's a lot less predictable,....
That can be debated, but after 2 seasons I'm pretty confident that it is worse for our specific purposes.
Indeed, your specific purposes. Even other people who relegated are against going back, yet you still remain bitter at the switch. 18 is without a doubt better than 15 overall
Specific purposes like having a sprinter and sprint train? lol. I didn't even suggest going back to 15 in my comment. I haven't suggested as much in over a year.
If you really believe that most people are satisfied with PCM 18 compared to how they were with 15, just look at this thread or any race thread from this year. I'm far from the only one who feels negatively about it.
RIP Exxon Duke, David Veilleux, Double Feature, and Monster Energy
TheManxMissile wrote:
15, 18, 20, our biggest and acutaly only issue, is inflation. With a normal PCM DB all three games work fine. The MG DB is broken and is the main cause of our problems, not the specific game.
While I certainly agree that the inflation exacerbates many of our problems (and should also be addressed), I think that's a reductive view. Yes, our stat matrix influences AI behavior, but it may do so differently across versions. For instance, if further testing confirms that not only the number of successful breakaways, but also the variation of teams that join them (both of which have been the source of frustration) is more akin to what we'd like to see, even with our DB, then that might be worth it, no? Or maybe if it turns out that in PCM20, sprint trains don't release the main guy way too early to the extent that trains can be a detriment. Yes, the relative strength of breakaway guys v peloton and top sprinters v secondary sprinters play a role in both cases, but the extent may vary.
TheManxMissile wrote:
15, 18, 20, our biggest and acutaly only issue, is inflation. With a normal PCM DB all three games work fine. The MG DB is broken and is the main cause of our problems, not the specific game.
While I certainly agree that the inflation exacerbates many of our problems (and should also be addressed), I think that's a reductive view. Yes, our stat matrix influences AI behavior, but it may do so differently across versions. For instance, if further testing confirms that not only the number of successful breakaways, but also the variation of teams that join them (both of which have been the source of frustration) is more akin to what we'd like to see, even with our DB, then that might be worth it, no? Or maybe if it turns out that in PCM20, sprint trains don't release the main guy way too early to the extent that trains can be a detriment. Yes, the relative strength of breakaway guys v peloton and top sprinters v secondary sprinters play a role in both cases, but the extent may vary.
As i said on the previous page, i prefer the look of the PCM20 results that were run vs those of PCM18. Not perfect but a step in the right direction.
But we sldo cannot pretend that moving game will make everything ok, because it won't. Just like moving from 15 to 18 didn't solve everything. We need to actually tackle the core problem.
But yeah, even i bore myself repeating the same thing over and over
Croatia14 wrote:
I still think the game is not worse than PCM15 to be completely honest. Well rounded riders score better (as they should), breakaways are A LOT more realistic, it's a lot less predictable,....
That can be debated, but after 2 seasons I'm pretty confident that it is worse for our specific purposes.
I'm going to be honest, I cant really recall what was so terrible about PCM 15 AI, apart from the complaint that ACC was too strong. Perhaps the issues are with Grand Tours, which I believe I was only involved in 1 while 15 was the game used, the Giro as a wildcard, where I thought it handled things well. Choi was far behind in the GC, and went after the KOM and took a stage win late in the race, and ended up finishing just outside the Top 10. The difference was he went on a solo attack, whereas now, multiple dangerous riders are allowed to gain time for some reason.
If we compare sprints/flat stages, I'd say they are definitely worse now, except for rolling flat classics where late attacks sometimes happen.
Hill stages can be argued either way probably, but with the MG DB, some riders became useless in 18. Im pretty sure long range attacks sometimes had success in 15. In 18, I think its 99% a death sentence, unless a group of favourites go together.
Mountain stages. Basically just a sprint up the last mountain now, before there was the odd long range attack, and downhill finishes were slightly better. Smarter AI in 15 imo.
Cobbles, probably the only thing definitely better in 18, I'll give it that.
TTs, are worse now, the gaps are sometimes massive, and then prologues are completely random now more or less. Hilly or MO TTs aren't as good either, atleast with this DB.
Grand Tour AI: I think the breaks having more success can be seen as better, but the actual AI of it is worse. Its the same 20 riders or so trying to get into the break everyday (from my experiences), and a rider in 9th OVL with a MO of 74 can be allowed sometimes, but a guy with 78 in 23rd GC isn't even if he's 45 mins down, and the rider in 9th overall is 4 mins down. The problem is a 74 MO rider in the Top 10 is almost for sure going to have a +1/2 for exceeding expectations in RDC, so they'll likely be a 78/79 rider anway, and thus its an endless cycle of overperforming for a "weak" rider.
Mountain stages are honestly pretty boring now as nothing ever happens until the last climb and the break is usually allowed a 20min gap, so the stage win never goes to a favourite. I know when I played 18 a lot, if the human didnt chase, the break never gets caught. On flat stages the sprinters teams should chase and only allow a break of 3/4 riders, but no, breaks of 15 are allowed sometimes. 18 always makes the catch too late, or not at all, which affects the sprint trains. If they're caught with 10km to go, sprints actually work out well. If its late, teams dont really form trains, just one long line. Stage 10 of the Vuelta, the gap was 1.5mins with like 30km to go. 20km later, the peloton lost 5 seconds. Theres really no reason for the peloton to not be making the catch there.
Race Day Condition: No comment needed here I think.
I'm obviously biased though, but if I read a comment that its because Im relegating, I would like to point out, I have been against 18 from the start, and complained about it in the season I promoted. I knew going into this season relegation was likely. Based on the Grand Tours, I think if things had gone better it would be a lot closer, but now theres really no chance I wont and thats where my frustration comes from, because teams nearly identical to mine are scoring 3x as much in some races because of luck/randomness/RDC/Bad AI and teams much worse also managed to outscore my team by a lot, with really no explanation.
Inflation is also a likely contributor to these issues too, as Manx has suggested.
The problem with going back to PCM15 is quite simple in my opinion. It will kill the ManGame. Why? Because we'd be stuck with that game forever. Cyanide are clearly not going back to the way races where handled on that game and all future games will have a behavior similar with PCM18.
alexkr00 wrote:
The problem with going back to PCM15 is quite simple in my opinion. It will kill the ManGame. Why? Because we'd be stuck with that game forever. Cyanide are clearly not going back to the way races where handled on that game and all future games will have a behavior similar with PCM18.
Certainly. It would be much more productive if people completely forgot about PCM15 and looked forward, tested new installments and thought of ways to adapt to them.
Going back to PCM 15 for the purposes of this conversation was a complete straw man to begin with, there's not much point in dwelling on it. We should be discussing PCM 18 vs PCM 20.
RIP Exxon Duke, David Veilleux, Double Feature, and Monster Energy
Maybe a stupid question: Why does PCM15 work for ICL (and isn't even questioned as far as I know), but not for MG? Is it the long reports vs. short result list approach that makes this difference?
Not that I'm in favor of going back, just wondering...
And just concerning the Giro: I actually replayed that 4th stage several times after the enitre race was done. Spilak only lost time to the main group on one other occasion - but with one exception always finished behind Talansky on that stage. PCM18 is really very, very bad at handling multiple riders of one team on the fav's list - if PCM20 does that better, that would already be a reason to move on for me.
Fabianski wrote:
Maybe a stupid question: Why does PCM15 work for ICL (and isn't even questioned as far as I know), but not for MG? Is it the long reports vs. short result list approach that makes this difference?
Not that I'm in favor of going back, just wondering...
And just concerning the Giro: I actually replayed that 4th stage several times after the enitre race was done. Spilak only lost time to the main group on one other occasion - but with one exception always finished behind Talansky on that stage. PCM18 is really very, very bad at handling multiple riders of one team on the fav's list - if PCM20 does that better, that would already be a reason to move on for me.
PCM 15 does work for MG. It worked for 3 years just fine. It wasn't without issues, but no PCM version ever has or ever will be. The main argument against it is what alexkr00 and quadsas brought up, that we should eventually progress through games, and that, given the new engine's persistence, PCM 15 will forever be a dead end in that.
RIP Exxon Duke, David Veilleux, Double Feature, and Monster Energy
A few reactions in no particular order. Because I ran these very quickly I won't claim to have paid attention to how the results occurred in all stages so can't comment on something like whether different riders were in the break.
1. Sprints. Good news is there seems to be more sprints because they catch the breakaway. Bad news is I don't think they are any prettier. Cio's head would have exploded with the way Demare was constantly out of position in the second test. I do think sprinting is where stat inflation (and lack of variation) is the biggest issue. I think we have had some nice sprints in PCM18 recently when there were more limited fields. (Also think that this may be another example of PCM being too realistic for the man game sprints have become very messy IRL).
2. GC Attacks. Very subjective but there were two really good stages I did pay attention to. 16 in Test 1 and 19 in Test 2. In the first Taaramae went crazy to get back the jersey and obliterated the field from a long way out. Putting over 2 minutes into everyone except Kritskiy. Kritskey attacked like 6 times in Test 2 S19 and eventually got a big gap only to see Taaramae solo back to him. Meanwhile Schleck fought out the stage win. So not sure it is any different but those stages seemed interesting to me.
3. I think the real difference was in the way some of the non-GC stages played out. Obviously more sprints but also on the hilly stages the peloton cared about the stage win. I might run the Vuelta to see how those first few stages work out.
4. Two leaders: Small sample size, I think the game addressed two leaders by making the second one work, Talansky finished pretty poorly in both test but he definitely did some work on the front for Spilak. This is probably better than 18 where it seemed to change its mind on the leader from stage to stage.
5. I tend to agree that there is nothing to lose by moving to 20. It isn't that different but if a little further testing confirms what I found then it seems worth it.
alexkr00 wrote:
The problem with going back to PCM15 is quite simple in my opinion. It will kill the ManGame. Why? Because we'd be stuck with that game forever. Cyanide are clearly not going back to the way races where handled on that game and all future games will have a behavior similar with PCM18.
Certainly. It would be much more productive if people completely forgot about PCM15 and looked forward, tested new installments and thought of ways to adapt to them.
I can understand this point of view for sure. Moving backwards sounds illogical. If what I read about the new games wasnt so hopeless (imo), I wouldn't be against moving to newer games at all.
The problem to me is new games will all still be based on 18's flawed AI. Why do all this work just to get the AI 10% better, if moving back to an older game that requires very little work and improves the AI by 30% for example. . Newer doesn't always mean better.
The thing that frustrates me with 18 and 20, is not the randomness of the results, but the way the randomness comes about. For example, sprints being slightly random, sounds great on paper, but its because sprint trains suck, leadouts/sprinters start their sprint way too early and then die before the line. Underrated riders get solid GC results in GTs, again sounds nice, but if its because they are constantly allowed in the breakaway when a struggling GC rider isn't allowed or cant even be bothered to fight for a stage win theres just a lack of logic to it all. Like what allows some teams with weak teams in GTs to be so active, but other teams with solid riders seem invisible. If these issues were fixed, I think people would have a lot less gripes with 18's AI.
If it would be possible to test out some of these issues I've uncovered from searching around on PCM 20, here they are:
Feel free to scroll through the Official Discord Bug Report Thread as well. Tons of issues with breakaways here. https://discord.c...te/9kxcCUy
- Breakaways were apparently addressed in a patch in July and Again in September I believe, so Im only referring to complaints after then.
- In non GTs I've read many things about a break only letting 1 single rider clear, getting a gap of 1min and then catching them halfway through the stage. The rest of the stage has no action until the sprint at the end.
- I've also read issues where sometimes the peloton just doesnt chase at all, and the break wins by 45mins, with the the break riders all taking the top GC spots.
- Other issues where the breakaway just stops riding for 2mins (I think it might have to do with a rider puncturing, but the AI is using the "follow" mechanic and isn't smart enough to keep riding, instead of following a motionless rider)
- If a favourite punctures 20km into a stage, the peloton will ride full gas the rest of the stage sometimes 100+km. The AI is programed to hurt a favourite that falls behind, but irl if a favourite punctures 20km into a stage, the peloton is going to let him back, since its more or less an unwritten rule to. Imagine this happening to your rider...
I think we're really just in a no-mans land here. 15 is probably the best AI (can anyone argue its not?), but going back makes no sense. At the same time, 18 seems pretty trash to me, and 20 is either marginally better or worse.
Ulrich's testing is already looking better than what I would have expected, but I think the testing needs to be a lot more open than last time (and it seems like it has been so far), as I recall many people shocked with how things went initially with PCM18. I'm really appreciating your insights btw, Ulrich
Edited by valverde321 on 10-03-2021 02:17
Not here to just rant again about AKA or sprints not working. It's all well documented and it's not gonna change.
However, I think we should have a discussion about the side effects. Do we have any reasonable way that OVL calculations and wage demands can make sense for sprinters again?
AKA has the second highest OVL in CT, but he is obviously not the second best rider. I think looking down the OVL list and rankings, we can generally agree that the order/gaps in sprinter OVL's (and even more so wages) are much greater than the actual difference in results, more so than any other terrain.
The question is: what can we do given that the sprinter field is so wide? We've seen high OVL's for sprinters lead to market collapse already, where 50k guys can easily be successful CT leaders. Can we make sprinter wages make sense when any free agent is a viable replacement?
Anybody have ideas?
RIP Exxon Duke, David Veilleux, Double Feature, and Monster Energy
As Roturn reference to on Page 1, I hope the Wage calculation is being looked at to be greater affected by Points vs OVL.
I think this would be good for more than just sprinters, and better adjust the game year-to-year on different game versions (especially as what i'm hearing is that PCM20 is much the same as now).
You have a rider who performs well, it makes sense to pay them according that that vs they have a good OVL. Especially as our MG OVL is self made and does not affect PCM. (vs. PCM does directly affect Points, so therefore Points should affect Wage)
![imgur.com/KlJUqDx.jpg[/img[img]]https://imgur.com/yUygrQ.jpg](https://imgur.com/KlJUqDx.jpg[/img[img]]https://imgur.com/yUygrQ.jpg)
