Tree of Savior Forum

UPDATED Hair Accessory Enchantment Odds and Prices

I have seen 2x aoe dmg ratio in total so far on the market and 1x ms

Yeah, I think now after looking at everything there would have to be some restructuring of the equation to accommodate the stats in that fashion. The simpler generalization is the assumption that:

  1. One of three checks are made to determine if it is a single, double, or triple stat roll.
  2. For each number of stats going to be on the item another check is then made to determine the stat type, being one of thirty-one possibilities.
  3. The final check is made to determine the value of the rolled stat from the range of possible limits.

From a statistical standpoint we can get pretty close with these assumptions. But I think there is weighting that isn’t quite… that general.

I mean… think about how many “AOE Defense Ratio” items you’ve seen out there. I know I’ve seen hundreds. And I’ve gotten some rolls of that myself. If it had the same weighting as Movement Speed, or AOE Attack Ratio, then we would have seen a lot more of those stats; after all they all cap out at one possible value. The fact that AOE Defense Ratio is so prevalent, where AOE Attack Ratio and Movement Speed are not, makes me feel it is more of an unknown weighting issue as, if it were truly pulling one random value from the entire pool of possible values of all stats combined, we wouldn’t be seeing AOE Defense Ratio as much as we do.

And thank you all, updating the OP a little, however there is more that I need to see before I make any further changes. For now, though, it is still far more accurate than guesswork, even if the weighting issues throw off the generalizations.

I didn’t even know there is AoE defense ratio stat on headgear o_0.

There are so few characters who’d like to have it.

I’ve opened like 15 of them and my best were 40 attack and 38 SP regen, they really help a lot, especially the last one. I’ve got 2 times double stats, but they were trash (low level useless stats) so I rerolled.

Never got “1 point” stats yet.

Yeah the stat has been seen a lot actually, although tbh the only value I see in that stat is PvP. And, even in PvP, it likely wouldn’t be as good as the other stats… unless maybe if everyone in a GvG had the stat. Then it could probably cripple the enemy guild’s advance very easily. With every player in your guild, on the field, having +3 extra AOE defense Ratio… those enemy spells are probably going to cap out at hitting 2-3 people tops instead of washing through the crowd =P

Yeah, it’s very nice stat for tank, actually. But only for tank and not as good as some other stats you can get from enhancements.

This is sillyyyyyyyyyyyyyyyyyy, nobody needs max stats on anything. If anything, that increases the odds to get a highly valued stat, for example HP. Personally I’d label each stat in 3 tiers: low, normal, high, and how many each tier has determines the odds. 1 attack is literally going to do NOTHING, it’s crazy to think about anyone who would literally aim for these “perfect stats” and disregarding nearly as good stats due to a few points.

Just my opinion

Reading your post, kinda gives a little hope that those p2w items are not that game breaking. Unless you are A) super rich in which case you probably dont care enough, or B) Legally insane.

At the time I initially wrote the post it was as the hype and concern around P2W being implemented in-game was getting to be a concern. I started out just being curious about the odds but, after some looking into it, quickly came to the conclusion that the “perfect” items everyone kept quoting at the time was pretty far out of reach.

Before our EA people were stating that stacking 3 hair accessories with perfect (or even near perfect stats) would be very P2W and commonplace but after looking at the odds to even rolling the three stats you want, regardless of their values, it’s refreshing to know that it isn’t so.

The formula sadly only goes so far as pointed out, because getting the exact three stats you want (even if your odds by 2 to accept a “higher than middle ground” roll) is what it solves, but it doesn’t solve for getting other acceptable stats at high values until someone adapts it to their needs. It may be a million to one odds to roll with a specific stat combo and value in mind and get it, but getting something equally efficient for another class may be your ticket to a ton of silver to just buy other things. Or you wanted specifically Holy type elemental attack but got a high roll of, say, Fire. It may not be exactly what you wanted but it will do almost as nicely for what you had in mind.

So far it seems more to be sheer luck than P2W. In-game a high roll of 3 stats you want is like winning the lottery. If someone were to roll, say, near perfect stats for MATK, MAMP, and HP Recovery (for stacking amp with Animus let’s say) they could keep it on market at 10+ mil and it would sell. But to get those stats in the first place would be luck as nobody is going to spend $100,000+ USD just to roll that combination with high values.

The formula should still be relevant but should also be slightly off now since a couple stats aren’t showing up as often as expected due to weighting (weighting is assured, though, through mathematical process). How it is weighted though… I’m not sure.

Overall though, yeah, I think it’s nice to see that the numbers don’t suggest these stats are something we can pay to win. It’s more like a lottery ticket where everyone is a winner, but, how much you win is where the odds are against you. Going forward people may still have cause to raise concerns over P2W, though, since this feature has not only been implemented but now accepted in-game.

Watching how IMC approaches their next cash-shop consumables will be interesting.

I’ve moved on from the post =( lol I don’t have a means to calculate every possible combination of acceptable stats unless someone were to code something for it. @Excrulon could, for instance, make something depicting how good a roll is in-game given what we know so far, and maybe even code (as someone possibly volunteered earlier) something that would allow us to plug in our acceptable stat values and see the odds of getting them before we make our rolls.

For instance we could say “Well, I want MATK to be high, I want MAMP to be high, and I’d really like some kind of elemental damage on it but I guess the value isn’t too important” and then have something we just select “Must have: MATK > X MAMP >Y Any value for: Elemental attack A B C D E F G, roll!” lol But as it is even typing out examples takes work by hand and not everyone would find what they were looking for (my idea of a “high” roll may not be your idea of a “high” roll) and, something else to consider, not every stat shares the same pool of possible values, meaning it is more likely to get a high roll of one stat and less likely to get a high roll of another based on that pool. So it’s not something a couple blanket examples could cover unlike with perfect rolls which can be more easily defined in a post.

Too much work to try to give enough examples =( At this point there is not much else to be done except attempt to observe the stats we see and figure out how AOE Attack Ratio and Movement Speed are weighted (if they were simply pooled as a sub pick then we’d see a Perfect HP value the same number of times we see those two, and it could very well be how those two stats are considered when rolling, but in this sense that would be weighting against those stats as it is separate from how we’ve observed other stats to function). There still are a lot of unknowns and maybe everything listed in the OP is 100% false lol we can only hope to get close to the truth though since we have only speculation and no way at all to test outside of spending thousands of dollars with diligent reporting and documentation of the rolls we’ve received lol

Unless again someone were to code an addon that also collected which rolls we’d received and transmitted them to a database where odds were directly determined not by formula but by actual practice. Over time it would get more and more accurate as more and more in-game and actual rolls are added… but it would possibly be intrusive and would therefor require not only the source to be released but also user-consent. Excrulon may be about the only person I know so far that could swing that effectively and coupled with being able to get the odds on a roll you want it would likely be the perfect Magic Scroll addon.

It is too much work, I agree, and the post is at least humorous, it’s just.

Kind of misleading? At least to people who skim the OP and nothing else, which was all I was pointing out.

Cool work regardless

Yeah, there is no point in talking about odds to get top tier stats, as you will never reroll hat with 3 stats that are even a bit useful, unless they are really low.

So there is a point in dividing stats by usefulness in 3 tiers, giving 1 point to low tier, 2 points to med tier and 3 points to high tier. 0 points if it’s useless stat. Also add 1 point for every stat that is exceptionally useful. And then just consider a hat succesful if the score is 4 or higher. If calculating this way, then odds of getting succesful hat are not very low. And some of those succesful hats can be really impactful.