webpov

[webpov]

OW: How much time and games do you need to climb?

Hello everybody. This is a short, simple guide to calculate how much time and games you need to climb to your desired SR.

1. Get a rough idea about how much SR you gain and lose

Here, we just need a rough average. If you usually gain and lose the same SR, then you are ready to go to the second step. If that doesn’t occur in your case, then you just need to calculate the average:

$$Average\;SR\;gain=\frac{SR_{gain} + SR_{lose}}{2}$$

2. Check your current win rate at overbuff.com

If you already know how to check your winrate using overbuff or another similar site, then you can skip this step. Otherwise, you should write your battletag in the search button and click on your profile. There, you just need to click the button “competitive” in the right top corner and you will see your win rate %.

Reminder: You need to divide that number by 100 to get the win rate. For example $52.5\%=0.525$.

3. Calculate how many games you need to reach your desired SR

By using your computer’s or phone calculator you can find how many games you need to win in order to reach your desired SR. You just need to use the following simple formula:

$$Wins\;to\;Climb =\frac{SR_{desired} - SR_{now}}{SR_{gain\;per\;game}}$$

If the number is not round, then round it up. These are the excess wins you need to have in order to climb to your desired SR.

4. Calculate how many games you need to climb according to your current win rate

Until now, we have calculated how many wins in the row we need to climb to our desired SR. Let’s see how many games we need to play given our current win rate.

$$Games\;needed=\frac{Wins\;to\;Climb}{2\cdot(Win\;Rate-0.5)}$$

This formula assumes that you are losing and winning games according to your win rate and you do not have unexpected win strikes.

5. Calculate how much time you need to climb there

According to my experience, an average competitive game is around 15 minutes. This includes the downtimes, such as watching the play of the game, waiting until the next game starts, the loading and the hero pick menu. Also, we usually do a lot of stuff between games, like relaxing after an intense game or trying to figure out of what went wrong.

Never mind. You can find how many hours you need to climb as follows:

$$Hours\;to\;Climb=\frac{time\;per\;game \cdot Games\;needed}{60}$$

Now, if you have a schedule of how many hours per day you play, then you can calculate how many days you need including your breaks:

$$Days\;to\;Climb=\frac{Hours\;to\;Climb}{Hours\;per\;Day-Breaks\;in\;Hours}$$

6. Example

In our example we have the following data:

I roughly gain/lose 25 SR per game;

I play 4 hours per day;

I have a 30 minutes break in my gaming sessions;

My win rate is 55%;

I want to climb from low gold (2100 SR) to mid Platinum (2700 SR);

On average I need 15 minutes to finish a game.

$Wins\;to\;Climb =\frac{SR_{desired} - SR_{now}}{SR_{gain\;per\;game}}= \frac{2700 - 2100}{25} = 24$

$Games\;needed=\frac{Wins\;to\;Climb}{2\cdot(Win\;Rate-0.5)}= \frac{24}{2\cdot(0.55-0.5)}=\frac{24}{0.1}=240$

$Hours\;to\;Climb=\frac{time\;per\;game \cdot Games\;needed}{60}= \frac{15\;mins \cdot 240}{60} = 60\;hrs$

$Days\;to\;Climb=\frac{Hours\;to\;Climb}{Hours\;per\;Day-Breaks\;in\;Hours}= \frac{60}{4-0.5} = 17.14 \approx 18 days$

[webpov]

OW: The truth behind win rates and popularity

Hero popularity and its win rates have become the most prominent metrics to argue in favor or against his standing in the current meta. While they are two completely different things, they are treated more or less in the same manner. Some argue that popularity is the one that determines its state, while others defend their favorite hero using its win rates.

The Great misconception

At the moment, the most popular source is Overbuff. Here’s a brief misconception list:

the popularity rates are standardized on roster, and they are not based on their substitutes. In short, if we sum up the pick rates for the roster, the result will always be 100%.

The system does not count the probability of having the same hero in both teams. The reason that lies behind this is the necessity of assuming the composition. Therefore, no assumptions, no joint probabilities.

The popularity determines by how much the real win rates are inflated or deflated towards 50%.

To sum things up, the current popularity and win rate metrics are mere tools to find the actual pick and win rates. Both of them - as they are - are misleading and they do not enable any sense of interpretation.

How to find the real pick rates

Overbuff provides the roster pick rates without making any assumption. Thus, it does not weight the roster pick rates according to a set of viable compositions. Given the current metrics, we can observe the following;

Tank picks: 33.24%

Support picks: 32.27%

Offense and Defense picks: 23.93% + 11% = 34.93%

These metrics explain the current trend to stick with a 2/2/2 composition, while the most variations include either replacements of tank or support. Nothing more.

Since we have got this insight, it is safe to assume that people tend to use 2/2/2 compositions. Therefore, it is like comparing oranges with apples when we use a common denominator between tanks, supports and dps.

On average, 2 players per team will pick a tank, 2 will pick a support and 2 dps.

Thus, the real chance to see at least one particular support hero in a match (enemy or ally) is 80%!* Also, whenever a hero is not in meta, then the pick rates of its direct substitutes are inflated, and it is not accommodated by the hero pool as a whole.

* This assumes that players are indifferent upon the hero selection.

The following example explains the actual pick rates of Genji:

Overbuff Overall Pick Rates: 4.89%

Overbuff Offense Pick Rates: 20.42%

Overall DPS Slot Pick Rate: 14.17%

Overall DPS Slot Probability: 28.35%

Flanker DPS Pick Rate: 48.22% (Tracer 40.04% & Sombra 11.74%)

Probability to engage 2 Genjis in a game: 8.04%

As you can see in the example above, I mentioned both Flanker and overall DPS slots. According to the assumptions, we get the corresponding results.

The win rate inflator & deflator

Thanks to Overbuff, we have several metrics to measure performance and evaluate the progress of the heroes. However, the win rates may mislead us to wrong results due to the lack of assumptions. As we already know, the win rate is calculated as follows:

However, there is a problem in this formula. What will happen if two of the same hero face each other in a match? In this case, one will certainly win and one will lose. So, assuming that the players try to form specific composition and they do not pick randomly heroes, then

The probability of engaging the same hero depends on the popularity. As you can see in Genji’s example, there’s the probability of engaging 2 Genjis in the same game. Given the above formula, Genji’s actual win rate is 50.38% and not 50.42%.

As the popularity increases, then the actual win rates are either inflated or deflated towards 50% even more, because the chance to have the same hero in both teams increases.

Conclusion

There is no safe way to determine whether a hero is over- or underpowered. Win rates - certainly - can give us useful insights for the current hero standings, but they are not enough. Popularity, on the other hand, either deflates or inflates the winrates towards 50% and the lack of sample size makes it impossible to calculate the error margins and evaluate our results.

[webpov]

Brigitte will give hard time to flankers, they said.

Amélie Lacroix

[webpov]

OW: Poster’s Fallacy Awards

Introduction

Reading carefully many threads per day, I was wondering why the majority of posters end up seeing things as “

black and white

” propositions, such as the aim is not a skill, Moira is either broken or perfectly balanced, Mercy is unplayable and so on. So, I have come here to discuss this peculiar phenomenon and listen to your opinion.

The sad Truth

For some reason, the mentality that “multiple truths coexist” dominates these forums. Indeed this can be observed in any topic, including elementary math.

We have learnt in school that 1+1=2. However, is it a universal truth? If we speak about: Calculus, then indeed 1+1=2 Binary entities, then 1+1=10 Roman numbers, then 1+1=11 Boolean logic, then 1 AND 1 = 1

However, is this the case? According to our example, it is impossible to do a mistake if we have set proper conditions to our problem. Will we ever use Boolean math if we are asked to count apples?

Therefore, there is always a possibility to find a single truth to the most of our problems, if we set the right conditions.

The Dominance of multiple truths

The explanation is rather simple. Social scientists, such as economists, describe this as “satisficing”. We speak about complex topics, which sometimes do not even have an answer. So, we try to give a solution to these problems by searching a satisfying and suffice solution.

Satisficing Coastline Problem

Let’s say that 2 different people decided to measure the US coastline. The first one found 200,000 km, while the second 260,000 km. However, both of them were right. The first found satisficing to measure it using a map with 1:2,000 scale and the second using one with 1:1,000. Therefore, both of them are correct as long as there are no coherent criteria.

Satisficing Overwatch Problem

Many Mercy mains have expressed their disappointment about the multiple resurrection rework. It was balanced, they claim. However, as we presented (check first example), in order to solve a maximization problem in a blurry environment, we need a satisfying and suffice solution. The first Mercy was sufficient in terms of mechanics, but she was not satisfying to play against. Therefore, there was a change that broke a relatively balanced hero, because the constraint was set wrongly.

As long as we cannot agree upon what is a suffice and satisfying solution, we will never have one.

The Nerf-Buff water Park

Sometimes because of stubbornness, some others because (s)he is our main, we try to retain the right of absolute truth. We make assumptions whenever it is convenient and then we break them. Let’s demonstrate two examples.

The “Moira” Example

It’s convenient to let aside Moira damage orb/click combo, because then she does the lowest dps in the game. On the other hand, the ones who oppose Moira’s current state, assume that she is always off cooldowns and she can use her offensive potential without trade-offs. Both of these statements are proposed in a way to satisfy the poster’s love or hate for a specific hero or situation.

The “Genji” Example

There are a lot around who demand Genji Nerf. According to my personal experience, the majority of these posts go around deflect ability and his burst potential. Some assume that if his deflect is visually broken, then it’s mechanically broken as well, which is a false statement. Other claim that supports are free kills. Empirically speaking, there are no evidence that Genji can reliably kill enemy supports in any given situation (that’s what free kill means at least). Sadly, there are not many Genji defenders to represent a reverse false statement.

Farming Likes

After the release of the new forums, I observed that there are more and more “like farmers”. Popular - false - threads and posts, which have as goal to get likes. It’s amazing how easily you can gather people and get likes if they feel you sympathize them. Also, you get a trust level reward if you have the time to poke enough and whisper “you are not alone in this life”. The most prominent example was the spam of “Mega-thread” titles for the same hero when the forums opened. Other examples are the “this hero is just fine” or “we need skins for this hero” threads.

[webpov]

Does Blizzard abuse the Code of Conduct?

Hello everybody. I enjoy a temporary ban from the U.S. new forums, so I have decided to share my story. Before that, I would like to start with an auto-generated sentence that comes along with any Blizzard Forum Moderation Team response to the appeals:

Please note that Blizzard Entertainment reserves the right to suspend access to these forums at any time for reasons that include, but are not necessarily limited to, repeated failure to abide by the Community Code of Conduct.

-  The Forum Moderation Team

The “Code of Conduct” is defined as follows:

Principles, values, standards, or rules of behaviour that guide the decisions, procedures and systems of an organization in a way that (a) contributes to the welfare of its key stakeholders, and (b) respects the rights of all constituents affected by its operations

Source: I.F.A.C. & Wikipedia

In short, Blizzard is not necessarily limited to its code of conduct and so, there is not any defined line that protects the rights of the constituents affected by its operations.

For example, does it makes any sense for the police to be able to hold guns and at the same time it is not to be limited by the law? If this is the case, then who can guarantee that the police officers will not abuse the power they hold?

Hence, there are several ethical considerations that arise in this issue. If the company can exempt cases and people at its convenience, then we are all subjects beyond any form of social contract, and therefore cumulative integrity can be achieved if, and only if, every single employee is tolerant and benevolent enough to endure our presence.

However, when it comes to the appeals, the problem becomes even more complex. As we all know, the current ban system is fully automated, prone to statistical errors, also known as “false positives”. The support team comes to read and evaluate the appeals to decide whether the case was a false or a true positive. Everything seems fine so far.

Although, a major problem arises if the subject of the appeal needs critical thinking from the behalf of the support. At this state, an automated system has already decided upon the fate of the charged, and only a single individual has the power to undo the running charge.

If the support is under a dilemma, then the possible outcomes are in a binary form i.e. remove a ban or not:

Take the risk as an individual, remove the ban and admit that your system is not perfect.

There is no smoke without fire. The chance of a “false positive” is already diminished, so whenever in doubt the easy way is to “trust the process”. After all, the company can exempt any user at any time from the Code of Conduct.

The repercussions of the current process

Sadly, the second outcome contains a crucial fallacy. If the individual does a single mistake, then inherently the user will never trust the process again. Not only the system failed him, but the company itself. Profound emotions, such as over-simplistic generalizations and other biased assumptions come to surface; distrust between the consumer and the company arises along with uncertainty of future behavior.

While there are numerous possible outcomes after this process, I would like to focus more on the two extremes. The first is the appearance of aggressive behaviour and the second is discomfort and distress. While aggression is self explanatory after this misjudgement, discomfort occurs because the user starts to behave with exhaustive to him caution.

Those repressive feelings are included and they are punishable according to the Code of Conduct:

Harassment takes many forms, and is not necessarily limited to the type of language used, but the intent. [...] The idea behind this is to prevent any one player from consistently being uncomfortable in the forums.

Source: https://eu.battle.net/en/community/conduct#harassment

In this case, does the company violate its own code of conduct?