• Announcements

    • IMPORTANT - REACH US IN THE NEW FORUM   05/04/2017

      Ladies and gentlemen ATTENTION please:
      It's time to move into a new house!
        As previously announced, from now on IT WON'T BE POSSIBLE TO CREATE THREADS OR REPLY in the old forums. From now on the old forums will be readable only. If you need to move/copy/migrate any post/material from here, feel free to contact the staff in the new home. We’ll be waiting for you in the NEW Forums!

      https://community.blackdesertonline.com/index.php

      *New features and amazing tools are waiting for you, even more is yet to come in the future.. just like world exploration in BDO leads to new possibilities.
      So don't be afraid about changes, click the link above and follow us!
      Enjoy and see you on the other side!  
    • WICHTIG: Das Forum ist umgezogen!   05/04/2017

      Damen und Herren, wir bitten um Eure Aufmerksamkeit, es ist an der Zeit umzuziehen!
        Wie wir bereits angekündigt hatten, ist es ab sofort nicht mehr möglich, neue Diskussionen in diesem Forum zu starten. Um Euch Zeit zu geben, laufende Diskussionen abzuschließen, könnt Ihr noch für zwei Wochen in offenen Diskussionen antworten. Danach geht dieses Forum hier in den Ruhestand und das NEUE FORUM übernimmt vollständig.
      Das Forum hier bleibt allerdings erhalten und lesbar.   Neue und verbesserte Funktionen warten auf Euch im neuen Forum und wir arbeiten bereits an weiteren Erweiterungen.
      Wir sehen uns auf der anderen Seite!

      https://community.blackdesertonline.com/index.php Update:
      Wie angekündigt könen ab sofort in diesem Forum auch keine neuen Beiträge mehr veröffentlicht werden.
    • IMPORTANT: Le nouveau forum   05/04/2017

      Aventurières, aventuriers, votre attention s'il vous plaît, il est grand temps de déménager!
      Comme nous vous l'avons déjà annoncé précédemment, il n'est désormais plus possible de créer de nouveau sujet ni de répondre aux anciens sur ce bon vieux forum.
      Venez visiter le nouveau forum!
      https://community.blackdesertonline.com
      De nouvelles fonctionnalités ainsi que de nouveaux outils vous attendent dès à présent et d'autres arriveront prochainement! N'ayez pas peur du changement et rejoignez-nous! Amusez-vous bien et a bientôt dans notre nouveau chez nous

How Horse Skill Gain Likely Works

16 posts in this topic

Posted

NOTE: This post is an extension of conversations taking place in the Breeder’s Discord, but because of its length could not be easily posted there.

 

What does it mean when my horse levels up and I get the message “chances of learning a skill has increased”? Why do horses learn skills early on but not many (if any) in their 20’s? Do costumes, branding stones, and the appearance change coupons work and how?

Skills: The Hidden XP Bar

When your horse is running, it gains xp towards its next level-up based on distance. This is an xp bar you can see and use to predict how long you will need to ride your horse before it levels.

Every time your horse levels, more xp is needed in order to level up again. By the time the horse level 28, you might have to autoloop an entire night on a high tier horse in order to hit level 30 as opposed to spending one night to get from level 1 to level 15.

Skills seem to work in a similar fashion.

Your horse is gaining skill xp every time it levels up, but this is an xp bar that we cannot see.

Every time your horse learns a skill, the next skill requires more xp in order to learn, just like how the horse levels. But because your horse is likely learning a set amount of skill xp on each level up, eventually the gap widens and it takes several level ups to learn a skill.

This is why early on your horse will learn many skills, then slow down in its teens, and then maybe only learn a couple skills in its 20s.

Why are some horses smart and some horses.. not?

The RNG comes into play when the horse is born. (Horses born pre-nerf will learn skills at a rate consistent with old skill gain rates. Early data is indicating that horses coming out of racing seal boxes may not be effected by the nerf either.)

Skill gain RNG is effecting one of two things, either of which would result in the patterns we see.

a) The set amount of Skill XP the horse gets on level up is determined at birth. A low roll means the horse will not be able to meet as many Skill XP thresholds, resulting in a lower number of skills learned.

b) The Skill XP thresholds that must be met in order to learn a new skill may be higher on certain horses, resulting in a lower number of skills learned.

In either case, costumes, branding stones, and change coupons will still have the same effect. The nerf probably made the amount of skill xp the horse learns on level up lower than before.

How do Costumes and other items that effect horse skills work?

Branding, costumes and appearance coupons likely work by boosting the total amount of skill xp your horse has.

Appearance coupons are unique in that even if your horse is level 30, if it was close enough to a Skill Threshold, and the coupon pushes your horse over that threshold, the horse is awarded with a skill.

Branding stones don’t seem to make much difference in skill gain, so the skill xp award is probably low.

Horse costumes seemed to be bugged for much of the early game, but certain ones are finally causing skill gains on horses. Confirmed with testing are the Lahr Arcien and the Karlstein. (Also, unequipping and re-equipping the costume on the same horse may be negating the bonus. It could be a bug.)

There is a lot more data related to this section, such as ideal times to equip a costume or use a coupon, but I feel more testing is needed before publishing.

14 people like this

Share this post


Link to post
Share on other sites

Posted

Excellent work, as always. Now if only we had a horse section to post this in xD

1 person likes this

Share this post


Link to post
Share on other sites

Posted (edited)

Indeed, a horse forum section would be wonderful for compiling all the horse info, discussions, and guides. Nice work, Capp!

 

Edited by Lightingale
1 person likes this

Share this post


Link to post
Share on other sites

Posted

That's really interesting. Thank you!

Share this post


Link to post
Share on other sites

Posted

I've been thinking about this a lot and I am skeptical that this is how things work.  Could you explain how you came to the conclusion that there was this hidden skill pool?

1 person likes this

Share this post


Link to post
Share on other sites

Posted

I've been thinking about this a lot and I am skeptical that this is how things work.  Could you explain how you came to the conclusion that there was this hidden skill pool?

To be fair, there is no way to be certain. However, I don't think it is just an RNG percentage chance for a horse to learn a skill and here is why:

1) If it was just chance, we wouldn't see skills clustered so much in the first 15 levels of the horse. The skills would be very random across all levels. But 2/3 of the skills are often learned at 50% of its level. The clustering of skills is a little easier to see on higher tiers where the total potential of the horse is high. I would regularly see t7/8 with 7 or 8 skills by level 10, then they would learn a couple more by 15, and then occasionally learn skills after that.

I looked at spreadsheet data from other breeders and on lower tier horses there were examples of skills being spread out a bit more on lower potential horses. I want to go back and level some again and track their skill gain to see if they truly break the pattern. So far, given that there IS an element of RNG in skill xp gain and/or the thresholds, I haven't seen anything that conclusively breaks the model yet.

2) People have argued that the reason the horse learns more skills early on is because certain skills are "common" and easily gotten. The thing is my horses have never just learned the common skills and then moved on to the rarer ones. My purebred t8s learned sprint (or IA) and sidestep by level 16. Both learned charge and drift before they were even level 10. One of them never learned Caution. 

I have a completely different theory about how specific skills are chosen, but I hesitate to share as I feel there is less evidence for this aspect. It isn't something the breeder can influence anyway, so it doesn't ultimately matter.

3) The fact that pearl horse items work the way they do is actually the biggest indicator of how skill gain works imo. If there were no skill xp thresholds, equipping a pearl horse set would not normalize skill gain on a horse after it hit a sticking point. If it was just percentage increase to roll, the skill gain would be all over the place especially at higher levels. But even with the pearl items, we still see that massive drop in skill gain in the horse's 20s.

4) If it was based on a percentage chance that increased every time your horse leveled and didn't manage to learn a skill, I would expect the pattern of learning skills to be reversed. Low at first and then increasing, but we see the opposite happen.

1 person likes this

Share this post


Link to post
Share on other sites

Posted

Your first point seems logical. But the fact that lower tiers are more spread out skill wise could also have to do with them being able to learn less (and less expensive) skills in certain cases.

The second seems to be true. I have a T8 (and recorded her progress) and the first 12 levels she learned 9 cheap skills. Then from 13-20 nothing, then drift and S:front. Then from 23 till 30 nothing. I wanted to test on her if it had something to do with the first day being her Lucky day but that was proven wrong. I tried leveling her a week later on the same day --> 0 skills learned so there is no pattern there. Total: 11 cheap skills , 1 mediocer skill --> drift. Therefor your point seems to be valid (but I only checked her progess I never tracked others).

Perhaps that's where the potato horses come into play. They are destined to become potato's the moment they were born.

Share this post


Link to post
Share on other sites

Posted

Good insights. Thanks! I'm currently training some T7s, aiming for some T8s! Hoping they learn some good skills. One of them has Drift, Sprint, and Two-Seater; the other has Drift, Sprint, and Instant Accel. I heard someone say once that if you breed smart horses together, they pass on a bit of their smarts to their foals. That seemed to only be true once I got to T6-T7s. So we'll see! I feel like it is highly dependent on the RNG?

Share this post


Link to post
Share on other sites

Posted

yes, i have a totally different theory, it was posted on the big thread about 600 pages ago so i will have to find it

it explained why you get less skills levels 20+

ok found it:

 

Ok here is my theory on horse skills.  I will just throw numbers out.  This theory comes from leveling over 300 horses (~150 tier 6s, 18 tier 7s, 2 tier 8) 

There are 3 RNG die rolls with a few modifiers:

1) at the tick of a new level there is a 10% to learn a new skill multiplied by the tier of the horse (n x 10% where n=tier)

2) if roll #1 is successful, RNG decides which skill to roll from current unlearned list

3) the skill that is chosen has its own success % roll based on its premium (just throwing numbers out there Caution and Start Accel could be 80%, Hind Kick and Fore Kick could be 75%, etc ... all the way down to Sideways Move and Two Seater which could be 15%                                  

--> horse brand adds 3% to rolls 1&3

--> full set pearl horse gear set adds 3% to rolls 1&3

this would explain BOTH why people don't think that the horse costume and brand work because they add so little to the RNG chance AND it would explain why folks have had success switching out caution/start accel for a premium skill once they ONLY had premium skills left, then relearning caution and rinse repeat.  As I explained a few dozen pages ago, only use the coupon switch theory when you finally train a horse that learns all the lesser skills and only has premiums left           

Share this post


Link to post
Share on other sites

Posted

Interesting for sure, but I can't make the percentage model fit what is actually happening when horses level for all of the reasons listed in an earlier post. If it was just percentage to roll a skill, we wouldn't see them grouped as we do. Plus, here is something reeeeeaaallly odd I have noticed several times now, and it has to do with how horses learn which skills and in my mind proves skill xp gain without a doubt.

Occasionally you will level 2 horses one right after the other (although some people have noticed it on multiple horses on the same wagon) that will learn the same number and same types of skills at the exact same levels until they enter their teens. Then they are still learning equivalent skills but sometimes a level gap will present (which makes perfect sense if there is an RNG in skill xp gain and/thresholds).

The RNG between the horses seems to be in which equivalent skill gets rolled. This happened most recently on 2 Tier 8s. Where one rolled Sprint, the other rolled IA. Where one rolled Quick Stop, the other rolled Quick Back, etc. Literally every skill they rolled was equivalent to what the other had rolled and at nearly the same level. (I have seen this happen on t6s as well.)

Then my black horse, who had presumably worse RNG on skill gain xp, fell one level behind my white t8 on skill gain. I disrupted this by equipping the horse costume. He caught up to the white horse, and then learned the next skill a level before the white one did. Same skills, but by increasing his skill xp with the costume, I kept him from lagging behind the white t8.

To me, these experiments conducted over different Tiers support that a) Skill XP and Skill Thresholds exist, b) a horse's potential is determined at birth, when it is randomly assigned a preset skill hierarchy and c) there are indeed equivalent skills and the RNG is only involved in choosing which equivalent skill is rolled when that threshold is reached.

1 person likes this

Share this post


Link to post
Share on other sites

Posted

This is a very interesting discussion and people have posed some very interesting theories. The one thing I'm wondering is in regards to the idea of thresholds (skill exp) to learning skills. If there are thresholds that needed to be reached to learn skills, and those thresholds gradually increase in amount as the horse levels (such that fewer skills are learned at the higher levels) then how do you explain the deviants. The horses that only learn 4 or so skills in the first 10-15 levels and then suddenly have a string of learned skills when they're older so that they end up with a normal number of skills for their tier. They're rarer than the horses that gradually learn fewer skills as they level but still exist. Also, how do you explain when a horse learns a string of skills one level after another in the higher levels? If you were to plot skill gain on a chart you wouldn't get a nice seemingly exponential curve like you do with experience gain. The points would be all over the place. I'm just curious how the idea of exp gain could conform to fit those findings.

Share this post


Link to post
Share on other sites

Posted (edited)

I think your theory on a hidden xp bar that builds up over time is simply faulty speculation. I get how you came to that conclusion. But sometimes a more simple explanation is more likely.

Based on the anecdotal evidence found interwebz wide... I think that first your horse has to roll for a chance to successfully learn a skill. If that succeeds, then another roll determines which skills can be possibly be learned. But that second roll/check can negate learning a skill altogether if the roll doesn't land within a range that has skills available.

For example (using arbitrary numbers and skills as an example for the sake of simplicity)...

  • Suppose horses have a base 40% chance of simply having a "chance" to learn a new skill.
    • First roll is 40% or more... pass. Horse has a chance to learn a new skill.
      • Next roll chooses the range of skills from which can be chosen.
        • Scenario 1: Roll is a 50
          • Only unlearned skills are Caution (assigned 30% or more), Hind Kick (assigned 10% or more), and Sprint (assigned 60% or more).
            • In this scenario the horse will indeed learn a new skill, but only Caution or Hind Kick can or will be chosen and learned, since Sprint is not an option from the pool of choices because it's chance percentage is higher than the roll's range.
        • Scenario 2: Roll is a 50
          • Only unlearned skills are Sprint (assigned 60% or more), Instant Excel (assigned 55% or more).
            • In this scenario, while the horse passed the first check at a chance to learn a new skill, the second roll negated that chance since none of the skills still available fell within the second roll's range.

Of course, that's all speculation on my part. But it's speculation based on the same informatin as the OP has. It also explains why most (not all, but most, afterall we gotta take RNG and the law of averages into consideration) horses typically learn skills 40%-50% of the time (aside from those oddball potatoes) up until their lvl 20's, when suddenly learning new skills stalls dramatically after the easier to learn skills are out of the way and all you have left are the harder to learn skills.

 

Edited by DenverRalphy
1 person likes this

Share this post


Link to post
Share on other sites

Posted

I've lvled 7 t8s. Only 1 got sprint, drift, and ia. Only 2 got 15 skills. Both were missing ia, sprint, drift, sideways, and 2 seater. My white t8 only got 11 skills. I'm very unlucky with t8 skills 😔 I took one that had 15 skills and spent alot to switch all of those. Even used a premium appearance change coupon on him and he did not get xp or another skill. But he went from being the red t8 to looking like a black t7 😄 Since he got two seater on the 1st swap I'm not planning on making him a t9. I'll just work on another for that

Share this post


Link to post
Share on other sites

Posted

This is a very interesting discussion and people have posed some very interesting theories. The one thing I'm wondering is in regards to the idea of thresholds (skill exp) to learning skills. If there are thresholds that needed to be reached to learn skills, and those thresholds gradually increase in amount as the horse levels (such that fewer skills are learned at the higher levels) then how do you explain the deviants. The horses that only learn 4 or so skills in the first 10-15 levels and then suddenly have a string of learned skills when they're older so that they end up with a normal number of skills for their tier. They're rarer than the horses that gradually learn fewer skills as they level but still exist. Also, how do you explain when a horse learns a string of skills one level after another in the higher levels? If you were to plot skill gain on a chart you wouldn't get a nice seemingly exponential curve like you do with experience gain. The points would be all over the place. I'm just curious how the idea of exp gain could conform to fit those findings.

I am loving the collaborative discussion and your work with breeding/color theory as well!

I have probably leveled several hundred horses at this point (as have you), one at a time for the most part. The only time I have seen high tier horses learn 2 skills one right after the other after passing the mid-teens is while wearing a pearl costume. I have never experienced this otherwise. Do I just have really weird RNG? How often have you experienced this? Genuinely asking here- not being facetious.

 If this does happen, and there are thresholds at work, it means that multiple skills rest at each threshold and are rolled until expired.

 

Of course, that's all speculation on my part. But it's speculation based on the same informatin as the OP has. It also explains why most (not all, but most, afterall we gotta take RNG and the law of averages into consideration) horses typically learn skills 40%-50% of the time (aside from those oddball potatoes) up until their lvl 20's, when suddenly learning new skills stalls dramatically after the easier to learn skills are out of the way and all you have left are the harder to learn skills.

 

Horses learn rare skills at anytime, not just after the easy ones are out of the way. In fact, the better horses tend to have better skills early. If your t7 has learned 9 basic skills and nothing else heading into the late teens/20s... rip potato.

The other problem with "chance to learn" based on percentages has to do with identical horses. How do we explain horses that learn the exact same skills at the exact same levels like a carbon copy of each other? If it was all just chance, that shouldn't be possible.

Share this post


Link to post
Share on other sites

Posted

I am loving the collaborative discussion and your work with breeding/color theory as well!

I have probably leveled several hundred horses at this point (as have you), one at a time for the most part. The only time I have seen high tier horses learn 2 skills one right after the other after passing the mid-teens is while wearing a pearl costume. I have never experienced this otherwise. Do I just have really weird RNG? How often have you experienced this? Genuinely asking here- not being facetious.

 If this does happen, and there are thresholds at work, it means that multiple skills rest at each threshold and are rolled until expired.

I'm honestly not certain how often it is because unfortunately I only record the level that skills are learned on a very few horses (mostly ones I level by hand). Out of the 62 horses I have recorded skills for only 7 of them have most or all of the levels that those skills were learned recorded (the majority are just yes or no). Of those 7, 5 of them had 2 or more skills learned in a row in the second half of their levels. These included: My T8 Freyja who learned roar at 18 and S: Front Kick at 19. My T7 Gloom who learned streak jump at 17 and sideways move at 18. My T7 Minni who learned Start Accel. at 14, High Jump at 15, and Drift at 16. My T6 Ambiguity who learned IA at level 22, high jump at level 23, sprint at level 26, and quick ride at 27.

To be honest the idea of rolling the chance to learn a skill and then rolling on the skill would be a more simple way to explain the results that I have seen. Just because a skill like instant accel has a low chance of being learned would not mean it cant be learned early (for example, lets say theoretically you had a 25% chance of learning Drift at level 5. If you leveled 5 horses to level 5 the probability that at least one would learn drift is 1 minus the probability of not learning drift to the power of the number of attempts: 1-(0.75^4) = 0.68 = 68%. Yes you're less likely to learn drift than something else with a higher probability but that doesn't mean that you have to learn the higher probability skill before you can learn drift). The fact that you can learn rare skills early does not disprove the idea of a fixed % chance.

The newer idea of skill inheritance could also explain why your may see horses that learn very similar groups of skills. The idea is that having parents with trained skills may increase the likelihood of learning that skill in the foal. That can be applied by simply increasing the percent. So for example, a foal that had parents with trained drift may have a 35% chance of learning drift at level 5 instead of just 25% or some other increase. No to mention that a certain number of identical horses can be attributed to chance. If you have a lot of horses that are identical that aren't related then you may be onto something but you have to remember that we're working with a limited number of skills so there's only so much variation you can get in different horses (especially in the higher levels where most horses learn the majority of the skills).

RNG is also a major aspect of practically all parts of this game. It wouldn't be a stretch to assume horse skills are determined by RNG as well.

That being said the idea of skill thresholds can fit with the data as well, but its much harder to create a pattern through it, which means it would probably be harder to code - the general assumption is the simpler option is usually more likely to be correct (unless you have someone that wants to make it complex on purpose which is entirely possible). Ultimately until we can identify a distinct pattern in the two that would allow us to predict probability accurately for different situations we cant really test them to determine which is actually at play, and considering we cant make that distinction, knowing how the skills are learned I'm afraid may be something more for us to contemplate then actually take advantage of.

If you do come up with theoretical ways to predict the results for the different ideas let me know because I would love to play around with the numbers and see how they turn out (I'll give you a little spoiler that I'm already working on a project looking at skill inheritance so keep an eye out for that).

2 people like this

Share this post


Link to post
Share on other sites

Posted (edited)

I leveled my good share of horses as well and I tried to find an explanation for skill learning as well. This is, in brief, how I believe it works (just my personal opinion):

1- there is a hidden multiplier that affects the change to learn skills that depends on the horse tier and, maybe, on a random roll at the horse birth. The random roll may or may not exist, it would mix things up a bit but is not necessary to explain the skills patterns I've seen.

2- every time a horse level, a random number determines if the horse is going to learn a skill or not. The chance to learn a skill is based on the multiplier mentioned before, and the amount of skills the horse already has. In a similar fashion as gear enchanting, the more skills you have (or the higher enchanting level), the smaller is your chance to learn another skill (or go up another enchanting level). This explains why at low level horses usually learn more, cause they have less skills, and later tend to learn less. However I've had horses do the exact opposite, so I don't believe the "skill exp bar" exist. If that was true, the patterns will be much more recognizable.

3- what do costume/branding do? I think they work similarly to failstacks. I noticed that the message "the chance to learn a skill has increased" always appear upon leveling an horse when the costume is on, while is not always there without the costume. I believe what the costume does, is giving you a guaranteed failstack at every level, which increases your chance, by a small %, of getting a skill when a random number is generated on leveling (point 2). This additional chance boost is small, and that's why is not that noticeable. It would probably noticeable by comparing 100 horses leveled with and without the costume.

4- which skill does an horse learn? I believe that if you roll success at point 2, your horse will learn a skill no matter what, there is no second roll that can decide that after all you don't learn anything. Each skill has a set probability of being learned. This probability definitely depends on the skin of the horse. Every time a skill has been learned its probability is set to zero, while other skills' probabilities are recalculated to fill the gap. Rare skills, which are also more expensive, have less chance to be learned. So after rolling at point 2, if successful, another random number is generated to decide which skill is going to be learned.

5- skin chance coupon: they probably roll point 2 and then point 4. There is likely a hidden multiplier, that works similarly to costume/bard.

Edited by Dzessito

Share this post


Link to post
Share on other sites