This WK/BJ situation is quite intriguing...I don't recall ever seeing a rider reacting quite like Jungels is for you. He's definitely the outlier in the example, as you pointed out. His high capital makes me think he's still "trying" to grow, but something is holding him back. It's almost as if Kelderman has followed what we'd expect to be a normal career trajectory, albeit a tad bit quick in regards to his progression ending. For Jungels however, it would be consistent with a rider who has reached all their max values, but is still below that 2000 training style threshold. However, with it being Jungels, I find that unlikely.

It makes me wonder if we're not actually a tad further from the "truth" then we'd like to be. Seeing as your so far in to that specific career Gentlemen, are there any other cases like this? If it's just a one-off perhaps it's just a glitch in the matrix -- but if this consistently occurs, then it's obviously something else we have to try and account for. With so much correlation between what we're finding as far as career development and the new BAP mode and DB sections, perhaps some of our secrets lie in that specific game mode. It does seem like Cyanide have overhauled everything to mesh up with how your rider in BAP develops, even if it is career mode.

@teamdoubledragon No, actually Jungels seems more normal than most in his profession. Julian Alaphilippe, for instance, is currently having a value_f_capital of a whopping 9607 and although he has hit most of his core max limits, there`s still room for i.e. his TT to develop (he`s on 72, could go to 75).

While Alaphilippe is the rider with the highest value_f_capital, Simon Yates is just as scary; he has 8717, but his value_i_training is still listed as 125, which means he`s still having the training style ground-breaking and gaining stats the fastest because of that (if the theory holds).

Of the 'old boys' Peter Sagan and Nairo Quintana are in the 9000-range as well, though they do not continue developing, seeing how their trainingstyle is in the 2000-range.

Coming back to Jungels/Kelderman, I`ll post their stats again, this time with their max. limit, so the comparison would hold up a bit better.



As you can see, their stats AND their limits don`t seem to differ all that much from each other.

Note that I said that Kelderman switched to Traditional early on, but that wasn`t the case. He`s still on Modern actually, but his value_f_capital still doesn`t progress as rapidly as Jungels.
Between 26/4/2018 and 15/7/2018 they progressed like this:

This despite having a pretty similar value_f_gain. So if their progression should be around 20 a week (based on your calculation), their progression should be around 200 points (value_f_gain * number of weeks), right? Based on those calculations, Jungels actually follows up on that. Kelderman, on the other hand, stays far behind, almost a factor 10. I think that`s odd. If I check how many riders have a higher value_f_capital than 7000, I`m getting 32 riders, OVR ranging from 76 tot 81. According to your breakdown, this should be a list of riders with an OVR > 80, which it isn`t.

So, can someone check if my DB is just messed up, or can they actually confirm what I found? Right now I`m in season 2019.

Sure, I have edited a few riders with the Fast Editor, but I doubt that would screw up things like this. Especially when I haven`t touched these guys.

Interesting point in the last post on page 1. Can anyone confirm that a rider will grow more on climbing when his other important stats are already on max and you train Climbing? Or will he incease less on overall because he can only improve on 1 stat? I was thinking the last thing.

@Gentleman When I get the chance either later tonight or over this weekend I plan on doing a pseudo-career that sims through 3-5 years, and checking how progression works in the long term. Thank goodness for weekly saves!! Makes this is but less tiresome. Perhaps this will also isolate some more unknowns.

@koningco I can't help but feel that while the process of allocating attributes is random, there is still a cap to how much a single stat will jump at once, regardless of training style. To explain, say you have Rider A: all of his stats are at their max potential except MO. Then you have Rider B: all of his stats are under their max potential cap, giving him the ability to gain in any of them. Both are training with the same coach, and have the climber skill as their style. I postulate that Rider B will see anywhere between a 0-3 point jump in MO plus an assortment of other 1-2 point jumps in his back up stats, much like Ulrich pointed out.

Now, for Rider A since he has all of his stats expect MO maxed, there is no alternative place to allocate points...BUT his MO stat still isn't going to jump more than that 0-3 threshold. After all, it wouldn't make much sense if Rider A had say 72 MO, and was going to get 12 points added to his stats, for ALL of them to go toward climbing just because everything else is full. 72 --> 84 in one jump just isn't logical, real or fake world.

So the Jungels - Kelderman example is a very interesting. I actually think it speaks to an aspect of the missing multiplier. I think it is function of comparing two riders with similar high maxes but pretty different starting points. If you have a save from the first week of your career it would be interesting to see what their starting capital values were.

I suspect they look something this (from the original game DB):

	OV. V	Capital
Kelderman	77.26667023	4843.052734
Jungels	74.06666565	3347.071533

So Kelderman starts 3 points higher but he is also therefore much closer to his max. So the game wants to slow his growth and speed up that of Jungels. But with young riders starting at these high values the quick progress level up points are already gone. So to improve Jungels a lot he has to increase his capital F value very fast because by the time you are in the 3000's there is only a level up value every 500 points+ and the game doesn't want him peaking at 32.

Now one mechanism the in the game is to lower or increase his gain value. It does tend to increase or decrease the gain value of riders it wants to advance or decline quickly or slowly (e.g. young potential 7 guys). But those values are still in a pretty narrow range and so it only impacts so much. So what I think this example shows is that it compensates by generating very high or low multipliers for riders who are too close or too far from their maximum. TeamDoubleDragons career test would prove this

So I would predict that now Jungels is reaching close to his potential he will cease to get the big multiplier.

I would also be interested to see if new iriders start at 0 capital. I assume they would which should solve this problem by giving them access to the rapid level up available in the lower values of the capital value (every 100 points) but at game start the game doesn't do that.

I actually think this points to a flaw in the process - when the game starts the capital value is based purely on a riders rating, not his rating and age. So a 30 YO with a value of 74 and a max of 75 starts at the same place in capital as a 23 YO with a value of 74 and a max of 81. Then the game has to compensate by giving the very high multipliers to the 23 YO rider.

@Ulrich That's a really great way of explaining it! I'll be honest, the correlation of what you said was always there in the back of my mind, but I never really connected it as factor. This makes me think, for this example career mode -- which I plan to start in a couple hours -- perhaps, along with overarching pro-/regression, I should pick a few different rider cases out. Like Jungels, a young rider with high potential who is quite far from his maxes. Then Quintana, a younger rider who is near most of his maxes. And perhaps a rider or two who are 4-5 potential as well. This should show us how these factors effect different age and potential groups as well. Looking forward to see how this all pans out over the next couple of days.

Exactly. But since you know Rider A has maxed out all of his other stats, you can switch to a Climber workplan (0-3 MO) instead of whatever he was training at the moment (say, Puncher, which has a 0-1 for MO). It increases the chance that you will get more points in MO with your next upgrade.

I often try this with my Punchers if I don`t know what their limits are. I have a very good Puncher with a HIL stat of 81, so I figure he won`t progress very much in his HIL stat anymore and he`s only 24. So I can either gamble on him getting to 82, or I can try to switch his training to see if he can improve on other stats, like MO, which gives him an edge on the longer climbs.

I admit, it is a gamble, but if you Puncher has an 81 for HIL, it is a safe bet to try for backup stats


I believe that they wanted to regulate that through the Training Style (Traditional, Modern, Ground-Breaking), considering our tests with that. But that is dependant on either age or a threshold in - probably a combination of value_f_capital and potential.

One thing I only now realize I`ve omitted (thanks, Ulrich!)is that I added the 'fix' which would stop the fast progress of riders in the BaP and make those other riders progress more reasonable.
I didn`t think of it because the superfast progression in BaP doesn`t seem to replicate in career mode (although career mode has different problems), but as it is a black-box solution for me, I now start to wonder what kind of impact it could have. It could just be a value_f_gain downgrade or something, but it could also do something else entirely. It another variable most of us probably have though.



As for Jungels, no such luck. He went up to 8515 by now. I`m currently in the Giro, he progressed from 7984 early December to 8515 mid May. I also took some time to check what Quintana and Sagan did in the early years ... in 2016 they had progressed almost a 1000 - from 6xxx to 7xxx. Now, in 2019 they`re over 9000 (pun not intented, they actually have that value).

But what is interesting is that Jungels sort of progresses the same as Kelderman, despite having almost 3000 more points. So at least his progress doesn`t skyrocket, although Kelderman hardly progresses at all nowadays.

Kelderman`s last stat update was before January 2017 (I only update my stats in January, so it happened somewhere during 2016), while we`re in 2019 now. Is that odd? Not really. With the stats he currently has, he`s a pretty good GT rider; not overpowering in the mountains, but he makes up for that in his TT - and he has a great recovery. I managed to win the Vuelta and the Giro with him, so I guess that proves my point

But while Kelderman doesn`t need to improve, he still does - in terms of value_f_gain at least. He`s not becoming a beast, but he certainly has improved over the years to gain these stats and becoming a GT contender if he focuses on it (well, in-game that is ... when he can beat Froome on a mountain that probably means Froome hasn`t had the GT as a primary focus).

The only worry I have is that Jungels will transform into a beast. But his progress actually hasn`t been too much different from Kelderman, despite his skyrocketing value_f_capital, his stats don`t really skyrocket. Sure, he progresses well, but he`s a 7 potential and he`s fairly young. Comparing them would come to (sue me for not being able to outline them):

MO HIL TT PRL END ACC RES REC

Kelderman +2 +1 +3 +1 +1 +0 +1 +3

Jungels +3 +2 +3 +1 +1 +1 +0 +4

With Jungels being 1,5 year younger, this isn`t an unrealistic growth. Kelderman got these stat updates over 4 seasons. Jungels needed 3 seasons for them, but considering he started in 2015 with a 68 REC, he really had to come from far.

His value_f_capital he had from my first save was 3688, so that seems to be pretty similar to Ulrich`s. Kelderman started Omloop end February with 4271 btw, but these values aren`t pre-filled on the PCMdaily DB, so I suppose they will be generated upon creating the limit values, rather than the actual values. Unless Ulrich didn`t use the PCMDaily EP v16.1 2014/2015NC, which would explain the difference.

Kelderman is 28 now and had his last stat update when he was 26, Jungels will be 27 in two months from now. I suppose his progress should be getting slower, but I`ll make up that balance by January 1st 2020 to see the difference a year since last one.


Try picking riders with a similar type. Jungels and Quintana are both GT riders, although Quintana is more of a Climber whereas Jungels is a Stage Racer. Picking along those lines will give some comparison in progression since they`ll have some sort of the similar targets and progression.

Agree - it will slow down soon, the age based reductions in the multiplier will begin to kick in.

I was using the original game db not PCM16. But just looked at the 16 DB and the gain and capital values are all zero in the base CDB file. The game sets them which is what I would have guessed. I might check if you did populate them would the game over write it.

I just started 2 custom careers, just to compare the stats and surprisingly the value_f_capital for Kelderman stays the same: 4247.99 - even the digits behind the dot are the same.
The max. limits for stats however, aren`t.

The first career gives him a max. of 83 for MO, but with a 74 REC. The second one gives him a max. of 81 for MO, but a 77 REC.
So there are differences based on every new career, although his value_f_capital stays the same. My assumption that this value is being calculated based on his max. limits is hereby busted; it seems more realistic that Ulrich`s theory - that it`s based on actual, current values - is the correct one. Even the hypothesis that it is based on max. - current doesn`t seem likely due to his value_f_capital being the static same on a new career.
I changed a stat and checked if the value_f_capital does change upon the start of a career and indeed: 4316.00 is the value after upgrading this REC from 74 to 75.

That probably means that the max. stats are based on a combination of current stats + potential, where value_f_capital might not have any influence.

I checked and confirmed the same for Jungels. One thing I noticed different though is the fact that his training style differs. In the first one he`s having a 4 - meaning Ground-breaking; in the other one he`s having 1108, meaning Modern. Considering our past conclusions, this should have an impact on how fast his progress should be, which should differ in both careers. Surprisingly, this happened with the same current stats and the same value_f_capital. So if progression is linked to training style, it seems to be (randomly?) predetermined regardless of the value_f_capital.

Sorry for the radio silence over the weekend, I had extended time on the job, so I'm a bit behind. But I am still simming through a career trying to get a few years in to try and get a more in depth look at detailed progression. It just may take a few days, I have a lot of real life stuff convening all at once.

In regards to the last couple of posts, yes, as proven, upon starting a new career there are a number of factors that are both random and static. value_f_capital seems to be derived from the riders overall stats in some way, shape or form; and thus is always the same at the start, pending the riders stats are the same. However, things like max stats, and training methods, along with rider contracts and the like, are randomized. Once again, I believe this is purposely done to make sure each career has its own unique feel. No two riders will progress the same way each and every time.

The more I think about it the less I think training style influences progression directly, but more gives you, the player, an idea of where riders are in the grand scheme of things. For example, while the majority of ground-breaking riders do have fast progressions, I specifically recall some who did have negative gain tallies. And the same goes for riders with modern training styles, although they were less likely to have large positive gain multipliers. Somewhere there is still something that causes, or directly assigns, those multipliers. The question really is, in my opinion, do we have access to them? Or has Cyanide created a hard coded variable which we currently can't access. Not to try and discredit all the other work we've been doing in unlocking the secrets around capital, gain, trainingstyle; I just feel the ultimate solution is beyond those yet still.

I think the problem is the multiplier is a formula that has multiple inputs - we have identified 3 factors at least - age, relationship of the rider's current stats to their potential and possibly training style. I too am skeptical that ground breaking is always going to create faster progression, seems like that would be making one training style better than another and that doesn't seem like the point.

Technically the difference between Traditional, Modern and Ground-breaking is how 'new' the training style is. If you remember that back in the days training merely involved hours and hours of just riding the bike, that seems a bit dated nowadays. Then someone introduced dieting, to take care of what you eat to make sure you have a particular weight and such, next to training programs for your muscles. Only much later they introduced training programs for testing bikes in a special environment (forgot the actual term for it, sorry) to measure the aerodynamica and form on the bike.

If you look at things this way, it is easy to understand that those are three entirely different ways of training. So, generally, the newer one is the one which lets you judge your form, how to best focus on growth and molding you into a particular sort of rider.
But I can imagine that after a while, when you grow older and more established, you have already become a particular rider. You don`t need all that extra stuff and you can just do some simple training on the bike and watch your diet. So all these molding techniques are not necessary anymore as you won`t make these growspurts anymore. A 27 year old rider usually doesn`t need to discover himself, unless he suddenly realizes he can climb (like Dumoulin). Just like not everyone fits a regid ground-breaking scientific training style Sky offers; some would actually get worse if they cannot deal with that kind of style.

While that wasn`t the point you were trying to make, I can easily understand why the developers would want to seperate those styles and thereby the progress growth. So I believe that somewhere, somehow, it influences the growth of a rider. If that happens by a hidden variable which triggers the training style, or that the training style triggers a hidden variable, that is currently unknown.

As for the hidden variable ... I don`t know if it is hardcoded or hidden somewhere in a DB or XML file. The fact that Cyanide likes to use abbreviations of French cycling terms doesn`t help spotting them either.

I also encountered a crash during a race which I had to figure out how to get past, simulating did the trick, but the logging pointed at missing commentary files which gave errors - probably resulting in a crash. Why I have never encountered those errors, I have no idea. It is not like Cyanide is terribly organized, so we might even overlook something obvious because it is hidden somewhere we wouldn`t even begin to look. An XML file with growth/decline ages/values which is hidden between the commentary audio for instance, obscured by a French filename I wouldn`t dare imagine it being related. Who knows? But I do like poking and discovering things and at least I`ve found a lot of things which give me more insight in how the game plays.

So at least for me it`s not a waste of time

I`ve just hit the point where new riders are generated (25 October 2019 iirc) and I looked at the DB to see what their value_f_capital is. Taking into account that there are a few youngsters which were already around, I just took the top 30 youngest riders and looked at their values. Somewhere around the 100.xx, although there are also a lot around 80.xx and a few which are higher - 126, 326, 394, 185.

So no - they do not start at 0, but definitely in the range where they can develop quickly.

Looking at the same riders around 1-1-2020, I don`t see the 100-values anymore. I`m now looking at riders of which most have around 240-250 values, with some having jumped to 450-560. One had an exceptional boost to 800, but he also has a 6 potential whereas the others varied between 3-4. The value_f_gain would explain some of that.

As for Jungels, he switched to Traditional training and the only progression he has made at the age of 28 is +1 FTR and +1 REC. But it looks like Jungels has gotten to the point where he now seems to top off, stopping his fast progression. Note that it looks like his progression slowed before he switched training styles, so it looks like the training style is an effect, not a cause.

Kelderman has made progression too, btw He had a +1 MON, +1 HIL, +1COB, +1 SP, +1 DH, +1 REC which makes it seem like he was ready to spend all his points accumulated over the past few years. Still, it`s a fairly significant jump considering his age and the fact that he gets a +1 in at least 2 core stats as a Stage Racer, although HIL and DH don`t really hurt either.

Coming back to the training style, I signed a young, 21 year old sprinter who has made substantial progression in core stats already (SP 78, ACC 75). Upon realizing he`s 21, while not having a known training style, I therefore assigned him to a Ground-breaking trainer. And a few weeks later he asked me to switch him to a different trainer. So I looked - and he has a Traditional training style. Apparently this is one of these cases where a young rider prefers that other style.
Does that mean he has progressed so fast that he now needs to top off? At 21 he certainly does have impressive stats for a sprinter, so it kind of makes sense. There are only 8 sprinters with an 80 SP and 5 with an even higher SP, so he`s already knocking on the door at the age of 21.

Thanks for the update this seems to confirm most of what we expected. Young riders don't start at zero but really pretty close as you note.