However when I open a PCM 2016 database (with Lachi's Editor) these columns cannot be found in STA_region. Only gene_f_chance_ "rider type" and gene_f_chance_ "temperature preferences" appear. So where to see/edit the distribution of potentials?
Secondly, in the youngridercarac xmls there are two set of intervals for all the individual ages, one in the top and one midway through the documents (potentialmax).
Are the ones in the top the ranges for starting/base attributes of a newgen and the potetialmax intervals further down the ones for potential attributes?
So for example in the ardennaises.xml (PCM Expansion Pack) an 18 year old with 8 potential, for the Flat attribute, he would have a 50 % chance of starting with 60 and a 50 % chance of starting with 61-62 [img][/img]
and for potential, a 30 % chance of reaching 71-73, a 60 % chance of 74-76, and a 10 % chance and of 77-79.[img][/img]
Is this understood correctly?
Lastly, can ranges for different values overlap? For example could a value 4 range be set as 74-77 and a value 5 as 76-79? In the PCM Expansion Pack xmls the maxpotential ranges never overlap, but I wonder if it is possible?
However when I open a PCM 2016 database (with Lachi's Editor) these columns cannot be found in STA_region. Only gene_f_chance_ "rider type" and gene_f_chance_ "temperature preferences" appear. So where to see/edit the distribution of potentials?
IIRC this is not possible in PCM 2016 anymore.
Secondly, in the youngridercarac xmls there are two set of intervals for all the individual ages, one in the top and one midway through the documents (potentialmax).
Are the ones in the top the ranges for starting/base attributes of a newgen and the potetialmax intervals further down the ones for potential attributes?
Indeed, the potentialmax sets ranges for the limits of the attributes, which is more appropriate name than potential attributes, although I think you might meant the same thing.
So for example in the ardennaises.xml (PCM Expansion Pack) an 18 year old with 8 potential, for the Flat attribute, he would have a 50 % chance of starting with 60 and a 50 % chance of starting with 61-62.
Yes
and for potential, a 30 % chance of reaching 71-73, a 60 % chance of 74-76, and a 10 % chance and of 77-79.
Not quite. It means 30 % chance of having limit 71-73, a 60 % chance of 74-76, and a 10 % chance and of 77-79. If you have a limit x, the value of the corresponding attribute cannot be higher than x, but it doesn't mean it will at some point of career be x. Depending on training etc., it could happen that the attribute never goes above x-10, say.
Lastly, can ranges for different values overlap? For example could a value 4 range be set as 74-77 and a value 5 as 76-79? In the PCM Expansion Pack xmls the maxpotential ranges never overlap, but I wonder if it is possible?
I believe they can, but it can create unrealistic probability distributions if used inappropriately, as I hope you can see.
So there is no (simple) way of finding out the distributions? I guess it's not really necessary to change the distributions as you could get the same results by tweaking the potential value ranges, but it would be nice to know what they are.
Not quite. It means 30 % chance of having limit...
Yeah that was what I meant.
I believe they can, but it can create unrealistic probability distributions if used inappropriately, as I hope you can see.
Yeah I understand, but couldn't you off-set this by making the ranges wider? Of course, if you let value 5 ranges overlap with value 6 ranges there will be significantly more very good riders, but if you also widen the low end of value 5 ranges (possibly also overlapping value 4 ranges) you would also produce more less good high potential riders. If properlybalanced the average distribution of stats across all newgens should remain the same, but it would be more varied and unpredictable.
Naturally, doing this requires knowledge of the percentage distribution of each potential level since without it there is no way to balance it out. Even with the knowledge I suppose it'd be a rather complicated endeavour.
Edited by Humlesnur on 10-03-2017 10:21
Based on different dbs it seems that you can assign a value from 0 to 5 for each attribute. Does higher values mean faster progression and vice versa, or is this completely off?
1. There's no way to find out and change the distributions, they're the same for all countries since PCM 15, from Belgium to Bermuda. It's unfair and it sucks for db makers, but there you go.
2. I suppose it's something that could be tested. The problem here as togo said is that then there might be cases where pot 4 riders end up with overall higher stats than pot 5 ones.
3. That column in sta_type_rider can be tweaked to assign different upgrades when a rider follows a certain training. I believe that the sum of those values needs to be between 25 and 30, I don't remember if the limit has to be a fixed number, a range, or it can even be unlimited if you decide to go crazy. The values need to be between 0 and 5. The order of the values follows the order of stats you see in DYN_cyclist, so flat, mountain, downhill, cobbles and so on.
2. If balanced properly this would be quite unlikely across all attributes, but even if it happened, would it really be a problem as long as the average distribution of attributes remains the same. The potential value effectively is nothing but a designated value (that makes scouting easier). If the ranges were less rigid potential value would be less significant as it be much less predictable, yet the number of good and bad riders would be around the same.
In theory, this should work well and hardly have any downside, if any (unless you count less predictability as a downside), but without knowing the expected number of riders of each potential level, it'd likely require too much tinkering with the distributions and ranges to be worthwhile.
3. So it doesn't have any effect on ai rider development, and is there any way in general to affect this? Regarding the values, I don't understand what they signify. What is the difference between 0 and 5? Slower/faster development when training?
The sum of the values in Anderis' db(from the link) greatly exceeds 30 for every rider type and each individual value is no less than 3. What is the effect of this?
I apologise for the pestering questions, but it is a little easier to ask here than to sim through several seasons comparing different setups. Edited by Humlesnur on 15-03-2017 13:59
3. So it doesn't have any effect on ai rider development, and is there any way in general to affect this? Regarding the values, I don't understand what they signify. What is the difference between 0 and 5? Slower/faster development when training?
5 = Fastest development, 0 = No development. When one of your riders is trained for cobbles, mountains will be further down in the priorities and cobbles at the top of the list of course, so one will be 0 and the other will be 5.
By changing the value_f_capital for riders you can affect their growth greatly. By increasing it, it becomes slower and vice versa. value_f_gain can also be used well for slowing growth, but for increasing growth it is less effective as it resets for every growth level.
In excel editor it's simple to do across multiple riders, but it naturally doesn't affect newly generated riders and thus would have to be repeated each time they are generated. Obviously for increasing growth, you'd wait a few years before lowering the value as it already starts low.
It's by no means an ideal method, but at least it is possible to control progression of all riders.
