# Talk:Mystic Clover

## Contents

- 1 Can some explain
- 2 RE: Relyk
- 3 Bottomline
- 4 Other drops?
- 5 drop rate
- 6 Obsidian Shards
- 7 Notes
- 8 1 or 10 clover recipe
- 9 spoiled data
- 10 Bernoulli
- 11 tier 6 refined materials
- 12 The note about the one recipe vs the ten recipe
- 13 New WvW-Reward-Tracks?
- 14 Fundamental assumption that is stated nowhere that I can see?

## Can some explain[edit]

Can some explain what is meant by "but the 1 clover recipe is less likely to vary from the average, and is therefore less of a gamble." as it makes little sense in its current state. Obviously if they have the same statistical chance of giving a clover or, for that matter the same number per skillpoint of crafting materials, the average should be the same. My only current guess is that the 10 clover recipe gives all or none (as in you either only get clovers or only crafting materials), though I haven't had time to test this. If anyone can understand what was meant by that phrase and prove it, please feel free to either clarify the phrase or leave a comment here. -Erukor (HoD, 20/9/2012) --*The preceding unsigned comment was added by* 192.168.104.79 (talk).

- If you try the 10x recipe 8 times, and the 1x recipe 80 times, you have a much higher chance of getting no clovers from the 10x, because you only did it 8 times. The 1x recipe will give you more trials, thus putting you closer to the statistical average. Of course, it's also more likely that you'd get 80 clovers doing the 10x 8 times than it is to get 80 doing the 1x 80 times, but as the chance of clovers is probably fairly small that's not a safe bet at all. Pie Formation 18:02, 20 September 2012 (UTC)
- Exactly what Pie Formation said. You're more likely to either hit the jackpot or bust if you use the 10x recipe, but you're more likely to get average results with the 1x recipe. It's a basic conclusion from the Law of Large Numbers. If you do enough trials, you eventually approach the average. The 1x recipe requires more trials, and is thus more likely to approach the average than the 10x recipe.Strill 08:53, 24 September 2012 (UTC)

Anyone have the data for the statistical claims? If that data exists it'd also give an approximate success rate which would be very useful. --24.28.69.166 18:33, 22 September 2012 (UTC)

- I came to ask the same thing. Could someone create a table in this talk page where we can aggregate statistical data? I'd create it but I am not sure how. I will happily add all of the results of my attempts once it exists.68.11.147.218 17:53, 23 September 2012 (UTC)
- I have seen no statistical basis for the claim that both recipes are the same. The claim that the 1x recipe is more likely to give average results than the 10x recipe assuming both have the same success chances is based on the Law of Large Numbers as I mentioned above.Strill 08:53, 24 September 2012 (UTC)

## RE: Relyk[edit]

What you said is not at all the same as what I said. This is what you said:

"Statistically, both recipes give the same number of clovers on average, but the 1 clover recipe is more reliable if you only need a single clover."

First of all, we have no information whatsoever to indicate that both recipes are the same chances. Second, "reliable" is vague. Third, the number of clovers the user needs is completely irrelevant. The 1 clover recipe is ALWAYS more likely to give average results than the 10 clover recipe assuming both have the same success chances.Strill 11:28, 24 September 2012 (UTC)

- Until you can show the probabilities are difference, it's safer to say that they are the same and doing in bunches of 10 are just for sake of convenience. That's pretty much what they do everywhere. See w:Reliability (statistics) for what reliable means; pick any word that describes a smaller standard deviation. "always more likely" is very poor language that doesn't mean anything to the average user and sounds dumb otherwise. Mentioning statistics at all is bullshit since no one knows the probabilities or if the 10 clover recipe is done on a single roll or 10 individual rolls.--
**Relyk**17:20, 24 September 2012 (UTC)- Why should we assume that the success chances for each recipe are the same and misrepresent this fact to readers when we don't know it ourselves? I see no reason why we can't make the current state of research clear, nor do I see how practices elsewhere have any bearing on this case. This is not a real world statistical case and there's nothing forcing the two recipes to have the same chances. For all we know, the 10x recipe could have a higher overall chance in order to balance out its inconsistency.
- For your second point, when did anyone say "always more likely"? I'm looking through the history and I don't see it. As far as my concern with the term "reliable" goes, my main problem is that when using it so concisely, its meaning will likely be lost or ignored. This is clearly not an intuitive concept, and I think it bears repeating at least once.
- For your third point, No one has yet reported even a single example of the 10 clover recipe giving any amount of clovers other than 10, and we have around 40 or so samples. It's safe to say that it's done using a single roll, which justifies the statistics as presented with their current caveat.Strill 18:01, 24 September 2012 (UTC)

I have done the 1 clover recipe 245 times and got 77 clovers. The succes rate was 31.4% for me. Of course this is not the general success rate, but it might be helpful for other players. I think rate is exactly 1/3, but more data is needed to prove this. Umnachtung 00:59, 25 September 2012 (UTC)

- I just completed my 77 clovers, 229 attempts at the 1 clover recipe for an average of 33.6% for me Zophar 21:00, 31 October 2012 (UTC)

## Bottomline[edit]

Bottomline being the 10clover recipe is for people feeling lucky, the 1clover recipe for people who don't want to take a risk and are more happy with a mostly predictable amount of stuff needed. Everyone with basic understanding of statistics knows this, however clarifying this to people without proper education seems difficult. Redirecting to any wiki page doesn't cut it imho, could some people please suggest ways to clarify this to people without relevant education.

I will try immediately by comparing die tosses by assumin the 10clover recipe also has a 1/3 chance:

- Getting Mystic Clovers is a gamble that can be completed through the 2 meanings of the word "fortune". Either you're very lucky or very rich. Either will get you the 77 Clovers needed. Any time you use one of the clover recipes you are rolling a die, every time a 5 or 6 comes up you will get 1 or 10 clovers depending on the recipe used. By using the 10 clover recipe a lucky player would only need 8 rolls of the die, and this is a chance that will happen once for every 6500 players. On the other hand, one very unlucky player in 6500 players will end up with nothing with all other players having between 10 or 70 clovers.
- The 1clover recipe needs to be successful 77 times. On average, getting a five or six on a die 77 times will require 231 rolls. Getting so unlucky to end up with no clovers at all in 231 tries is such a small chance that it probably won't happen ever in the lifetime of the game (although improbable things happen). On the other hand, getting all 77 clovers in exactly 77 rolls of the die will require the luck of one person in about a billion earths which is also very unlikely. Over 90% of people will end up with at least 70 clovers after those 231 rolls of the die.
- So what does your "fortune" look like? Do you feel lucky or are you rich? Someone feeling lucky could try the 10clover recipe, although this player should know he's gambling with high risks. A rich person would be better of using the 1clover recipe and being fairly sure to get the 77 clovers required.
- Calculation can be found on this page.

Someone please clean this up more and I'd also love someone with credentials making a proper page with statistical analysis so people with non-mathematical educations can trust it. As far as losing 231 consecutive gambles goes, a few funny things: it's actually so unlikely it won't happen in the lifetime of the universe although that's silly. It's also more unlikely to happen than someone correctly calling every single card of a well-shuffled deck without cheating.62.159.14.62 13:55, 6 November 2012 (UTC)

- If you mean by credentials, a mathematician, statistician, or just someone who can do the math? I'm not posting my cert here :P but either way I'd say its good enough to use knowledge of math to easily illustrate this 10 versus 1 odds. I'll try to rehash the costs and odds in a bit.A Liability 01:41, 7 November 2012 (UTC)
- With a 1/3 chance of success for both the 1 clover and 10 clover recipe the costs associated with obtaining 77 clovers for a gift of fortune is equivalent for the average result. With the 10 clover recipe it is more likely to either have a lower overall cost or higher overall cost. This is because a 10 clovers recipe has a 1/3 chance to get 10 clovers and a 2/3 chance to get no clovers compared to the 1 clover recipe which has a very small chance (one in 3^10) to receive all 10 clovers and a small chance of receiving no clovers (two in 3^10) with the average around 3.33 clovers per 10 tries. Simply put a 1 clover recipe averages risk over the course of obtaining 77 clovers at the detriment of having to expend resources close to the average, where as a 10 clover recipe has a higher potential to be cheaper or more expensive to obtain the same said 77 clovers.A Liability 02:44, 7 November 2012 (UTC)
- It's all saying the same thing. 10 clover recipe will have more variance because of the smaller n-value. The average user won't be able to tell the difference anyways since the probability, hence expected average, for both are identical (ignoring the "we don't know" bullshit).--
**Relyk**02:58, 7 November 2012 (UTC)

- It's all saying the same thing. 10 clover recipe will have more variance because of the smaller n-value. The average user won't be able to tell the difference anyways since the probability, hence expected average, for both are identical (ignoring the "we don't know" bullshit).--

- With a 1/3 chance of success for both the 1 clover and 10 clover recipe the costs associated with obtaining 77 clovers for a gift of fortune is equivalent for the average result. With the 10 clover recipe it is more likely to either have a lower overall cost or higher overall cost. This is because a 10 clovers recipe has a 1/3 chance to get 10 clovers and a 2/3 chance to get no clovers compared to the 1 clover recipe which has a very small chance (one in 3^10) to receive all 10 clovers and a small chance of receiving no clovers (two in 3^10) with the average around 3.33 clovers per 10 tries. Simply put a 1 clover recipe averages risk over the course of obtaining 77 clovers at the detriment of having to expend resources close to the average, where as a 10 clover recipe has a higher potential to be cheaper or more expensive to obtain the same said 77 clovers.A Liability 02:44, 7 November 2012 (UTC)

- I don't think it's appropriate to
*assume*that the recipes deliver clover at the same rate:- GWW has long followed the policy to avoid posting drop rates unless the data supports them. There is data for a 1/3 chance of clover from the short recipe, but none for the long recipe (at least none published or linked here).
- ANet's quant people are clever enough to know that the law of large number would favor the short recipe, so why would they bother creating a nearly identical & more expensive recipe using the same odds? This isn't meant to
*prove*that there are two rates, only that there is good reason to avoid making the assumption that there is only one rate.

- 75.36.177.28 10:08, 9 November 2012 (UTC)
- The two recipes cost exactly. Crafting 10 at a time is simply for convenience. If you don't want to assume anything, we have no data for 10 clover recipe, so comparisons between the two recipes shouldn't even be there. The fact they cost the same and that you need ~80 for the legendary weapon is like Arenanet holding up a big red sign that says "WE UNDERSTAND CRAFTING ONE CLOVER AT A TIME IS RETARDED, SO WE ARE GOING TO LET YOU CRAFT 10 AT A TIME!"--
**Relyk**10:32, 9 November 2012 (UTC)- Didn't read the whole topic, and don't have much time. I just wanted to say that I am a mathematician with emphesis on statistics. So you can formulate a question, I can give you the answer. - Yandere 12:35, 9 November 2012 (UTC)
- Did the 10x clover recipe 28 times and it's about 33%. It's gambling though since you have a high chance of getting excess garbage instead of lodestones. 24.216.239.40 21:18, 10 November 2012 (UTC)
- Just did it another 22 times, got clovers 7 times with the "10-clover recipe". 24.216.239.40 05:12, 11 November 2012 (UTC)

- Didn't read the whole topic, and don't have much time. I just wanted to say that I am a mathematician with emphesis on statistics. So you can formulate a question, I can give you the answer. - Yandere 12:35, 9 November 2012 (UTC)

- The two recipes cost exactly. Crafting 10 at a time is simply for convenience. If you don't want to assume anything, we have no data for 10 clover recipe, so comparisons between the two recipes shouldn't even be there. The fact they cost the same and that you need ~80 for the legendary weapon is like Arenanet holding up a big red sign that says "WE UNDERSTAND CRAFTING ONE CLOVER AT A TIME IS RETARDED, SO WE ARE GOING TO LET YOU CRAFT 10 AT A TIME!"--

- I don't think it's appropriate to

- The expected amount of tries is too low to have the two ways 'statistically average out'. 87.99.14.138 13:48, 23 November 2012 (UTC)

- I feel like it's worth mentioning that the 20-50 rule for the 10 stack is incorrect, on my first attempt I got 17 ancient wood logs (yeah miserable fail, ill only be doing singles from now on.) But if singles are a guaranteed 2-5, that already puts it at a better value recuperation rate than the 10. --
*The preceding unsigned comment was added by*71.195.219.175 (talk).

## Other drops?[edit]

I tried the 10 clover recipe and got 10 Crystal Lodestones. Do these count as a Tier 6 material or should the wiki page be amended? Svartalf 14:48, 29 November 2012 (UTC)

- Having looked up the page I linked, it appears they are a Tier 5 material. The wiki page should definitely be changed to reflect this (it appears all drops are nerfed game-wide, at the moment. Svartalf 14:51, 29 November 2012 (UTC)

I got 3 Mystic Coins whith the one-clover recipe, and also Lodestones -- Crazy 07:13, 6 December 2012 (UTC)

I've got 20 globs of ecto with ten-clover recipe. --Lordnex 13:09, 15 December 2012 (UTC)

- Same as Kirbman -- Crazy 09:53, 22 December 2012 (UTC)

- and 5 Mystic Coins
*talk*05:12, 27 December 2012 (UTC)

- and 5 Mystic Coins

- I got 50 Mystic Coins —
*The preceding unsigned comment was added by*75.65.254.218 (talk • contribs) at 03:46, 8 April 2013 (UTC). - I also got 50 coins --Geekers (talk) 18:22, 29 March 2015 (UTC)
- I just got 50 mystic coins from the 10x recipe, and 2 T6 fangs from the 1x recipe 00:04 - 04 October 2016 (UTC) Banrigh

- I got 50 Mystic Coins —

## drop rate[edit]

Mystic Clover/Drop rate for tracking items received from 10 clover recipe. There will be people who want each type of crafting material and amounts, you can make tables for that too.--**Relyk** 08:39, 6 December 2012 (UTC)

Somebody changed the drop rate from 33% to 20%. Is there solid evidence behind this? If so can it be linked here? I only see the original source and my source for drop rates, both of which are > 20% by no small margin. Kojin 05:07, 8 January 2013 (UTC)

I did the change, based on applying the recipe 150 times on Jan 17, and getting 29 clovers. Before the January 10 release I applied the recipe 100 times and got 35 clovers. I would say it is a pretty good statistical sample. Not sure what kind of "solid evidence" you are looking for, nor do I care you removed potential useful information. Do as you please.

- Can confirm this a little, I would say I minimum put 300 ectos for 77 clovers. Unfortunately I have not wrote down that exactly. Also I'd like to add, I believe, whenever you "start" your "forging", you have to spend more to get to the result. like, 8 times crap and only after that some clovers --213.196.136.71 02:26, 12 January 2013 (UTC)

- Also confirmed rate drop to 20%. Before Jan 10 update I did 100 batches of the 1 clover recipe, resulting in 35 clovers. On January 17 I did 150 batches resulting in 29 clovers. Did not notice any other changes in drops.

- April 5, 2013 - 71 Clover from 190 attempts using single recipe- ~37% Grishkafool 14:49, 6 April 2013 (UTC)

Drop rate certainly is well below what it says here. On my second legendary, don't have exact figures, but closer to 20% than 37.... the figures given here are overly optimistic. Update 9 months later: On my 5th legendary, always around 20% drop rate. I'm aware that this could be a fluke, but then: The offical stats could be a fluke. Still don't believe in the 37% drop rate. Poom 121.72.223.98 07:11, 28 August 2013 (UTC)

- You can see here that quite a few people contributed/documented their results, which is the basis for the 30-32% assumption.
- Out of interest: Which recipe did you use? If you used the one with 10, the potential deviation is a lot higher than with the 1 recipe and could easily explain your low rate, even with the 385 clovers needed for 5 legendaries. --aRTy 22:00, 24 June 2014 (UTC)
- The forge likes giving me clovers, so my drop rate with the 1 recipe is closer to 50%. YMMV. Vili 点 00:27, 25 June 2014 (UTC)

- As of recently (Dec-2015) the rate is 10% or less.

## Obsidian Shards[edit]

Just got 2 obsidian shards when I tried to make a clover...is this new? 80.112.181.245 20:17, 5 January 2013 (UTC)

- I tried the 10x recipe for the first time and got 20 crystals in return. Never felt so ripped off. (Posted my findings to the drop rate discussion page)76.105.103.27 10:56, 21 January 2013 (UTC)

## Notes[edit]

It says that on average it costs 5 ectos/shards/coins/skill points per clover. Unless theres something I'm missing on average it takes 3 mystic clover recipes to get 1, which gives 3 ectos/3shards/3coins/2.4 skill points. Update these numbers, feel free to revert if im wrong Calach 06:21, 12 January 2013 (UTC)

- If someone wants to change the numbers, provide the correct formula for calculating the average,
**do not**simply replace the equation with an integer.--**Relyk****~**talk > 22:21, 24 January 2013 (UTC)

## 1 or 10 clover recipe[edit]

Which one is the best recipe in your opinion?

- Depends on how much time you have. 1 clover recipe requires more clicks to produce 77 clovers, 10 clover recipe requires higher risks - you can get like 20 useless gossamer bolts.
**MalGalad**21:23, 25 April 2013 (UTC)

- Then, time apart, 1 clover is safer?

- I'm not an expert in probability theory, but yes, on such small numbers it's safer.
**MalGalad**21:22, 25 April 2013 (UTC)- Say you are choosing between the 1-clover recipe 100x or the 10-clover recipe 10x. It's easy to see being unlucky 5 times in a row, lucky once, and unlucky 4 more times. It's harder to get unlucky 90 times out of 100 tries. You can experiment with this using a google spreadsheet and its random number generator and see that the 1x recipe is more likely to provide average results for 100 tries compared to the 10x recipe used ten times.

- I'm not an expert in probability theory, but yes, on such small numbers it's safer.

- On the other hand, if you are on some sort of lucky streak, then it's easier for the 10x recipe to give you awesome results. As noted above, 1x recipe is less likely to give you horrible luck, but the 10x recipe is less work. The main thing to remember is, like all gambling, the house always wins -- it's a gold & material sink, so expect to be unlucky. 22:12, 25 April 2013 (UTC)

## spoiled data[edit]

**←***moved to Talk:Mystic Clover/Drop rate#spoiled data*

## Bernoulli[edit]

**→***moved from Mystic Clover*

The Mystic Clover is a Bernoulli experiment with Success=Clover and Fail=No Clover. Because of this we can use the Binomial Cumulative Distribution Function to determine our chance to successfully acquire what we need. The community has found that the success rate is 30.71%. Using this data we can determine;

- The chance to get
**AT LEAST 8 success**out 20 trials is 24.99% (10 Clover Recipe) - The chance to get
**AT LEAST 77 success**out 200 trials is 1.15% (Single Clover Recipe) - The chance to get
**AT LEAST 8 success**out 25 trials is 51.93% (10 Clover Recipe) - The chance to get
**AT LEAST 77 success**out 250 trials is 51.15% (Single Clover Recipe) - The chance to get
**AT LEAST 8 success**out 30 trials is 74.62% (10 Clover Recipe) - The chance to get
**AT LEAST 77 success**out 300 trials is 97.625% (Single Clover Recipe)

The conclusions we can draw from this data is that on average it will take about 250 trials to get the Clovers you need. Below the 250 trial mark the 10 Clover recipe is superior to the single clover recipe, and above the 250 trial mark the reverse. However below 200 trials both recipes have low chances to produce the quantity of clovers needed. Since at our critical value of 250 trials both are equal, and above the critical value the single is superior we can conclude;

**The most optimal way to obtain your 77 Clovers is to use the single clover recipe. You can expect to need between 250 and 300 trials worth of materials.**

—*The preceding unsigned comment was added by* KitenMitens (talk • contribs) at 14:41, 21 October 2013 (UTC).

- I moved this from the main article because it's evidence in support of the conclusion that the theoretically "optimal" strategy is to use the single clover recipe 250-300 times. 75.37.16.111 16:42, 21 October 2013 (UTC)
- I don't get why the conclusion is 250-300, instead of 200-300. Either we say it's about 250 on average, or we take the full interval of +/− 50. --aRTy 17:18, 21 October 2013 (UTC)
- The idea is you should not assume it will take any less than 250 trials. The CDF shows that ~50% of players will succeed in getting 77 clovers by 250th trial. The other 50% will not and will need more trials. So yes it would be appropriate to assume 250 as the average # of trials needed. However my intent was to show definitive evidence that the single clover recipe has the greatest chance to successfully produce 77 clovers with as few trials as possible. 184.88.58.21 20:10, 21 October 2013 (UTC)
- That's the expected value, you don't need use a CDF to calculate that. That's why it's simple enough to suggest the average cost of each clover. There are obvious reasons the single clover recipe is optimal. You only need 77 instead of 80 clovers and around 250 trials to obtain the clovers instead of 25 trials. The difference in probability for the two recipes would have to be pretty large for 10 clover recipe (as you pointed out with the CDF values). None of this is worth the main page, as you can tell from previous discussion, people will disagree or most users won't understand what it means.--
**Relyk****~**talk < 00:47, 22 October 2013 (UTC)

- That's the expected value, you don't need use a CDF to calculate that. That's why it's simple enough to suggest the average cost of each clover. There are obvious reasons the single clover recipe is optimal. You only need 77 instead of 80 clovers and around 250 trials to obtain the clovers instead of 25 trials. The difference in probability for the two recipes would have to be pretty large for 10 clover recipe (as you pointed out with the CDF values). None of this is worth the main page, as you can tell from previous discussion, people will disagree or most users won't understand what it means.--

- The idea is you should not assume it will take any less than 250 trials. The CDF shows that ~50% of players will succeed in getting 77 clovers by 250th trial. The other 50% will not and will need more trials. So yes it would be appropriate to assume 250 as the average # of trials needed. However my intent was to show definitive evidence that the single clover recipe has the greatest chance to successfully produce 77 clovers with as few trials as possible. 184.88.58.21 20:10, 21 October 2013 (UTC)

- I don't get why the conclusion is 250-300, instead of 200-300. Either we say it's about 250 on average, or we take the full interval of +/− 50. --aRTy 17:18, 21 October 2013 (UTC)

## tier 6 refined materials[edit]

I think the amount stated in the article (1-5) is wrong. From what I see the listing was carried over from the mystic forge article a long time ago and because people did not differ between the 1 and 10 recipe properly back then, the numbers got a bit mixed up. However, there is a google doc sheet link in the clover article history (there you go) with 267 tries and I have my own table (from when two friends and I forged our legendary weapons) with another 428 documented tries in total. For all four refined materials both tables state amounts of only 1-3 without an exception.

I don't have data about the 10 recipe at hand, but I strongly assume it is 10-30, accordingly. If anyone ever got more than 3 resp. 30 of the refined materials, please leave a comment. For now, I'll just go ahead and edit the article. --aRTy 05:34, 4 February 2014 (UTC)

## The note about the one recipe vs the ten recipe[edit]

It should state that the one recipe will most likely be closest to the estimated about of materials predicted.... not less materials used.

Since you need to repeat the one recipe so many times, it will be much closer to the predicted amount of materials needed than the ten recipe which can have a wilder variation.

...this just puts your actual results very close to the predicted results. It does not save you any shards, coins, skill points, or ectos as the note states.

If anything, since the ten recipe has such a wide variation, it is more likely that you will save materials using it rather than the one. 76.16.108.117 03:55, 10 March 2014 (UTC)

Results *on average* will be the same for both. As (ip) stated, the difference will be the variation : using the 1 recipe will probably lead you closer to average cost, while using the 10 recipe is more likely to lead you to much lower of much higher cost. Eowin (talk) 21:34, 29 March 2015 (UTC)

## New WvW-Reward-Tracks?[edit]

Do they also drop from the new WvW-reward-tracks? --62.155.144.85

## Fundamental assumption that is stated nowhere that I can see?[edit]

All this talk about the 10 recipe being higher variance is true <==> the 10 recipe *always* produces 10 of the same item. If it calculates rewards one-by-one, then the claim is patently false. I have no idea which is the case, but evidence for this assumption stated or justified nowhere on this talk page or the main wiki page. I'm assuming it's one of these things which is just "common knowledge"?. --Notso (talk) 12:06, 6 June 2016 (UTC)