Good to see test results that look promising, however I have to ask with us already running almost a year behind, how long would it take to adjust all the stages? Side note here is that the time spent on testing could possibly also be used on reporting.
Another thing we need to think about is how far we go with this, will we also change races like Milan - San Remo with the iconic Poggio descend and chicane at the end?
Ollfardh wrote:
Good to see test results that look promising, however I have to ask with us already running almost a year behind, how long would it take to adjust all the stages? Side note here is that the time spent on testing could possibly also be used on reporting.
Another thing we need to think about is how far we go with this, will we also change races like Milan - San Remo with the iconic Poggio descend and chicane at the end?
Basically not needed to change races that went pretty well imo.
This year saw a mix of sprinters and roleurs go for it with a late attack by the strongest puncheur being rewarded. This makes MSR special and different from pure sprinter classics.
Editing the road wide though should be possible easily, even in MSR.
Crunched some of the numbers and results from alex's test (perfect layout and info! Really well done and helpful)
Original Startlist:
5 runs, only once did the leadout get the win. Three times it was better to sit behind and benefit from another teams work. Once we were lacking in trains to really tell.
Degenkolb was the winning leadout, but in all 5 runs he never dopped out the Top5, still claiming a win surfing another team.
The main riders being leadout were almost always outside the Top5 in the end. Coquard + Swift combined for 4 Top5's out of a possible 10, Ewan with the most consistent leadout also only making the Top5 twice, and one of those was the no-leadout run. Game related conclusion: better to have no leadout and be Degenkolb
Change 1, More FL focused leadout for Festina:
2 runs, things go worse for Coquad who drops out the Top5 when he gets a leadout. The run with leadouts, again surfing another team is better with the focus sprinters not making the Top10. No leadouts at all goes better for the top sprinters, although in both runs Degenkolb and Ewan are just gone from the Top5 and we have a much bigger variation in Top5 finishers. Game related conclusion: The +FL -SP leadout doesn't seem to work that well, and still better to not have a leadout dedicated to you
Change 2, Upping the SP for Festina:
Ok we only get one run here, but there is a clear improvement for Coquard this time. That said Degenkolb wins again (he's loving these tests!) by surfing other teams work. Ewan now hasn't made the Top5 in 6 straight runs despite usually being given a leadout! Game related conclusion: Still better to not get a leadout, if you're Ewan you are hating your leadout by now and wishing them all fired!
Change 3, RES Boost for Podium:
2 runs, and this has definitely helped Ewan out who gets his first Top5 in 6 tries, and Top5's both times. Although he doesn't have the better leadout in either case and does better surfing off other teams work. Festina still struggling to get Coquard a win, after 10 tries they just can't get him there. Degenkolb has vanished in these runs, whilst Grosu and Ahlstrand are becoming the kings of consistency by surfing other trains. Game related conclusion: The RES and FL boost has definitel helped Podium, but it's still better to not have a leadout made for you and use other teams.
Change 4, Ewan and Coquard Switch Teams:
5 runs here. Holy heck does this make a difference for Coquard as he bags two wins immediately as the leadout works twice in a row! Ewan, however, will only make the Top5 once in all 5 runs. But for the last three runs, we get a mixed leadout and a win for Swift, followed by two cases where the leadout doesn't help and Swift wins and so does Degenkolb. Grosu and Swift are now the most consistent riders, and Groenewgen has suddenly become a regular Top5 rider as Ahlstrand goes backwards. Game related conclusion: I should buy Coquard from Festina! We're getting close to how sprints with leadouts should work, we just needed to put the best sprinter in the game with the best leadout man in the game...
Change 5, Removing Saber from Podium:
3 runs, and somehow this works even better for a Coquard led Podium. Twice his leadout works and he wins. Ewan is also doing better at Festina twice in the Top5. Swift hates this set-up and goes away to join Ahlstrand further down the Top10. Grosu is also loving this and is just behind Degenkolb for consistency. We do see the leadout not work once and Coquard go away from the Top10. Game related conclusion: Maybe Saber was too good, we seem to have found the perfect leadout balance for Coquard (Supa-Nova's high FL, Aberasturi's mid-RES and sub-80 SP/ACC combo (i think that's right?))
Change 6, Adding more Sprint to Festina:
2 final runs, but Podium are still class of the field delivering a win for Coquard once. Grosu gets a leadout in the final run to win, and pretty much ties Degenkolb for consistency as a result. Although both of them benefited by surfing other teams trains most of the time. Ewan can only make the Top5 once, he just doesn't like Festina. Game related conclusion: Festina just can't make a good train, although why Grosu can get a working one right at the end throws a massive spanner in the workings out...
Overall conclusion:
About 7 times, out of 20, the leadout works and delivers a winner. About 3 occasions we get messy leadouts, or no leadouts, and these finishes offer the biggest variation in results. 10 times, that's half the tests, it was much better to surf other teams leadouts. And i mean much better, in these 10 cases the ledout rider doesn't even make the Top5 on 6 occasions!
For a leadout, the game likes that 3rd-man for be high FL rather than a sprinter. Your 2nd-man seems to work best with mid-FL, good RES, and SP 3-5 points lower than your leader.
And even then, in 2 out of 5 finishes with that set-up you're still better off not having the leadout at all.
And of course, no idea what would happen if we had a bump near the finish, or any hills or mountains on the way, and your perfect leadouts struggled with MO or HL...
Game related suggestions: Wage changes could go a long way towards balancing this, upping the cost of those Tier 2 sprinters or lowering that of the Tier 1 guys.
Offering more race day clashes/choices for the top sprinters would help. This would help create bigger stat gaps between riders at races and help reduce the randomness.
Doing something quite radical to reduce SP stat inflation, for the same effect.
____
Feel free to disagree with my conclusions, this is what i read from those results. Happy to talk about them and reach a compromise solution for the game.
These tests are re-inforcing the idea that a leadout is not the best way to go, even that having a top sprinter might not be the way to go and stacking more tier2 guys would give a better points return.
Some more tests. Trying to figure out exactly why some teams have trains most of the time, while others only at times. Also, trying to get a better picture as to what type of rider works better for the first and the second man in the train.
Sadly, I have no exports for the last 11 tests as it looks like the name of the file was beginning to be to big too handle either for my computer or for the game. I will still leave the notes on to give an idea of what was going on.
Startlist change Original startlist, but Aberasturi replaces Nelson. Gorazd Per replaces Petit. I'm making this change because this would make Barbier Festina's first man in the train and he is quite similar to Podium's Alaphilippe.
test 1: podium and volvo. no attempt from festina. zabel, coquard, swift and grosu are behind podium
Spoiler
1
Jonas Ahlstrand
Volvo acc. by Spotify
4h00'08
2
Fabio Silvestre
Berg Cycles
s.t.
3
Caleb Ewan
Podium Ambition
s.t.
4
Bryan Coquard
Festina - OAKA
s.t.
5
Rick Zabel
Desigual
s.t.
6
Ben Swift
Aker - MOT
s.t.
7
Maurits Lammertink
eBuddy
s.t.
8
Yudai Arashiro
Rakuten Pro Cycling
s.t.
9
Ho-Ting Kwok
Air France - KLM
s.t.
10
Eduard Grosu
eBuddy
s.t.
11
Johann van Zyl
Kraftwerk Man Machine
s.t.
12
Marildo Yzeiraj
Aker - MOT
s.t.
13
Kris Boeckmans
Campari/Asahi development
s.t.
14
Juan Abenhamar Gallego Martin
Iberia - Team Degenkolb
s.t.
15
Mark Cavendish
Kraftwerk Man Machine
s.t.
16
Sten Stenersen
Volvo acc. by Spotify
s.t.
17
Jon Aberasturi
Podium Ambition
s.t.
18
Barry Markus
Volvo acc. by Spotify
s.t.
19
Willie Smit
EA Vesuvio
s.t.
20
Kenji Itami
Rakuten Pro Cycling
s.t.
test 2: only podium make a train and with coquard in ewan's wheel they get a slight gap to the rest. grosu leads the "chasing" train behind
Spoiler
1
Bryan Coquard
Festina - OAKA
4h00'56
2
Ben Swift
Aker - MOT
s.t.
3
Caleb Ewan
Podium Ambition
s.t.
4
Jonas Ahlstrand
Volvo acc. by Spotify
s.t.
5
Mark Cavendish
Kraftwerk Man Machine
s.t.
6
John Degenkolb
Iberia - Team Degenkolb
s.t.
7
Romain Vanderbiest
Carrefour - ESPN
s.t.
8
Eduard Grosu
eBuddy
s.t.
9
Kenji Itami
Rakuten Pro Cycling
s.t.
10
Oscar Guerao
SPAR - Siam Cement
s.t.
11
Andrea Guardini
Fablok - Bank BGZ
s.t.
12
Oscar Avelino
Duolingo
s.t.
13
Jon Aberasturi
Podium Ambition
s.t.
14
Rick Zabel
Desigual
s.t.
15
Dylan Groenewegen
Aegon - Lavazza
s.t.
16
Mohamed Harrif Salleh
Sauber Petronas Racing
s.t.
17
Kris Boeckmans
Campari/Asahi development
s.t.
18
Peter Kennaugh
Air France - KLM
s.t.
19
Sten Stenersen
Volvo acc. by Spotify
s.t.
20
Fabio Silvestre
Berg Cycles
s.t.
test 3: festina are the strongest. podium and volvo also make trains but not as strong. swift is behind the festina train
Spoiler
1
Bryan Coquard
Festina - OAKA
3h59'25
2
Ben Swift
Aker - MOT
s.t.
3
Caleb Ewan
Podium Ambition
s.t.
4
Oscar Guerao
SPAR - Siam Cement
s.t.
5
Peter Kennaugh
Air France - KLM
s.t.
6
Jonas Ahlstrand
Volvo acc. by Spotify
s.t.
7
Romain Vanderbiest
Carrefour - ESPN
s.t.
8
Mohamed Harrif Salleh
Sauber Petronas Racing
s.t.
9
John Degenkolb
Iberia - Team Degenkolb
s.t.
10
Dylan Groenewegen
Aegon - Lavazza
s.t.
11
Fabio Silvestre
Berg Cycles
s.t.
12
Mark Cavendish
Kraftwerk Man Machine
s.t.
13
Rick Zabel
Desigual
s.t.
14
Oscar Avelino
Duolingo
s.t.
15
Maxime Vantomme
Minions
s.t.
16
Eduard Grosu
eBuddy
s.t.
17
Davide Appollonio
Carlsberg - Danske Bank
s.t.
18
Kris Boeckmans
Campari/Asahi development
s.t.
19
Danilo Kupfernagel
Berg Cycles
s.t.
20
Pedro Merino Criado
Iberia - Team Degenkolb
s.t.
test 4: podium and festina are the strongest. iberia and ebuddy are also trying but not as fast, but all four trains are kind of messy, especially the leading two, the sprinters are not very well co-ordinated with their leadouts
Spoiler
1
Ben Swift
Aker - MOT
4h32'36
2
Mark Cavendish
Kraftwerk Man Machine
s.t.
3
Eduard Grosu
eBuddy
s.t.
4
Andrea Guardini
Fablok - Bank BGZ
s.t.
5
Romain Vanderbiest
Carrefour - ESPN
s.t.
6
John Degenkolb
Iberia - Team Degenkolb
s.t.
7
Dylan Groenewegen
Aegon - Lavazza
s.t.
8
Bryan Coquard
Festina - OAKA
s.t.
9
Oscar Avelino
Duolingo
s.t.
10
Rick Zabel
Desigual
s.t.
11
Jonas Ahlstrand
Volvo acc. by Spotify
s.t.
12
Caleb Ewan
Podium Ambition
s.t.
13
Kris Boeckmans
Campari/Asahi development
s.t.
14
Jon Aberasturi
Podium Ambition
s.t.
15
Andreas Stauff
Festina - OAKA
s.t.
16
Peter Kennaugh
Air France - KLM
s.t.
17
Kenji Itami
Rakuten Pro Cycling
s.t.
18
Jurgen Roelandts
Volvo acc. by Spotify
s.t.
19
Barry Markus
Volvo acc. by Spotify
s.t.
20
Marco Brus
eBuddy
s.t.
test 5: split with 10km to go. Everyone came back together with 6km to go, but some leadouts have definitely been caught too far behind. Stauff and Brus are tryint to control the sprint but there's not coherent train. coquard left in the wind alone with 1.5k to go
Spoiler
1
Eduard Grosu
eBuddy
4h03'05
2
Dylan Groenewegen
Aegon - Lavazza
s.t.
3
John Degenkolb
Iberia - Team Degenkolb
s.t.
4
Ben Swift
Aker - MOT
s.t.
5
Peter Kennaugh
Air France - KLM
s.t.
6
Andrea Guardini
Fablok - Bank BGZ
s.t.
7
Kenji Itami
Rakuten Pro Cycling
s.t.
8
Jonas Ahlstrand
Volvo acc. by Spotify
s.t.
9
Kris Boeckmans
Campari/Asahi development
s.t.
10
Caleb Ewan
Podium Ambition
s.t.
11
Bryan Coquard
Festina - OAKA
s.t.
12
Oscar Avelino
Duolingo
s.t.
13
Mohamed Harrif Salleh
Sauber Petronas Racing
s.t.
14
Marildo Yzeiraj
Aker - MOT
s.t.
15
Oscar Guerao
SPAR - Siam Cement
s.t.
16
Andreas Stauff
Festina - OAKA
s.t.
17
Benedikt Mundle
Sauber Petronas Racing
s.t.
18
Ho-Ting Kwok
Air France - KLM
s.t.
19
Yudai Arashiro
Rakuten Pro Cycling
s.t.
20
Mark Cavendish
Kraftwerk Man Machine
s.t.
test 6: ebuddy vs festina. podium are also there but they seem to have some trouble with coordonitation. The wind is very strong and the riders that waited for a late launch are definitely helped by this.
Spoiler
1
Ben Swift
Aker - MOT
4h14'57
2
John Degenkolb
Iberia - Team Degenkolb
s.t.
3
Dylan Groenewegen
Aegon - Lavazza
s.t.
4
Oscar Guerao
SPAR - Siam Cement
s.t.
5
Kris Boeckmans
Campari/Asahi development
s.t.
6
Romain Vanderbiest
Carrefour - ESPN
s.t.
7
Mohamed Harrif Salleh
Sauber Petronas Racing
s.t.
8
Caleb Ewan
Podium Ambition
s.t.
9
Eduard Grosu
eBuddy
s.t.
10
Peter Kennaugh
Air France - KLM
s.t.
11
Rick Zabel
Desigual
s.t.
12
Oscar Avelino
Duolingo
s.t.
13
Kenji Itami
Rakuten Pro Cycling
s.t.
14
Mark Cavendish
Kraftwerk Man Machine
s.t.
15
Bryan Coquard
Festina - OAKA
s.t.
16
Andreas Stauff
Festina - OAKA
s.t.
17
Jonas Ahlstrand
Volvo acc. by Spotify
s.t.
18
Andrea Guardini
Fablok - Bank BGZ
s.t.
19
Ho-Ting Kwok
Air France - KLM
s.t.
20
Jon Aberasturi
Podium Ambition
s.t.
test 7: split in the pack with 10k to go. podium vs ebuddy, but grosu only has vemeltfoort in front of him. coquard is behind grosu
Spoiler
1
Kenji Itami
Rakuten Pro Cycling
4h09'09
2
Peter Kennaugh
Air France - KLM
s.t.
3
Caleb Ewan
Podium Ambition
s.t.
4
Bryan Coquard
Festina - OAKA
s.t.
5
Rick Zabel
Desigual
s.t.
6
Mark Cavendish
Kraftwerk Man Machine
s.t.
7
Eduard Grosu
eBuddy
s.t.
8
Kris Boeckmans
Campari/Asahi development
s.t.
9
Dylan Groenewegen
Aegon - Lavazza
s.t.
10
Ben Swift
Aker - MOT
s.t.
11
Andrea Guardini
Fablok - Bank BGZ
s.t.
12
Oscar Avelino
Duolingo
s.t.
13
Marco Brus
eBuddy
s.t.
14
Jonas Ahlstrand
Volvo acc. by Spotify
s.t.
15
Mohamed Harrif Salleh
Sauber Petronas Racing
s.t.
16
John Degenkolb
Iberia - Team Degenkolb
s.t.
17
Oscar Guerao
SPAR - Siam Cement
s.t.
18
Romain Vanderbiest
Carrefour - ESPN
s.t.
19
Jon Aberasturi
Podium Ambition
s.t.
20
Ho-Ting Kwok
Air France - KLM
s.t.
test 8: volvo vs festina. ebuddy a bit slower. swift and degenkolb attempt to jump to the volvo train
Spoiler
1
Bryan Coquard
Festina - OAKA
3h55'05
2
Jonas Ahlstrand
Volvo acc. by Spotify
s.t.
3
Ben Swift
Aker - MOT
s.t.
4
Eduard Grosu
eBuddy
s.t.
5
Peter Kennaugh
Air France - KLM
s.t.
6
Mohamed Harrif Salleh
Sauber Petronas Racing
s.t.
7
Ho-Ting Kwok
Air France - KLM
s.t.
8
Oscar Guerao
SPAR - Siam Cement
s.t.
9
Andrea Guardini
Fablok - Bank BGZ
s.t.
10
Caleb Ewan
Podium Ambition
s.t.
11
Kenji Itami
Rakuten Pro Cycling
s.t.
12
Mark Cavendish
Kraftwerk Man Machine
s.t.
13
Oscar Avelino
Duolingo
s.t.
14
Yudai Arashiro
Rakuten Pro Cycling
s.t.
15
Fabio Silvestre
Berg Cycles
s.t.
16
John Degenkolb
Iberia - Team Degenkolb
s.t.
17
Rick Zabel
Desigual
s.t.
18
Dylan Groenewegen
Aegon - Lavazza
s.t.
19
Davide Appollonio
Carlsberg - Danske Bank
s.t.
20
Kris Boeckmans
Campari/Asahi development
s.t.
test 9: aker and podium. But aker only have oss in the train who tires himselfout pretty quickly and swift and those following him are quick to jump on ewan's wheel. swift, coquard, kennaugh, degenkolb, groenewegen and grosu are all following the podium train. I think the podium train worked to perfection here. Aberasturi was still there under the 1k flag, it's just that Coquard, Dege and Swift are superior sprinters. Ewan held on well and finish was very close.
Spoiler
1
Bryan Coquard
Festina - OAKA
3h58'33
2
Ben Swift
Aker - MOT
s.t.
3
John Degenkolb
Iberia - Team Degenkolb
s.t.
4
Caleb Ewan
Podium Ambition
s.t.
5
Peter Kennaugh
Air France - KLM
s.t.
6
Dylan Groenewegen
Aegon - Lavazza
s.t.
7
Eduard Grosu
eBuddy
s.t.
8
Kris Boeckmans
Campari/Asahi development
s.t.
9
Oscar Avelino
Duolingo
s.t.
10
Jon Aberasturi
Podium Ambition
s.t.
11
Yudai Arashiro
Rakuten Pro Cycling
s.t.
12
Marildo Yzeiraj
Aker - MOT
s.t.
13
Ho-Ting Kwok
Air France - KLM
s.t.
14
Vojtech Hacecky
Fablok - Bank BGZ
s.t.
15
Janis Dakteris
Aker - MOT
s.t.
16
Ben King
Iberia - Team Degenkolb
s.t.
17
Coen Vermeltfoort
eBuddy
s.t.
18
Jamalidin Novardianto
Podium Ambition
s.t.
19
Simon Clarke
Air France - KLM
s.t.
20
Benedikt Mundle
Sauber Petronas Racing
s.t.
test 10: volvo vs podium. iberia are there too but only with 1 man. festina trying but they are bit furter back. Weird sprint with big dinstances between the sprinters and their lead outs with 2 k to go. Aberasturi and Roelants were flying, but Ewan and Alhstrad were not immediately in their wheels.
Startlist change Barbier-Stauff seems to be working better than Petit-Stauff, but not as consistent as Alaphilippe-Aberasturi. Is ala the the king of locomotives? Ala and Barbier switch teams.
test 11: festina are by far the only train that its shaped up here. podium are trying too but Ewan only has one rider with him and it's... Postlberger. Itami, Dege and Cav are behind Ewan, Swift and ahlstrand behind coquard
barbier and aberasturi eventually made a train on the other side of the road. ewan is quick to jump on it, but it's very short lived and he is forced to go back to the festina line of slipstream
test 12: podium vs volvo. but ewan only had aberaturi with him and ahlstand only had roelants. left in the wind too early, sprinters that are behind take advantage of this inside last km
test 13: volvo are the strongest here. coquard only has stauff with him. roelants is a beast and creates a split. he takes ahlstrand and coquard with a him
test 14: festina lead the way. volvo, podium and ebuddy are trying but not as fast. swift is behind coquard, while volov and podium are all over the place
Startlist change Stauff is there for Coquard, but to me he seems to be the weak link due to his low res. Stauff out, Petit in. French dream team for Festina
test 15: no sprint trains. late attacks. everyone is all over the place
test 16: festina, volvo, ebuddy, podium are all trying but everyone is packed and it's hard to follow the wheels of your team-mates
test 17: late attacks. no trains. every man for himself
test 18: festina vs aker. festina look sharper though as swift only has oss with him. Groenewegen behind coquard, degenkolb and others are following swift, but as oss tires himself out, everyone jumps ships to festina. petit takes over at 1.6, coquard goes at 0.8. this is the perfect lead out in my opinion. coquard jump at the perfect time and never looked like slowing down but swift was right behind coquard and resulted in a very close finish.
test 19: volvo are the only train. they get a gap to the others as they take degenkolb with them. ewan and grosu lead the "Chase"
test 20: again a very unorganized pack. volvo are the only train. everyone follows it
TheManxMissile wrote: Overall conclusion:
About 7 times, out of 20, the leadout works and delivers a winner. About 3 occasions we get messy leadouts, or no leadouts, and these finishes offer the biggest variation in results. 10 times, that's half the tests, it was much better to surf other teams leadouts. And i mean much better, in these 10 cases the ledout rider doesn't even make the Top5 on 6 occasions!
I mentioned this earlier somewhere, but just keep in mind that when you're comparing wins with leadout vs. wins without, you're comparing the wins of two to at most three riders to the wins of everyone else. I haven't had the time to look at the results specifically, but hypothetically if you have two riders consistently with leadouts and they score 4 wins each in a sample of 20 races, and then the other 12 wins are spread among maybe 6 riders, then you have the guys with leadouts winning just 8 of 20, but still individually double the chance of winning than any single rider without a leadout.
Of course that is a very simplified example which doesn't take into account whether the guys with leadouts are actually stronger sprinters and wouldn't change the fact that riders with leadouts may end up outside of the Top 5 more often than you'd like, but it's still something to note when evaluating the absolute numbers I think.
TheManxMissile wrote:____
Feel free to disagree with my conclusions, this is what i read from those results. Happy to talk about them and reach a compromise solution for the game.
These tests are re-inforcing the idea that a leadout is not the best way to go, even that having a top sprinter might not be the way to go and stacking more tier2 guys would give a better points return.
It's really all still very hard to pin point exactly what works and does not. I think a big impact also is made by when the breakaway is caught in terms of how well trains are made.
For me, it shows that lead-out trains do work. But if you do not have a reliable lead-out you are indeed better off following someone else. It's very important that the lead out does not leave the sprint alone before the final kilometer. But even then, if you are Coquard and you have Swift behind you, sometimes you can held him off, sometimes you can't. The top 6 (Coquard, Swift, Degenkolb, Ewan, Ahlstrand and Grosu) are very close between each other so every extra second spent in the slipstream could count in this case.
I think the role of a lead-out is to bring the sprinter in a position where he can fight for the win. After that, it's up to the sprinter to deliver. And these tests do show that a reliable layout can put you in a good position.
It's important to look if a good lead-out delivers more consistency, not necessary the victory. A big problem for Coquard is with his current set up he misses out on the top 10 quite often. But with a proper lead out this was rarely the case.
cunego59 wrote:
[quote][url=https://pcmdaily.com/forum/viewthread.php?thread_id=55199&
I mentioned this earlier somewhere, but just keep in mind that when you're comparing wins with leadout vs. wins without, you're comparing the wins of two to at most three riders to the wins of everyone else. I haven't had the time to look at the results specifically, but hypothetically if you have two riders consistently with leadouts and they score 4 wins each in a sample of 20 races, and then the other 12 wins are spread among maybe 6 riders, then you have the guys with leadouts winning just 8 of 20, but still individually double the chance of winning than any single rider without a leadout.
Of course that is a very simplified example which doesn't take into account whether the guys with leadouts are actually stronger sprinters and wouldn't change the fact that riders with leadouts may end up outside of the Top 5 more often than you'd like, but it's still something to note when evaluating the absolute numbers I think.
If you take Coquard, Swift and Degenkolb as the top3 sprinters you would expect them to do roughly the same across the tests.
Degenkolb comes out as by far the most consistent with his surfing method. Coquard is second, but is kind of an all or nothing, winning thanks to his leadout or being swamped thanks to his leadout.
Swift comes along with a mixed bag from surfing.
These three are combining for majority of wins, although almost all Coquards good result come after he changes teams + gets given the ideal leadout. Degenkolb's wins come almost entirely off surfing another wheel.
Ewan is a bit behind in terms of ability, but gets a lot of leadout. If a leadout was beneficial he should be able to match the better sprinters without leadouts across the tests. But he doesn't. He's a ways behind Degenkolb, can't match Coquard for wins and is passed by Grosu surfing most days.
After the switch to Festina he does worse, which throws the result a bit. Want to see him with the Aberasturi leadout support.
He does come off better than Swift overall and Kennaugh, Groenewgen, Cavendish and most other riders across the tests, but no more so than you'd expect looking at stats. The leadout might hurt compared to them, but it's not helping either...
From what we're seeing in these tests, being number 1 or 2 behind the main leadout guy is the better way to go. It gives you more wins, more podiums and more consistent Top5's vs that led out sprinter.
I'm also trying to look at the wider picture. Saber, Aberasturi, Petit and these leadouts we talk about are not cheap! I know, i've tried to buy most of them in the past couple of years. And in FA their wages far outweigh the points gain Ewan would get vs picking up Nelson or Kek.
It's why i'm trying to focus more on OVL + Renewal changes in my suggestions, because the AI we can't change.
If we come back to Cunego's tests: he shows us that if you give the Top3 sprinters leadouts, and no-one else, they are distinctly worse in results than the rest. Their basic strength as sprinters does not help overcome the benefit weaker sprinters get surfing behind them.
Alex's tests support this theory. You can get more consistent, and better results compared to your riders skill, by surfing than with a leadout.
~~~
Unless you specifically design the perfect leadout for the best sprinter. I would like to see a similar test to these were we give the perfect lead out to Avelino or Cav, and nothing to Coquard/Ewan/Grosu.
____
@Alex
Easier to edit this than quote another post
Your tests have shown a lot more about what makes a good leadout train, but also that even then you are still at the mercy of the game AI and decision making. And that PCM18 AI, like every PCM AI, is poor when it comes to sprinting.
If people think this is fine, i can't argue to change against the majority of managers. What i can, and hope to, do is just run with my theory to stack sprinters and not try to assemble a leadout for races. See if across a season 6 80+ sprinters can deliver enough results to keep me in the PCT
(SotD, you need to sell me Coquard and watch him destroy! The game wants it to happen!)
Edited by TheManxMissile on 26-04-2020 14:49
Look at my latest tests where we do have results for. Out of the five times Ewan got a lead out (a proper one not by Ricki "64 resistance" Nelson) he finished 3 times on the podium and once he was fourth. His results without a train include a podium but not much beyond that. I think that shows consistency when it comes to trains.
alexkr00 wrote:
Look at my latest tests where we do have results for. Out of the five times Ewan got a lead out (a proper one not by Ricki "64 resistance" Nelson) he finished 3 times on the podium and once he was fourth. His results without a train include a podium but not much beyond that. I think that shows consistency when it comes to trains.
You're timing was pefect
I'm not really annoyed with Ewan in the actual season, Top3 ranking sprinter still.
And we seem to get more consistent leadouts with the redesigned stage!
Figuring out the importance of RES is good to know, something that can be fed into OVL which in turn feeds into Renewals.
TheManxMissile wrote:
If we come back to Cunego's tests: he shows us that if you give the Top3 sprinters leadouts, and no-one else, they are distinctly worse in results than the rest. Their basic strength as sprinters does not help overcome the benefit weaker sprinters get surfing behind them.
Just one quick note: The problem wasn't that the Top 3 sprinters had leadouts and were overtaken. The problem was that eventhough they were the only ones with good sprinters as teammates, they didn't form trains and instead sat back with those teammates, hardly ever participating in the sprints or if they did coming from too far behind. Other teams set up sprint trains despite not having sprinters and were more successful. As I said, in my tests, the winner was from a sprint train ~70% of the time, it was just not clear to me why certain teams were the ones setting them up. But alex's tests are not as clear, so I still struggle to draw conclusions.
TMM said:
It's why i'm trying to focus more on OVL + Renewal changes in my suggestions, because the AI we can't change.
This is something likely to happen I guess.
We once upped the sprinters OVL due to them being quite overpowered point/wage wise.
Now we see the opposite really and so could move back to the "old" system, where sprinters had a lower OVL and this way less wage and/or more race days.
In especially the upped OVL for leadout riders might be reduced.
Surely calculations must be made, but this is very likely a good method indeed.
That all said, difficult really as this also means, that those secondary sprinters right now, are scoring pretty decent and would be even cheaper as well. And also they are leaders in lower divisions, e.g. Krieger/Moser in CT doing pretty well and would be on a cheaper wage then as well.
alexkr00 wrote:
Look at my latest tests where we do have results for. Out of the five times Ewan got a lead out (a proper one not by Ricki "64 resistance" Nelson) he finished 3 times on the podium and once he was fourth. His results without a train include a podium but not much beyond that. I think that shows consistency when it comes to trains.
You're timing was pefect
I'm not really annoyed with Ewan in the actual season, Top3 ranking sprinter still.
And we seem to get more consistent leadouts with the redesigned stage!
Figuring out the importance of RES is good to know, something that can be fed into OVL which in turn feeds into Renewals.
I don't know if RES is necessarily a make or break stat, but it's definitely good to have one above 70. 64 is definitely way too low to for Nelson to lead-out properly as he burns way too fast when he is left in the win. I'm suspecting if he were to follow wheels it wouldn't be so noticeable (since Guerao and Avelino still can make the good result from time to time).
For the first man in the train STA I believe to play a bigger role, but I'm only basing this on looking at the riders that seem to be leading their train better than others. STA and Fighter stat are the only stats where Alaphilippe is better than Barbier, yet he is a way better train rider than the Festina rider. Vermeltfoort and Markus (eBuddy and Volvo train leaders) also seem to be good riders to pick for leading their train and they also have a good STA. Don't know much RES counts here since Ala is only 69.
Again this is just speculation and nothing concrete.
Interesting work going on here guys. While reading through I jotted down a few thoughts:
* If Dege is being perhaps the most successful overall thus far (by 'surfing' ), what happens if he is given the position of the 'best' leadout. Does his success now drop, or is he actually the 'best' sprinter?
* I remain unsure why certain teams end up being those with a lead-out. Volvo in many of the tests seem to keep getting the opportunity at building a lead-out for Ahlstrand for example - is it related to the AI's perception of the best lead-out train fit to the model it is seeking, or the overall strength/suitability of the team? What happens if you take the most consistently occuring lead-out teams and sub in 5 lvl 1 riders for the 5 squad members not participating in the lead-out. (This may also have the impact though of meaning that the BotD does not get caught, but that too would be relevant to get some feel for as well given that many MG GT teams for instance have 2-3 dev riders in their line-up).
* Have there been many tests where an absolute flat beast is subbed in as first lead-out (in the tests would seem that Ala is working real well as man 1 - the conclusion thus far seems that too strong a man 1 (in terms of sprint) is not working well). Say Dzamastagic or someone similar stat-wise (or if easier in terms of trying to draw conclusions, just boost Ala's FL in test DB to 77/78 to see if makes big impact). The final part of this would then be to reduce the SP/ACC of the man 1 rider so that more of the Oss model Fl 78, SP/ACC = 73/74.
I think that Sprints are by far the biggest PCM AI issue in the context of the MG. As someone with some experience of building numerical models of complex dynamic systems, I think that this is unlikely to be an area where Cyanide are able to make significant improvements beyond where they are now given how many goes at it they've had already! Basically if the MG community can be presented with as good a view as possible as to what seems to statistically 'work' more regularly vs. not work, that'll then lead to automatic response of managers in off-season FA/TM activity to re-value SPR type riders. Folks here seem to be doing an admirable job of achieving this, so thanks very much for all the work/discussion being put forward.
I'll finish with what may well be a somewhat controversial view, but I believe that PCM 18 alterations to the Mo/Hi/Acc balance make complete and logical sense. All professional cyclists can be defined using a power curve - nowadays the basis of good coaching practise. No professional (road) rider has a power curve that can in anyway be mapped to the 1-d puncheur types we have in the MG DB. The closest I can compare those stat lines to logically would be the guys on the British hill-climb circuit (end of season TT's up steep hills of 2-5km length). Often the top 3 in the British Champs have little to no road results even at amateur level, but can beat road professionals in these very specific events (in the same way that modern day track sprinters cant/dont moonlight as road professionals, but would if all stage lengths were less than 5km!). It is a significant change in the perceived wisdom of the MG and in particular the way 'hill' riders have more prevalently been developed - this though was only in response to having gained an understanding of what the PCM AI (was) valuing. As this previous model we had all been learning to attune to has no reasonable basis, think how confusing this has continuously been for new players joining the MG over the last number of seasons! The true puncheur type 1-day races seem to be working not dis-similar to previously. Stage races not so much, but in my opinion a lot of this is only defined by preconception. If the intention is that the Tour of the Basque Country for example should be a PT race tailored to puncheurs, then it will necessitate not using routes based upon the real race (which is undoubtedly a parcours for stage racers).
TheManxMissile wrote: Game related suggestions: Wage changes could go a long way towards balancing this, upping the cost of those Tier 2 sprinters or lowering that of the Tier 1 guys.
Offering more race day clashes/choices for the top sprinters would help. This would help create bigger stat gaps between riders at races and help reduce the randomness.
Doing something quite radical to reduce SP stat inflation, for the same effect.
Only problem with this solution is that who qualifies as "tier 1" and "tier 2" is not static. Those tier two sprinters will be the top sprinters if they end up in lower divisions. The matrix is always relative to divisions, but it's one market with plenty of overlap, which makes things more difficult to control than stratifying by stats alone.
Totally agree with more race day clashes if we feel that bigger stat gaps between leaders gives the favorites more of an advantage (which in theory of course should always be the case).
It seems to me that any difficulty stage difficulty, including wind, leads to the leadout rider and hence their sprinter launching far too early. Trains will not slow down their effort to accommodate for timing, they just burn out. So less difficulty means more successful leadouts.
Can I ask someone more knowledgeable if we know exactly what the Sprint and Acceleration statistics (and Flat and Resistance for that matter) do in PCM 18? To my understanding in PCM 15 sprint controlled top speed and acceleration controlled the time it took to increase speed, but I know that wasn't always the case. In PCM 11 for instance acceleration controlled how fast red bar drained, like resistance for attacks. Have their been any changes in PCM 18 that fundamentally change how these stats work?
Add Stamina and Fighter into the mix as well I suppose. I was under the understanding that those stats should have basically no effect on a rider's leadout duties with how we've understood them in the past, at least unless the stage was very long, but Alaphillipe's stardom potentially showed otherwise.
My inclination is still that top speed matters less and slipstream matters more in PCM 18. Within a reasonable stat differential, it seems the sprinter on another's wheel will almost always have enough speed to pop out and overtake, it's just a matter of whether they time it right. A better leadout for the rider in front prevents them from having enough time. Maybe this is just a matter of energy depleting much faster in sprints than in years early (especially yellow), or maybe it's something else in terms of mechanics. Or maybe I'm completely wrong. I'm just spitballing.
RIP Exxon Duke, David Veilleux, Double Feature, and Monster Energy
Scorchio wrote:
* I remain unsure why certain teams end up being those with a lead-out. Volvo in many of the tests seem to keep getting the opportunity at building a lead-out for Ahlstrand for example - is it related to the AI's perception of the best lead-out train fit to the model it is seeking, or the overall strength/suitability of the team? What happens if you take the most consistently occuring lead-out teams and sub in 5 lvl 1 riders for the 5 squad members not participating in the lead-out. (This may also have the impact though of meaning that the BotD does not get caught, but that too would be relevant to get some feel for as well given that many MG GT teams for instance have 2-3 dev riders in their line-up).
Podium Ambition were almost always getting a train. After I took Alaphilippe out of the team, the number of times they got a train versus the number of time they didn't pretty much reversed.
What I think the game does is looks at the best 5-6 sprinters (I have yet to see in the tests train made for riders other than Ewan, Coquard, Grosu, Ahlstrand, Degenkolb and Swift) and picks them as the teams that will be "organizing" the sprint.
Then it's up to the strength of the riders in those lead-outs how well it goes. And it looks like if the train is set up correctly is often dependent of the first rider in the train. But we can't ignore the second one either since in the few tests I did with Petit and Gaviria they seemed to work well together, while this season, not only the tests show us that Petit doesn't work as well with Stauff.
To reiterate on my last post regarding sta and res. I just played two sprints in my career. One in a hilly stage that flattens out towards the end and a flat classic. I usually don't look at what bar empties first but this time I did and the results were quite similar in both occasions.
I believe the yellow bar is sta and red bar is res, but someone more knowledgeable of the game can perhaps correct me on this.
The first rider in the train that does not actually sprint is definitely running out of yellow faster than red. Only when increasing the intensity of the effort close to 99 would the red bar run faster. 99 is basically an attack or a sprint.
At an effort of 85 the yellow bar was running faster than the red bar. Not by much, but enough to be noticeable. Since this kind of effort usually puts him a bit above the rest of the bunch or at the same level as the rest of the group, I'd like to make the assumption that somewhere around 85 or even a bit lower is where the AI usually sets up the first man in the train.
Now, when the actual sprint started the rider in the wind was running very fast out of red bar, while the one behind him was running out of it as well but at with a slower speed.
Thanks alex. Still not 100% clear to me though whether the AI's perceived strength of the lead-out or overall team strength is determinant - you replaced Ala with a dev rider I think, this might impact both inputs (PA lead-out strength and PA overall team strength).
Without wishing to make things even more complicated, if we are able to confidently dismiss the overall strength of a team in terms of lead-out selection, if that team is in fact weak, some of the lead-out riders might get used up (or at least somewhat) if they end up having to participate in chasing the BotD during the stage. and hence underperform their lead-out potential.
Of course is very hard to isolate specific factors in these tests without running 100's of times in each case to get a sound statistical basis! Chapeau to those providing good evidence towards the discussion
In response to BBL - while trying to understand these things, this was the best input I could find on Daily: https://pcmdaily.com/forum/viewthread...rowstart=0 - note the change in perception of how STA is working
EDIT As a further note re: the link above - the description of the impact of TT vs Prologue stat makes me inclined to suggest we need to incorporate this somewhat into rider OVR calc. When playing in the MG DB spreadsheet, I get practically no impact on OVR from changing the prologue stat whereas TT has a major impact.
Edited by Scorchio on 26-04-2020 16:37
That clears things up a lot actually, thank you Alex.
I was under the impression that resistance still affected yellow and stamina green. If that has changed, it explains some of my confusion. I knew there was now the element of yellow bar decreasing its cap throughout the race, so I figured that would be the stamina influence, rather than actual yellow bar usage. And with resistance taking on the old role of acceleration, that means it is way more important than ever before in sprints.
So perhaps it's not about difficulty of the stage, but simply how energy diminishes in sprints now. In past PCM's, as long as you went from <1.2 km out or so in a pure flat stage, you would not crack before the finish. If someone tried to come out from your wheel, they would get a slight slipstream boost and not be losing energy as fast just like now, but I found energy wasn't super important if you were both pretty full to begin with. As long as you didn't go super early, neither sprinter would run out of yellow, and possibly neither would run out of red. So it was possible for several sprinters to be holding their top speed concurrently, and the rider in front to remain competitive speed wise with the rider coming from behind him. This will theoretically reward top sprinters more, since they have the higher top speeds. Whether that's a good thing is debatable obviously.
Clearly that's not the case anymore, because momentum is playing a bigger role and a lot of sprinters do crack by the end of the sprint because they run entirely out of energy. This means they can lose big places more easily if they go too early. Hence why a leadout train can be very helpful if left late enough, but be catastrophic for it's sprinter if it leaves them in the wind too early. This was always the case of course, but the margin for error now is less forgiving.
If what you say about stats is true, this will most definitely be affected by the low stamina and resistance stats of sprinters and their leadouts, more so than ever before. The question is whether an increase of these stats overall would lead to more consistency in sprint results if the distribution stayed the same. And of course, would it hurt sprint trains? My gut says it would help them last longer and make them more effective, but it could also mean top sprinters won't need them at all. Potentially a double edged sword. Plus, even if it is beneficial, how could it be implemented?
Anybody know what flat does now, by the way?
RIP Exxon Duke, David Veilleux, Double Feature, and Monster Energy
Scorchio wrote:
Thanks alex. Still not 100% clear to me though whether the AI's perceived strength of the lead-out or overall team strength is determinant - you replaced Ala with a dev rider I think, this might impact both inputs (PA lead-out strength and PA overall team strength).
Without wishing to make things even more complicated, if we are able to confidently dismiss the overall strength of a team in terms of lead-out selection, if that team is in fact weak, some of the lead-out riders might get used up (or at least somewhat) if they end up having to participate in chasing the BotD during the stage. and hence underperform their lead-out potential.
I swapped Ala for Barbier, who had the same role at Festina in that startlist. I mentioned it.
Festina is indeed weaker than Podium altogether and I'm quite sure I saw Ala pacing after the breakaway once while riding for Festina. No train for Festina that time.
@bbl I wouldn't take the things I said as 100% percent accurate. Flat and acceleration could still be a big factor.
But by looking at Ala vs Barbier the conclusion for me it's that the sta makes the difference. And then for Nelson and Stauff I think it's pretty clear it's the res stat. This seems to be highly affected though only when it's up to Nelson or Stauff to be leading the sprint. It seems like Brus is working a bit better for Grosu (despite a low res) because when they do make the train, eBuddy are not the dominant one, but just putting Grosu into a good position. And even then it's obvious he burns out very fast.
EDIT: now looking at the link with the stats discussion. For me it does not make sense for res to be the yellow bar. In attacks and sprints its the red bar that runs out faster. In longer less intense efforts, it's the yellow bar that runs out.
Edited by alexkr00 on 26-04-2020 16:57
My mistake in identifying Ala swap. My concern with using Barbier as the comparator (even in this really quite flat stage) is Mo = 50; unclear to me whether this will mean Barbier is 100% fresh.
![imgur.com/KlJUqDx.jpg[/img[img]]https://imgur.com/yUygrQ.jpg](https://imgur.com/KlJUqDx.jpg[/img[img]]https://imgur.com/yUygrQ.jpg)
