There's still a ton of value in Evenepoel just reliably getting over 2000 points. That's such a high amount for a single pick that even at 2500 he'll still be pretty good and not a serious detriment for a team hoping to win it all. Just didn't think the upside was significant enough to gamble on. Think he'll just end up being a decent enough pick, but since I'm bound to also have a few busts I aimed for more high risk - high reward picks like Roglic.
@Armchair Cyclist - Thank you for sending me a list linking PCS to CQ names. It's not completely valid going forward as names are subject to change and ideally its based on identifiers but works well enough.
Just realized what you just implemented!
Such a good function! Thanks really for that!
No more srcolling through PCS startlists anymore. What should I do with the 10% of my working time that has been freed up? My boss will thank you for that!
@Armchair Cyclist - Thank you for sending me a list linking PCS to CQ names. It's not completely valid going forward as names are subject to change and ideally its based on identifiers but works well enough.
That’s a magnificent feature. Looking at it for now my riders are being very lazy and waiting until closer to spring to start their seasons unless they are just being shy about announcing their plans for the early season schedule.
I wonder if someone at PCS reads this thread. Shortly after me and @Kazistuta discussed that AWP interview from mid-December where his participation in the opening weekend was confirmed, he was added to the startlist for those on PCS.
Yeah, to overcome this time, we can start to discuss the pre-startlist of the race:
These should be the riders, who are picked in this game and will participate. If you wonder about the sorting, it´s popularity multiplied with 2025 points, which is a good indicator in my eyes, which riders are the most important participating riders for this game. Quite a huge number! Nearly half of the startlist had been picked!
@Shakes: To receive this overview, when you click on a race on your website, would be fantastic!
There's still a ton of value in Evenepoel just reliably getting over 2000 points. That's such a high amount for a single pick that even at 2500 he'll still be pretty good and not a serious detriment for a team hoping to win it all. Just didn't think the upside was significant enough to gamble on. Think he'll just end up being a decent enough pick, but since I'm bound to also have a few busts I aimed for more high risk - high reward picks like Roglic.
I'm not sure how true that is. At 2500 points he has a profit of 30%. That contributed to a team that scores overall 9720 if the whole team performed like that. And it's a significant proportion of the teams cost.
Looking at Salvaranis winning team last year and seeing how many riders returned that little and seeing how much they cost it lands on 1958 (by being generous and counting Plapp at 34% profit). That's basically the same cost as Evenepoel but spread over 7 "underperforming" riders. That means a team with Remco scoring 2500 could not afford to have any other flops below 30% profit if you want to have a similar failure rate as the last years winning team.
That seems like a rather significant handicap for the rest of the team to compensate.
Raising Remco's score to 3000 points instead means a profit of 55% and that raises the cost of riders in Salvaranis team scoring less than that to 2878 points. That gives the Remco team another 950 points worth of players who can also afford to score lower than 55% profit in order to have a similar failure rate. Now that's at least starting to look a little more reasonable.
Yeah, to overcome this time, we can start to discuss the pre-startlist of the race:
These should be the riders, who are picked in this game and will participate. If you wonder about the sorting, it´s popularity multiplied with 2025 points, which is a good indicator in my eyes, which riders are the most important participating riders for this game. Quite a huge number! Nearly half of the startlist had been picked!
@Shakes: To receive this overview, when you click on a race on your website, would be fantastic!
I'm not sure how true that is. At 2500 points he has a profit of 30%. That contributed to a team that scores overall 9720 if the whole team performed like that. And it's a significant proportion of the teams cost.
Looking at Salvaranis winning team last year and seeing how many riders returned that little and seeing how much they cost it lands on 1958 (by being generous and counting Plapp at 34% profit). That's basically the same cost as Evenepoel but spread over 7 "underperforming" riders. That means a team with Remco scoring 2500 could not afford to have any other flops below 30% profit if you want to have a similar failure rate as the last years winning team.
That seems like a rather significant handicap for the rest of the team to compensate.
Raising Remco's score to 3000 points instead means a profit of 55% and that raises the cost of riders in Salvaranis team scoring less than that to 2878 points. That gives the Remco team another 950 points worth of players who can also afford to score lower than 55% profit in order to have a similar failure rate. Now that's at least starting to look a little more reasonable.
I did some numbers on this some years back in the context of picking Valverde when he was fairly popular. One important thing to consider is that a lower budget is expected to give a larger % return. So a Remco team's other 32 riders will likely have a higher % return in total than the 33 riders on a non-Remco team.
Doing some extrapolation with the optimal team's returns to get percentages for an average 15000-scoring team, I calculated that Valverde barely had to make a profit to keep up with the non-Valverde teams if everything else went 'as expected'. Same goes for Remco if one is to use that calculation again.
The kicker is of course where the best picks are in a given year, who you are sacrificing to accommodate the expensive guy, and whether the extra cheap picks that you need are of the required quality.
If you wonder about the sorting, it´s popularity multiplied with 2025 points, which is a good indicator in my eyes, which riders are the most important participating riders for this game.
I have some recollection of a theory being espoused that the relevant metric maximised riders who are selected by half the entrants. A rider who is chosen once will only make a difference to one team: a rider on nearly every team has his relevance to the game effectively obliterated: he might change the size of the scores, but not much difference to the rankings. We shouldn't multiply by price such that free picks are irrelevant, no matter how often they were picked, so something needs to be added to get a non-zero rating: lets make that 100. (I tried it as 1, but that made cheap, very high frequency picks like Williams and Agostinacchio too far down the rankings: in the end, I thought if the minimum expectation of a rider is to return a 200 point profit, I'd post half of that)
So if we take (45-|frequency-45|) x (price+100), and then index that so that the top score is 100 and the free picks that are on only one team have 1, the importance metric is:
picks
price
importance index
WITHEN PHILIPSEN Albert
44
459
100.0
EVENEPOEL Remco
12
1929
99.0
NYS Thibau
46
423
93.6
O'CONNOR Ben
28
563
75.6
MAS NICOLAU Enric
27
560
72.6
TARLING Joshua
36
328
62.9
GAUDU David
37
315
62.7
VAN GILS Maxim
58
362
60.3
PITHIE Laurence
46
228
58.9
BRENNAN Matthew
11
1115
54.6
SEIXAS Paul
73
661
52.9
BLACKMORE Joseph
31
277
47.8
LAPORTE Christophe
59
273
47.3
MOHORIC Matej
28
290
44.7
TIBERI Antonio
16
555
43.0
DEL GROSSO Tibor
15
598
42.9
POGACAR Tadej
2
5080
42.5
PHILIPSEN Jasper
7
1269
39.3
COSNEFROY Benoit
64
268
39.3
POOLE Max
60
215
38.8
VAN EETVELT Lennert
68
302
36.3
MARTINEZ POVEDA Daniel Felipe
40
121
36.3
DECOMBLE Maxime
35
152
36.2
VLASOV Aleksandr
28
209
35.6
NORDHAGEN Jørgen
22
293
35.5
OMRZEL Jakob
19
345
34.8
LAMPERTI Luke
25
235
34.4
LANDA MEANA Mikel
15
433
32.9
TORRES ARIAS Pablo
34
130
32.2
PELLIZZARI Giulio
10
680
32.1
THIJSSEN Gerben
26
192
31.3
HAGENES Per Strand
36
106
30.6
BISIAUX Léo
21
246
30.0
WIDAR Jarno
69
245
29.9
SÖDERQVIST Jakob
17
323
29.7
RODRIGUEZ CANO Carlos
71
262
28.4
PERICAS CAPDEVILA Adria
25
162
27.1
HOBBS Noah
34
90
26.7
ZINGLE Axel
65
155
26.4
FINN Lorenzo
20
211
25.7
AYUSO PESQUERA Juan
4
1422
25.2
VERMAERKE Kevin
18
219
23.8
KRON Andreas Lorentz
38
50
23.6
REMIJN Senna
37
52
23.3
VALTER Attila
23
137
22.6
HIRSCHI Marc
6
800
22.4
ROGLIC Primoz
5
956
21.9
DE KLEIJN Arvid
22
134
21.4
UIJTDEBROEKS Cian
7
614
20.8
JAKOBSEN Fabio
44
11
20.3
GIRMAY HAILU Biniam
5
843
19.7
VANSEVENANT Mauri
22
113
19.5
SANCHEZ MAYO Pelayo
46
5
19.3
ZANA Filippo
13
237
18.3
VACEK Mathias
7
517
18.1
VAN WILDER Ilan
6
619
18.0
WILLIAMS Stephen
51
10
17.9
KÜNG Stefan
8
433
17.8
DINHAM Matthew
33
25
17.3
RICCITELLO Matthew
6
576
17.0
BEHRENS Niklas
33
16
16.1
DUNBAR Edward
14
167
15.7
MORGADO Antonio Tomas
5
635
15.4
GEE-WEST Derek
4
783
14.9
LIPOWITZ Florian
3
1065
14.7
EULALIO Afonso
9
242
13.0
BERCKMOES Jenno
6
408
12.9
AGOSTINACCHIO Mattia
30
0
12.7
NOVAK Pavel
15
97
12.5
GLOAG Thomas
9
222
12.3
AUGUST Andrew
11
162
12.2
RYAN Archie
9
215
12.1
GAUTHERAT Pierre
9
213
12.0
TEUTENBERG Tim Torn
7
298
11.9
KRIJNSEN Jelte
19
45
11.7
BRENNER Marco
9
191
11.2
POLITT Nils
11
137
11.1
KOOIJ Olav
2
1182
11.0
MILAN Jonathan
2
1179
10.9
PICKERING Finlay
12
112
10.9
MADOUAS Valentin
5
407
10.8
TUCKWELL Luke
17
46
10.6
VAN AERT Wout
2
1130
10.5
STEINHAUSER Georg
13
86
10.4
PEDERSEN Mads
1
2232
10.0
MARTINEZ Lenny
2
1066
10.0
404 riders in the game are of less than 10 importance rankings, and 222 of them are below 2 (remember, 1 is the minimum possible)
I'm not sure how true that is. At 2500 points he has a profit of 30%. That contributed to a team that scores overall 9720 if the whole team performed like that. And it's a significant proportion of the teams cost.
Looking at Salvaranis winning team last year and seeing how many riders returned that little and seeing how much they cost it lands on 1958 (by being generous and counting Plapp at 34% profit). That's basically the same cost as Evenepoel but spread over 7 "underperforming" riders. That means a team with Remco scoring 2500 could not afford to have any other flops below 30% profit if you want to have a similar failure rate as the last years winning team.
That seems like a rather significant handicap for the rest of the team to compensate.
Raising Remco's score to 3000 points instead means a profit of 55% and that raises the cost of riders in Salvaranis team scoring less than that to 2878 points. That gives the Remco team another 950 points worth of players who can also afford to score lower than 55% profit in order to have a similar failure rate. Now that's at least starting to look a little more reasonable.
I'm not sure you can compare things in that way. We are comparing Remco's value split over 7 riders vs. just Remco on his own.
That would neglect that Remco still leaves you with 6 spaces. In the Remco scenario you paid 1929 and 1 spot for 2500 points while in the other scenario you paid 1958 and 7 spots for the same thing. I wouldn't consider cheaper picks going +30% as good as a 1929-cost pick going +30%.
That being said I'm really only looking at this from the perspective of what you'd need a cyclist to score to end up with 16.000 - 17.000 points as that would typically make your team competitive.
I'm trying to calculate how much they'd need to score so that at the end of the year you don't look at them as being particularly bad or particularly good, but just acceptable, for the level required to try and win this game.
(If we played it with our nieces and nephews our picks wouldn't need to be so good).
e.g. Seixas didn't have to score so much in order to be considered a good pick.
I applied this to your team from last year and got a total of 16.590 points. This is what it would look like if every cyclist pulled their weight into getting you to win the game, but none of them carried anything extra:
Cyclist
Cost
Score
Profit
VINGEGAARD HANSEN Jonas
1760
2514
142,8%
VAN AERT Wout
1259
1823
144,8%
PLAPP Lucas
423
742
175,4%
SCHMID Mauro
420
732
174,3%
RICCITELLO Matthew
333
630
189,2%
LEKNESSUND Andreas
289
572
197,9%
LECERF William Junior
286
569
199,0%
TILLER Rasmus Fossum
269
551
204,8%
HAYTER Ethan
263
544
206,8%
VERNON Ethan
244
519
212,7%
JAKOBSEN Fabio
189
449
237,6%
UIJTDEBROEKS Cian
176
433
246,0%
NORDHAGEN Jørgen
165
423
256,4%
GEOGHEGAN HART Tao
137
392
286,1%
RONDEL Mathys
131
382
291,6%
KULSET Johannes
123
373
303,3%
TORRES ARIAS Pablo
122
372
304,9%
KÄMNA Lennard
108
354
327,8%
BAGIOLI Andrea
106
350
330,2%
LOUVEL Matîs
101
346
342,6%
KAJAMINI Florian Samuel
96
338
352,1%
CAVAGNA Rémi
80
318
397,5%
DONALDSON Robert
97
341
351,5%
KRON Andreas Lorentz
68
297
436,8%
VERMEERSCH Florian
51
278
545,1%
KAMP Alexander
49
272
555,1%
GLOAG Thomas
46
267
580,4%
SENECHAL Florian
37
257
694,6%
VAN BAARLE Dylan
35
254
725,7%
TULETT Ben
27
248
918,5%
GESBERT Elie
10
230
2300,0%
WITHEN PHILIPSEN Albert
0
210
SEIXAS Paul
0
210
So, what about Evenepoel? How much would we want him to score so that at the end of the year we should think he was a fine inclusion, yet nothing crazy.
It's actually 2756 (+42,9%) so it is indeed a bit higher than 2500.
If you manage to get a ~2750 Evenepoel then he's not going to be the reason you don't win.
Some extra reference points:
If a rider costs 100 you'd want ~350 points (x3.5)
If a rider costs 250 you'd want ~530 points (x2.1)
If a rider costs 500 you'd want ~830 points (x1.7)
If a rider costs 750 you'd want ~1150 points (x1.5)
If a rider costs 1000 you'd want ~1490 points (x1.5)
If a rider costs 1500 you'd want ~2150 points (x1.4)
If a rider costs 2000 you'd want ~2850 points (x1.4)
If a rider costs 2500 you'd want ~3600 points (x1.4)
If a rider costs 3000 you'd want ~4380 points (x1.5)
If a rider costs 5000 you'd want ~9910 points (x2)
If a rider costs 7500 you'd want ~18090 points (x2.4)
A 40-50% improvement seems to be the minimum. I'd indeed expect it to increase again at the top as too expensive picks begin having much more extreme downsides (e.g. a 7500-pick would mean you can only pick 0-pointers for the rest of the squad). It breaks at the top because of the method used.
Obviously since you'll have duds too, you'll hope that some of your picks far exceed these numbers.
Method: A mix of ideas, basically it roughly calculates ~80% of the score that a cyclist would at least need to make the optimal team as that more-or-less matches with what the winners score.
(Note: Their total score is not ~80% that of the actual optimal team, but rather it is ~80% of the optimal team where each rider scored exactly the amount of points he needed to make the optimal team. Unlike someone like Seixas or Del Toro who clearly scored a lot more. This is a more fair comparison as you really don't need picks to be otherworldly good in order to win).
Unless Brennan also scores 2700 and then AWP (or whoever else the Remco team didn't pick) chips in with 1000 as well.
It's just all down to estimated value as always. Expensive rider + cheap riders is just as valid a strategy as several somewhat expensive riders. Just depends who are the better picks. In practice, the expected value is almost never the same (although I'd say it's kinda close this year with the Remco strat vs. alternatives).
And there are different risks included with both. We've had the risk discussion a million times and there are differing views, but in my opinion the expensive rider is the least risky because it's easy to predict that Remco will score a lot of points. The risk comes from potentially missing out on some 700 guy going nuclear and/or your big hitter not delivering for some reason. The risk involved with the other strategy is how likely are you to predict all of your semi-expensive guys correctly. There is maybe three times the risk of someone getting injured or just not doing well compared to the risk of Remco not doing well.
In that hypothetical the person with Brennan and AWP instead would have gotten an advantage, but it would at least not be due to Evenepoel being bad or there being a situation like with Lazkano last year.
In the end I mostly do this to try and calculate what I'd need to get from a pick to make them worthwhile as it differs a lot based on their cost and I don't always find it intuitive.
Facebook
Twitter
Instagram