|
calculated stats based on FINAL results
|
| TankNL |
Posted on 24-09-2014 21:04
|
Domestique
Posts: 440
Joined: 19-03-2007
PCM$: 200.00
|
matt17br wrote:
Well, with totally messed up I mean that Froome can't have the top recuperation stat. It is wrong. Simply wrong. In the last week of tour 2013 he was always beaten by Quintana and the same at this year's Vuelta. He has his top form in the 2nd week. Quintana, conversely, is 2 points lower than Froome, while he is clearly the best in the 3rd week.
Another thing that I don't like is too much riders on the 80's stats. But I think that this could easily be fixed since it is a problem of the matrix.
Of course the same of Froome is applied to Valverde and Rodriguez.
I will look into it, but will be hard, but i'll try. |
| |
|
|
| matt17br |
Posted on 24-09-2014 21:11
|
Directeur Sportif
Posts: 10525
Joined: 28-09-2013
PCM$: 200.00
|
TankNL wrote:
matt17br wrote:
Well, with totally messed up I mean that Froome can't have the top recuperation stat. It is wrong. Simply wrong. In the last week of tour 2013 he was always beaten by Quintana and the same at this year's Vuelta. He has his top form in the 2nd week. Quintana, conversely, is 2 points lower than Froome, while he is clearly the best in the 3rd week.
Another thing that I don't like is too much riders on the 80's stats. But I think that this could easily be fixed since it is a problem of the matrix.
Of course the same of Froome is applied to Valverde and Rodriguez.
I will look into it, but will be hard, but i'll try.
It is already a great matrix, but that will make it better. Ofc is nearly impossible basing everything on the results. BTW, what you did for the domestiques?
|
| |
|
|
| Ad Bot |
Posted on 29-09-2024 16:27
|
Bot Agent
Posts: Countless
Joined: 23.11.09
|
|
| IP: None |
|
|
| TankNL |
Posted on 24-09-2014 21:15
|
Domestique
Posts: 440
Joined: 19-03-2007
PCM$: 200.00
|
For the domestiques I have given them a percentage of the leader points in the stage or race. Percentage and minimal result in order to get points varies per race category. |
| |
|
|
| sammyt93 |
Posted on 24-09-2014 22:59
|
Classics Specialist
Posts: 3625
Joined: 03-07-2012
PCM$: 300.00
|
TankNL wrote:
matt17br wrote:
Well, with totally messed up I mean that Froome can't have the top recuperation stat. It is wrong. Simply wrong. In the last week of tour 2013 he was always beaten by Quintana and the same at this year's Vuelta. He has his top form in the 2nd week. Quintana, conversely, is 2 points lower than Froome, while he is clearly the best in the 3rd week.
Another thing that I don't like is too much riders on the 80's stats. But I think that this could easily be fixed since it is a problem of the matrix.
Of course the same of Froome is applied to Valverde and Rodriguez.
I will look into it, but will be hard, but i'll try.
I have absolutely no idea how to word this so bare with me here.
Is it possible for the percentage of how far through the race the stage is to have an effect, with say a bonus 20% (capped at 100) for anything in the third week and a bonus 10% for anything in the second week?
e.g. you have a general rec stat calculation for a stage (no idea what that formula would be) then you get a percentage of that for how far into the race the stage is, maybe with a minimum % and a minimum race length in which you can earn Rec points.
Please tell me at least some of that makes sense as the more I read it or try to reword it the less understandable it becomes.
|
| |
|
|
| matt17br |
Posted on 25-09-2014 12:46
|
Directeur Sportif
Posts: 10525
Joined: 28-09-2013
PCM$: 200.00
|
sammyt93 wrote:
TankNL wrote:
matt17br wrote:
Well, with totally messed up I mean that Froome can't have the top recuperation stat. It is wrong. Simply wrong. In the last week of tour 2013 he was always beaten by Quintana and the same at this year's Vuelta. He has his top form in the 2nd week. Quintana, conversely, is 2 points lower than Froome, while he is clearly the best in the 3rd week.
Another thing that I don't like is too much riders on the 80's stats. But I think that this could easily be fixed since it is a problem of the matrix.
Of course the same of Froome is applied to Valverde and Rodriguez.
I will look into it, but will be hard, but i'll try.
I have absolutely no idea how to word this so bare with me here.
Is it possible for the percentage of how far through the race the stage is to have an effect, with say a bonus 20% (capped at 100) for anything in the third week and a bonus 10% for anything in the second week?
e.g. you have a general rec stat calculation for a stage (no idea what that formula would be) then you get a percentage of that for how far into the race the stage is, maybe with a minimum % and a minimum race length in which you can earn Rec points.
Please tell me at least some of that makes sense as the more I read it or try to reword it the less understandable it becomes.
This is a good idea, if this is possible will make the matrix way more realistic.
An excellent middle way between what I say and the actual matrix.
|
| |
|
|
| TankNL |
Posted on 25-09-2014 19:03
|
Domestique
Posts: 440
Joined: 19-03-2007
PCM$: 200.00
|
@sammyt93 and matt17r;
Thanks for the feedback and inspiration.
Now the recuperation is based on the GC points, based on the assumption that the higher recuperation riders are able to win GC's.
What I also could do is build in a multiplier based on the length of the tour;
stage_date - start_date_tour / 21 * 100% * points for the race.
I could even cap it that only from 8 days or more the recuperation points are awarded. How does that sound. And should that be on top of the points that are collected on GC's, or should that be ignored altogether?
Tank
Edited by TankNL on 25-09-2014 19:03
|
| |
|
|
| matt17br |
Posted on 25-09-2014 19:12
|
Directeur Sportif
Posts: 10525
Joined: 28-09-2013
PCM$: 200.00
|
TankNL wrote:
@sammyt93 and matt17r;
Thanks for the feedback and inspiration.
Now the recuperation is based on the GC points, based on the assumption that the higher recuperation riders are able to win GC's.
What I also could do is build in a multiplier based on the length of the tour;
stage_date - start_date_tour / 21 * 100% * points for the race.
I could even cap it that only from 8 days or more the recuperation points are awarded. How does that sound. And should that be on top of the points that are collected on GC's, or should that be ignored altogether?
Tank
This sounds great! Glad that our suggestions are going to improve the rec stat matrix now!
|
| |
|
|
| TankNL |
Posted on 27-09-2014 19:04
|
Domestique
Posts: 440
Joined: 19-03-2007
PCM$: 200.00
|
I implemented the suggestion today, but now the recuperation of Quintana has gotten worse and Valverde and Froome still have 82 and 81 recuperation.
The reason for this is that they still have a lot of good results (at least top 10) in the last week of grand tours... more than somebody like Quintana.
so putting that in place, the points scored in the 2nd and 3rd week is in my opinion less accurate, because now also somebody like Majka gets a high recuperation, and with his stats could be a GC contender... which might be accurate for the future, but not right now. |
| |
|
|
| matt17br |
Posted on 27-09-2014 19:07
|
Directeur Sportif
Posts: 10525
Joined: 28-09-2013
PCM$: 200.00
|
TankNL wrote:
I implemented the suggestion today, but now the recuperation of Quintana has gotten worse and Valverde and Froome still have 82 and 81 recuperation.
The reason for this is that they still have a lot of good results (at least top 10) in the last week of grand tours... more than somebody like Quintana.
so putting that in place, the points scored in the 2nd and 3rd week is in my opinion less accurate, because now also somebody like Majka gets a high recuperation, and with his stats could be a GC contender... which might be accurate for the future, but not right now.
Crap. But how this is possible? Quintana never had a bad day in the 3rd week 
|
| |
|
|
| TankNL |
Posted on 28-03-2015 18:23
|
Domestique
Posts: 440
Joined: 19-03-2007
PCM$: 200.00
|
Hi all, I've kept adding races to the excel and now have a the stats up until now.
will post the links in this thread later tonight. |
| |
|
|
| TankNL |
Posted on 28-03-2015 19:37
|
Domestique
Posts: 440
Joined: 19-03-2007
PCM$: 200.00
|
Edited the first post.
update 28-03-2015
I've updated the big file with all the races up until now. In total 2135 results are taken into account to calculate the stats of all the riders.
https://dl.dropbo...%20v1.xlsx
There is a sheet called; calStats with the raw calculated stats. I've modified a couple before I loaded them into the database (sheet dbStats) and the alterations are shown in yellow.
I've edited the 1.75 database of Jesleyh, just to see how it would run. I think the results are very accurate. Will do some more testing.
For those who are interested;
https://dl.dropbo...cStats.cdb |
| |
|
|
| TankNL |
Posted on 29-03-2015 11:10
|
Domestique
Posts: 440
Joined: 19-03-2007
PCM$: 200.00
|
Here are a couple of screens of stats in game;
Mountains;
Hills;
Cobbles;
Sprint;
 |
| |
|
|
| Fresh D |
Posted on 29-03-2015 11:30
|
Free Agent
Posts: 104
Joined: 08-12-2007
PCM$: 200.00
|
Nice idea from you and good work
I noticed all sprinters are rate 70 in climbing and everybody timetrials and sprints good so 70~.
Is it possible to diversify the results of your calculation a bit more for each stat, so that the range will go from maybe 60-85?
But nice tool nonetheless.
Good for editting continental team i guess. |
| |
|
|
| matt17br |
Posted on 29-03-2015 11:33
|
Directeur Sportif
Posts: 10525
Joined: 28-09-2013
PCM$: 200.00
|
Splendid work as always. Your matrix gets nearer and nearer to perfect with every release. Huge kudos Tank.
Just a question: why it happens that quite a lot of sprinters (and not only) have 70 mountain? Is that to make them get the mountains passed more easily in the gts? It's the only thing that doesn't look good to me. Kittel wouldn't deserve a 55 either 
EDIT: Ha zabel'd 
Edited by matt17br on 29-03-2015 11:35
|
| |
|
|
| TankNL |
Posted on 29-03-2015 12:16
|
Domestique
Posts: 440
Joined: 19-03-2007
PCM$: 200.00
|
Hi, thanks for the compliment matt17br. And yes as you and Fresh D pointed out, there seem to be a number of riders with stats around 70. I will look into that. I have a algorithm, that gives points to the top between 76 and the max of the stat and it also had a sub range of 69 to 76. I might do some tests with the algorithm and lower that second range to 65 to 76. But I will have to see what the effect is.
if you look in the excel, there are around 4500 riders with stats, and there are a lot of riders with stats lower than 69... but they might not be in a database But I will come back to that later tonight. |
| |
|
|
| TankNL |
Posted on 29-03-2015 19:06
|
Domestique
Posts: 440
Joined: 19-03-2007
PCM$: 200.00
|
Hi,
did some alterations on the algorithm, but it is hard to keep the balance. Especially since it applies to all the stats. So what works for mountains, messes up the cobbles stat. I will have to do some more testing.
so far the sprinters are now;
Spoiler | CyclistName | FLA | MOU | HIL | ITT | PRO | COB | SPR | ACC | DWN | FIG | STA | RES | REC | pop | | GREIPEL André | 79 | 67 | 70 | 62 | 71 | 69 | 83 | 82 | 73 | 73 | 78 | 72 | 73 | 82 | | CAVENDISH Mark | 79 | 69 | 70 | 69 | 74 | 70 | 83 | 81 | 76 | 73 | 77 | 72 | 74 | 82 | | KITTEL Marcel | 78 | 66 | 69 | 69 | 77 | 58 | 83 | 82 | 71 | 73 | 77 | 69 | 72 | 82 | | KRISTOFF Alexander | 81 | 69 | 74 | 68 | 71 | 78 | 82 | 81 | 78 | 73 | 82 | 77 | 75 | 85 | | DEGENKOLB John | 81 | 69 | 75 | 69 | 76 | 79 | 82 | 80 | 79 | 73 | 81 | 78 | 74 | 85 | | BOUHANNI Nacer | 79 | 64 | 71 | 61 | 70 | 64 | 81 | 82 | 74 | 74 | 77 | 73 | 71 | 82 | | SAGAN Peter | 81 | 70 | 79 | 72 | 79 | 80 | 80 | 81 | 82 | 73 | 82 | 80 | 77 | 87 | | VIVIANI Elia | 77 | 69 | 70 | 64 | 71 | 69 | 80 | 79 | 73 | 71 | 76 | 72 | 71 | 80 | | DEMARE Arnaud | 78 | 69 | 70 | 64 | 70 | 76 | 80 | 79 | 76 | 71 | 77 | 74 | 73 | 81 | | NIZZOLO Giacomo | 77 | 69 | 70 | 69 | 77 | 61 | 79 | 79 | 72 | 72 | 77 | 72 | 71 | 80 | | DUMOULIN Samuel | 77 | 69 | 74 | 61 | 69 | 59 | 78 | 77 | 73 | 72 | 77 | 76 | 71 | 80 | | FARRAR Tyler | 77 | 69 | 69 | 65 | 71 | 74 | 78 | 78 | 73 | 71 | 76 | 72 | 71 | 79 | | RENSHAW Mark | 76 | 64 | 67 | 62 | 69 | 69 | 78 | 77 | 70 | 75 | 72 | 70 | 70 | 70 | | ROJAS GIL Jose Joaquin | 77 | 71 | 74 | 65 | 70 | 69 | 78 | 77 | 75 | 77 | 74 | 76 | 72 | 80 | | FERRARI Roberto | 76 | 69 | 69 | 64 | 69 | 51 | 78 | 77 | 71 | 72 | 74 | 70 | 71 | 79 | | MODOLO Sacha | 76 | 63 | 69 | 61 | 59 | 63 | 78 | 77 | 70 | 71 | 74 | 70 | 69 | 73 | | VAN STAEYEN Michael | 76 | 58 | 69 | 56 | 50 | 61 | 78 | 77 | 69 | 75 | 73 | 69 | 63 | 66 | | MATTHEWS Michael | 78 | 69 | 77 | 70 | 75 | 52 | 78 | 78 | 74 | 72 | 77 | 77 | 71 | 81 | | MEZGEC Luka | 76 | 66 | 70 | 63 | 69 | 61 | 78 | 77 | 71 | 71 | 73 | 72 | 70 | 79 | | VAN ASBROECK Tom | 77 | 61 | 71 | 51 | 50 | 72 | 78 | 77 | 75 | 75 | 76 | 74 | 69 | 79 | | COQUARD Bryan | 77 | 65 | 70 | 61 | 69 | 66 | 78 | 78 | 71 | 71 | 74 | 71 | 71 | 79 | | VALVERDE BELMONTE Alejandro | 79 | 81 | 83 | 77 | 76 | 69 | 77 | 78 | 80 | 72 | 82 | 80 | 82 | 87 | | PETACCHI Alessandro | 73 | 64 | 63 | 61 | 59 | 61 | 77 | 77 | 66 | 74 | 73 | 66 | 69 | 48 | | CHICCHI Francesco | 72 | 59 | 61 | 50 | 51 | 61 | 77 | 76 | 60 | 72 | 70 | 64 | 59 | 39 | | VAN HUMMEL Kenny | 73 | 61 | 67 | 52 | 52 | 61 | 77 | 77 | 66 | 71 | 70 | 69 | 69 | 51 |
Edited by TankNL on 29-03-2015 19:07
|
| |
|
|
| matt17br |
Posted on 29-03-2015 19:22
|
Directeur Sportif
Posts: 10525
Joined: 28-09-2013
PCM$: 200.00
|
It definitely looks much better now. Some adjustaments to do here and there but having a completely perfect matrix is not the point of the whole project 
Keep it up!
|
| |
|
|
| TankNL |
Posted on 29-03-2015 23:01
|
Domestique
Posts: 440
Joined: 19-03-2007
PCM$: 200.00
|
matt17br wrote:
It definitely looks much better now. Some adjustaments to do here and there but having a completely perfect matrix is not the point of the whole project
Keep it up!
Thanks, you are correct. The point was not to set it up to score the obvious cyclist, but rather the less known riders. I think this is looking better and better.
Tank |
| |
|
|
| nightguy |
Posted on 31-03-2015 19:47
|
Free Agent
Posts: 100
Joined: 14-09-2009
PCM$: 200.00
|
Have to say this is great idea and I really love it!
I have taken a brief look at the excel. Just a quick suggestion that perhaps you can consider.
Would it be useful to have different ratios based on the rider type? For instance, a climber would have a bigger ratio than a classics rider for instance for getting a good result on a mountain stage. That would help to lower the mountain points of sprinters for example. |
| |
|
|
| TankNL |
Posted on 02-04-2015 19:42
|
Domestique
Posts: 440
Joined: 19-03-2007
PCM$: 200.00
|
nightguy wrote:
Have to say this is great idea and I really love it!
I have taken a brief look at the excel. Just a quick suggestion that perhaps you can consider.
Would it be useful to have different ratios based on the rider type? For instance, a climber would have a bigger ratio than a classics rider for instance for getting a good result on a mountain stage. That would help to lower the mountain points of sprinters for example.
I think It will cause a circular reference. Because the rider profile would be computed based on results and your suggestion would make the results depending on the rider profile.
As you have seen in the Excel, almost every thing is computed automatically. The only thing that needs a refresh is the pivot table... everything else are formula's that compute on demand. |
| |
|
