PCM.daily banner
22-11-2024 13:51
PCM.daily
Users Online
· Guests Online: 73

· Members Online: 0

· Total Members: 161,784
· Newest Member: Anthonyslamb
View Thread
PCM.daily » Pro Cycling Manager 2006-2020 » Pro Cycling Manager 2012
 Print Thread
PCM.daily DB: Support and Comments
TankNL
Still working on the new race program. It's coming allong nicely, but will have to do some remodeling to the fitness scheme's too. For instance; there is a leader_specialisation; Giro_noworld and Giro_world. Both have the same fitness program. That doesn't seem logic at all. Will also correct that.
 
TankNL
Still working on the new race program. It's coming allong nicely, but will have to do some remodeling to the fitness scheme's too. For instance; there is a leader_specialisation; Giro_noworld and Giro_world. Both have the same fitness program. That doesn't seem logic at all. Will also correct that.
 
Ad Bot
Posted on 22-11-2024 13:51
Bot Agent

Posts: Countless
Joined: 23.11.09

IP: None  
TankNL
I'm still working on the new program's and new fitness program's, but I need your feedback.

For instance. Lets take the leader_type 15 (cya_giro). This has 6 specializations;
11 GIRO_SOLO_NOWORLD
12 GIRO_SOLO_WORLD
13 GIRO_ARDENNAISES
14 GIRO_TOUR_NOWORLD
15 GIRO_TOUR_WORLD
16 GIRO_VUELTA

There is a possibility to link fitness programs to the specializations. But when I look at the giro fitness programs now, then most of them also have a fitness_peak in September or October. I will need to link them to the specializations with the worlds in them, but what about specializations like 11 (GIRO_SOLO_NOWORLD). Should this be peak in May and holiday for the rest of the season. If not, what kind of races would you expect in their program's?

Need some inspiration, so please share your thoughts.

Tank
Edited by TankNL on 23-02-2013 18:42
 
Jesleyh
Maybe Paris-Nice/Tirreno and Suisse/Dauphiné - NC's?
Tour Down Under?
Or if it's the puncheur type of stage-racer(like Valverde) the Canadian classics?
Edited by Jesleyh on 23-02-2013 18:51
i1326.photobucket.com/albums/u660/jesleyh/Junk/0ca5fb14-ed59-44b1-8eb0-596097ba5c01_zps8e97f370.jpg

Feyenoord(football) and Kelderman fanboy


PCMdaily Awards: 12x nomination, 9x runner-up, 0x win.
 
lluuiiggii
I don't see a reason to make a schedule where there's only a peak in May and holiday for the rest of the season. First, keep in mind that not only actual Giro contenders use the schedules; a lot of riders from Pro Conti, Continental teams and even FAs also use it. Second, the fact that the specialization says 'no world' doesn't mean he can't have late season races where he might have a shot at a good result, with the best example being the Giro di Lombardia, which is race that all riders with leader type 15, with no exception, will ride according to STA_program_info Wink

Fitness schedules should maximize the rider's performances, so as long as it doesn't affect his main goal (a peak before the Giro will be really bad for the Giro itself, but a peak after the Giro won't change a thing), I don't see why riders shouldn't be in good form.

Lastly, I don't think that having a specialization "giro_solo_world" assures that the rider will ride the Worlds, he's got to be selected after all. But I'm not sure if the AI uses the specializations to choose riders for the national teams (f.e., a rider with "noworld" will for sure reject going for the Worlds, or a rider with "world" will have priority for the Worlds even if there are better riders with other programs that don't specify focusing or not on the Worlds).
 
Blueprint
In my opinion, GIRO_SOLO_NOWORLD can have several options:

1. Tirreno/Sanremo - Giro - build up to Lombardia
2. Tirreno/Sanremo - Giro - Eneco / Hamburg /Plouay

Both can be enriched with Canadian races too.

Or for smaller Italian teams:

3. Tirreno/Sanremo - Giro - Lombarda / Padania etc...
4. Tirreno/Sanremo - Giro - August Italian classics
5. Tirreno/Sanremo - Giro - Lombardia, Emilia etc...

I'm sure you can fill in the details.

TankNL wrote:
I'm still working on the new program's and new fitness program's, but I need your feedback.

For instance. Lets take the leader_type 15 (cya_giro). This has 6 specializations;
11 GIRO_SOLO_NOWORLD
12 GIRO_SOLO_WORLD
13 GIRO_ARDENNAISES
14 GIRO_TOUR_NOWORLD
15 GIRO_TOUR_WORLD
16 GIRO_VUELTA

There is a possibility to link fitness programs to the specializations. But when I look at the giro fitness programs now, then most of them also have a fitness_peak in September or October. I will need to link them to the specializations with the worlds in them, but what about specializations like 11 (GIRO_SOLO_NOWORLD). Should this be peak in May and holiday for the rest of the season. If not, what kind of races would you expect in their program's?

Need some inspiration, so please share your thoughts.

Tank
 
TankNL
Good to see people joining the discussion. Let me share what i've learned so far.

STA_type_leader;
IDtype_leader CONSTANT
14 cya_polyvalent
15 cya_giro
16 cya_tdf
17 cya_vuelta
21 cya_clas_val
22 cya_clas_pav
23 cya_sprinters
25 cya_eq_tours
26 cya_eq_clas
28 pro_sprinters
29 pro_francais
30 pro_italiens
31 pro_espagnols
32 pro_benelux
33 pro_eq

We have;
- STA_type_leader with a relation to STA_type_leader_specialization with relations to STA_fitness_program with a relation to STA_fitness_program_week

The first relation between leader and leader_specialization is not something I found in the files, but I see the combination exclusively.

assumption 1; The game gives the cyclist according to their stats and relative strength to the competition, a type_leader value
assumption 2; The game gives the cyclist according to their stats a type_leader_specialization value.

After this the game has in STA_program and STA_program_info a fixed calendar for a number of type_leaders (14,15,16,17,21,22,23). The calendar consists of possible races to ride in a month and the choice of those race combinations can be influenced by the leader_specialization.

assumption 3; The game selects every month the program for the rider and does this completely random. The rider will ride the races that his team is participating in, regardless of his fitness level.

There are no programs for the other type_leaders and there for, race participation is completely random.

I will do some analyzing of games started with the new PCMDaily2013, in order to see how many riders get what leader roll and specialization. From the first glance, there are only about 8 riders who are a GIRO type_leader and approximately 20 who have the TOUR type leader. I will try to gather some stats on that.

This will determine my approach with the fitness program's and the program_info. because if it remains an exclusive group, than I will make exclusive fitness and race programs.

Tank
Edited by TankNL on 24-02-2013 11:18
 
lluuiiggii
TankNL wrote:
assumption 1; The game gives the cyclist according to their stats and relative strength to the competition, a type_leader value
assumption 2; The game gives the cyclist according to their stats a type_leader_specialization value.

I'd strongly believe that nationality also plays a role on choosing the type_leader value. And a bit of random too.

As for the type_leader specialization, if I had to guess I'd say it's just random. Otherwise the same riders should have the same specialization in all careers, which doesn't happen.

And btw, if you'll be testing these with the 2013 DB you might want to "reset" it before testing. The non-save .cdb file already defines the type_leader of some riders, according to real life (e.g. Purito for TdF, Wiggins for Giro, etc), that would mess up any conclusions that can be taken from testing Pfft
 
TankNL
I've created 10 career files, all went 1 day ahead and used FullEditor_V400 to get the dyn_cyclist files. I've analysed them all and the first finding. The leader_types for GIRO, TOUR and VUELTA (15,16,17) are only used by 47 cyclists. Only Talansky, and Van Garderen switched between GIRO and VUELTA and Danielson switched between GIRO and TOUR, The rest were constant in their leader_type.

The specialization was very random, not static at all. I will post the results.
 
TankNL
Leader type = 15 (GIRO)
Number of times that the specialization was seen for this cyclist.

Name111213141615
A.Talansky341
B.Intxausti14122
B.Wiggins2512
D.Cataldo2152
D.Cunego118
D.Pozzovivo231211
I.Basso631
J.Brajkovič133111
J.J.Cobo73
M.Scarponi12511
R.Gesink52111
R.Hesjedal3223
R.Kiserlovski121222
R.Majka82
R.Urán1621
S.Sánchez73
T.Danielson221
T.Machado1522
T.Van Garderen34
V.Nibali82


11 GIRO_SOLO_NOWORLD
12 GIRO_SOLO_WORLD
13 GIRO_ARDENNAISES
14 GIRO_TOUR_NOWORLD
15 GIRO_TOUR_WORLD
16 GIRO_VUELTA

from this table, Cunego is 8 out of 10 times a specialization 13 (GIRO_ARDENNAISES), so that looks good. But Nibali also is 8 out of 10 times specialization 13 and that seems stranger.
 
Jesleyh
Gesink's one (He aims for Giro + Tour just 2/10) seems a bit strange too Frown
Edited by Jesleyh on 24-02-2013 18:39
i1326.photobucket.com/albums/u660/jesleyh/Junk/0ca5fb14-ed59-44b1-8eb0-596097ba5c01_zps8e97f370.jpg

Feyenoord(football) and Kelderman fanboy


PCMdaily Awards: 12x nomination, 9x runner-up, 0x win.
 
TankNL
Leader type = 16 (TOUR)
Number of times that the specialization was seen for this cyclist.

Name21222324252627
A.Contador2143
A.Schleck12331
A.Valverde3241
B.Mollema316
C.Evans217
C.Froome442
D.Menchov11251
E.Gasparotto1171
F.Schleck1612
J.Coppel82
J.Fuglsang46
J.Peraud12511
J.Rodríguez3214
J.Van den Broeck217
J.Vanendert212131
M.Nieve316
N.Roche361
P.Fédrigo322111
P.Rolland181
R.Taaramäe136
T.Danielson23
T.De Gendt112231
T.Pinot433
T.Voeckler3151

21 TOUR_SOLO_NOWORLD
22 TOUR_SOLO_WORLD
23 TOUR_SOLO_ARDENNAISES_NOWORLD
24 TOUR_SOLO_ARDENNAISES_WORLD
25 TOUR_GIRO
26 TOUR_VUELTA
27 TOUR_VUELTA_ARDENNAISES
 
TankNL
Leader type = 17 (VUELTA)
Number of times that the specialization was seen for this cyclist.

Name3233
A.Talansky11
D.Moreno91
H.Zubeldia91
I.Antón91
S.Henao64
T.Van Garderen3


31 VUELTA_SOLO
32 VUELTA_GIRO
33 VUELTA_TOUR

Only 6 cyclists with this leader_type. And only the specialization with another grand tour and no VUELTA_SOLO (31).
 
TankNL
lluuiiggii wrote:
And btw, if you'll be testing these with the 2013 DB you might want to "reset" it before testing. The non-save .cdb file already defines the type_leader of some riders, according to real life (e.g. Purito for TdF, Wiggins for Giro, etc), that would mess up any conclusions that can be taken from testing Pfft


That pre-defined works perfectly. The 10 games above are with your already defined leader_types. Can you also pre-define the specialization?

I will also conduct the same analysis without the predefined leader types.

Tank
 
TankNL
Ok, now without the predefined leader_types.

Leader type = 15 (GIRO)

Name111213141516
A.Talansky617
D.Cataldo1313210
D.Pozzovivo11132210
I.Basso22
J.Brajkovič11
J.Van den Broeck112
M.Scarponi3121310
P.Rolland112
R.Gesink112
R.Hesjedal11114
R.Porte1214
R.Taaramäe112
S.Henao1124
T.Danielson213
T.De Gendt123
T.Pinot1517
T.Van Garderen1427
Eindtotaal7241712101080

Leader type = 16 (TOUR)

Name21222324252627
A.Contador2132210
A.Klöden43310
A.Schleck1151210
A.Valverde2121410
C.Froome215210
C.Horner231410
F.Schleck27110
I.Basso178
J.Brajkovič14218
J.Van den Broeck12317
N.Roche26210
P.Rolland2147
R.Gesink121217
R.Hesjedal11316
R.Porte1326
R.Taaramäe2417
S.Henao55
S.Sánchez13610
T.Danielson2136
T.De Gendt1236
V.Nibali1910
Eindtotaal2632216571714173

Leader type = 17 (VUELTA)

Name313233
A.Talansky213
D.Moreno9110
H.Zubeldia8210
J.Brajkovič11
J.J.Cobo15410
J.Van den Broeck11
P.Rolland11
R.Gesink11
R.Taaramäe11
S.Henao11
T.Danielson11
T.De Gendt11
T.Pinot33
T.Van Garderen33

Edited by TankNL on 24-02-2013 20:12
 
TankNL
Leader type 14 (cya_polyvalent) shows more changes

first with predefined leader_types;

Name13456
A.Klöden12710
C.Horner36110
F.Kessiakoff134210
L.L.Sánchez5510
M.Rogers9110
R.Costa531110
R.Porte17210
T.Martin23510
Eindtotaal1092982480

and without predefined leader_types;

Name12345678
B.Mollema1332110
B.Wiggins44210
C.Evans8210
D.Menchov7310
F.Kessiakoff18110
J.Fuglsang131121110
L.L.Sánchez6410
M.Rogers25310
R.Costa521210
R.Kreuziger261110
R.Urán333110
T.Martin25310
V.Kiryienka141410
Eindtotaal1422945112234130


For me the biggest supprise is Wiggens. I don't get it why the game thinks that he should go for the smaller tours. Same goes for Mollema, Evans and Menchov... might be that there are already strong cyclists in the same team?

1 POLY_PARISNICE_ARDENNAISES
2 POLY_TIRRENO_ARDENNAISES
3 POLY_ARDENNAISES_TDS
4 POLY_ARDENNAISES_DAUPHINE
5 POLY_PARISNICE_DAUPHINE
6 POLY_PARISNICE_TDS
7 POLY_TIRRENO_DAUPHINE
8 POLY_TIRRENO_TDS

however in those STA_type_leader_specialisation records, the Tour is given as a yes, so the game might think that it's ok for cyclists to go that route. Will do some further testing.
Edited by TankNL on 24-02-2013 20:50
 
Ulrich Ulriksen
Playing with this while waiting for the PCM 13DB to download.

I noticed Wiggins' exclusion just before you posted

There are exactly 30 riders in every test, my guess is the game selects those riders using some criteria and Wiggins doesn't make the cut. I don't think it is by team because Sky has 4 riders on the list and radio shack has 5. Menchov and Evans are also odd exclusions.

I combined the 3 "undefined" tables from the prior post to see how it looked if you combine leader types 15, 16 and 17. Then I counted the number of occurrences of each of the Giro, Tour, Vuelta and Ardennaises. A value of 10 means every possibility includes that race. See table below.

Nationality clearly plays a role - all those with a 10 in the Giro are Italian and all those with a 10 in Vuelta are Spanish. Also, only 3 riders have a significant value on the Ardennaises, the Schlecks plus Sanchez. Doesn't make total sense - what about Valverde a Spanish rider with interest in the Ardennaises but his only high value is the tour. Could favorite races have an impact?

I think your rework of the fitness schedules is a great thing. I think the poor fitness on many races is the biggest issue with the AI - particularly the smaller races.


TGVA
Name
A.Talansky2731
D.Cataldo41023
D.Pozzovivo51021
I.Basso8903
J.Brajkovič9320
J.Van den Broeck8524
M.Scarponi31031
P.Rolland8611
R.Gesink9423
R.Hesjedal7722
R.Porte9420
R.Taaramäe9621
S.Henao6930
T.Danielson7611
T.De Gendt8712
T.Pinot3731
T.Van Garderen0732
A.Contador10342
A.Klöden10030
A.Schleck10028
A.Valverde10403
C.Froome10520
C.Horner10041
F.Schleck10018
N.Roche10020
S.Sánchez10069
V.Nibali10901
D.Moreno90100
H.Zubeldia80100
J.J.Cobo51100
 
TankNL
@Ulrich Ulriksen; Thank you for your comment.
Can you do the same exercise with leader_type 14 included?
 
Ulrich Ulriksen
Here you go.

Not a lot of logic here - Menchov loves the Dauphine. LLS loves Paris-Nice and likes the Ardennaises. I think he won PN a few years back so maybe that is it but wouldn't have thought of him as big gun for LBL.

Can't think what TDS stands for. Am I being slow?

Also where is Rodriguez? Would be interested to see what leader type he has.


TGVAPNTATDSCD
Name
A.Talansky27310000
D.Cataldo410230000
D.Pozzovivo510210000
I.Basso89030000
J.Brajkovič93200000
J.Van den Broeck85240000
M.Scarponi310310000
P.Rolland86110000
R.Gesink94230000
R.Hesjedal77220000
R.Porte94200000
R.Taaramäe96210000
S.Henao69300000
T.Danielson76110000
T.De Gendt87120000
T.Pinot37310000
T.Van Garderen07320000
A.Contador103420000
A.Klöden100300000
A.Schleck100280000
A.Valverde104030000
C.Froome105200000
C.Horner100410000
F.Schleck100180000
N.Roche100200000
S.Sánchez100690000
V.Nibali109010000
D.Moreno901000000
H.Zubeldia801000000
J.J.Cobo511000000
B.Mollema00072254
B.Wiggins00080264
C.Evans000100082
D.Menchov000730010
F.Kessiakoff00091028
J.Fuglsang00053363
L.L.Sánchez000610040
M.Rogers00073055
R.Costa00078041
R.Kreuziger00080237
R.Urán00094043
T.Martin00028037
V.Kiryienka00055055
 
TankNL
TDS = Tour du Suisse.
My guess is that he's a leader type 21. I'll check that tomorrow. I'm off to sleep now.
 
Jump to Forum:
Login
Username

Password



Not a member yet?
Click here to register.

Forgotten your password?
Request a new one here.
Latest content
Screenshots
Promo : Astana TT
Promo : Astana TT
PCM 08: PCM.daily's 2009 DB
Fantasy Betting
Current bets:
No bets available.
Best gamblers:
bullet fighti... 18,376 PCM$
bullet df_Trek 17,374 PCM$
bullet Marcovdw 15,345 PCM$
bullet jseadog1 13,552 PCM$
bullet baseba... 10,439 PCM$

bullet Main Fantasy Betting page
bullet Rankings: Top 100
ManGame Betting
Current bets:
No bets available.
Best gamblers:
bullet Ollfardh 21,890 PCM$
bullet df_Trek 15,520 PCM$
bullet Marcovdw 14,800 PCM$
bullet jseadog1 13,500 PCM$
bullet baseball... 7,332 PCM$

bullet Main MG Betting page
bullet Get weekly MG PCM$
bullet Rankings: Top 100
Render time: 0.32 seconds