User:Coran Ironclaw/Finding the best solution for gladiator title

Note: This article is incomplete and it is stopped. Ryan listened to my points. You may still post comments on the talk page if you want.

Reason to care

I don't know what methodology Anet is using to confront the issue of updating the gladiator title. But I find the 3 proposals to be speculative only, leaving aside analysis that can approximate desirable values such as the total effort a player has already invested on that title. Without that type of analysis there are high risks to:

• Devalue the title if the new system results in much less effort required than the previus one. (Giving some more points to the current title holders does not help much, since you are lowering the total life span of the title.)
• Implement an unfair system for any of the two arenas. In other words, that it can be a lot easier to ascend on the title on one arena than the other. Such a system will eventually bring a lot of complains from the less favorable arena community.
• Overrate the effort needed to get high tiers, making them unreachable. (this do not apply to the current 3 proposals made by anet since the gladiator title would be clearly easier to attain)
• Failing at making the title ascension smoother. This is what I consider is the main reason for updating the gladiator title, to bring a system smoother, less chaotic, rewarding more constantly and not just every hit of luck. That way new players have more incentive. This is what making the title more accesible really means "we want to maintain Gladiator as a skill title, but we want to make it more accessible also (Ryan Scott)"
• Implement a title that differentiate little from low skilled players to high skilled ones. In other words, not implementing a skill title but a grinding one.

Design Guidelines

I now write what in my opinion should be the design guidelines for updating the gladiator title, (Ryan please feedback me on this):

• Ra and Ta must have the same title. (splitting title brings other serious issues anet dont want to deal with now).
• Since gladiator is not a new title, any change on it should reflect all the effort the players have invested. (This brings the issue of measuring the player effort).
• Ra and Ta are two different valid pvp gamestyles, two equal skilled players who prefer different arena should receive a similar reward if they invest the same time, being able to get to its highest tier. (This brings the issue of looking for a proposal that reflects an effort for each skill level).
• Any new proposal needs to have a smoother ascension than the current one. What this means is that every winning battle should be rewarded (save maybe for the very first one to prevent leeching or even second if it becomes a huge problem), and please you must undertand one thing: Not because every win is rewarded the title is easier to obtain, we are not saying that every win will be equally rewarded. The effort must be maintained as the second guideline implies. The title will be equally difficult, it will just reward smoother, less chaotic, making it more accesible. That way new people will know they are advancing, maybe a little but they will be, receaving an incentive to continue.
• Any new proposal needs to differenciate low skilled players from high skilled ones, rewarding better the second. The same way as the current system does (or even better). This way you guarantee it is not a grind only title, but a skill one. (of course, it should also be time consuming to a certain degree).

Maybe many are thinking it is impossible to fulfill all that guidelines, or there is no way to measure what it is required. But it can be done, the most difficult part it is not the analysis, but the data required to start it. As I previusly stated on the topic "Do you monitor RA? Could you? Statistical Analysis" on User talk:Ryan Scott page.

Measuring the effort current gladiators had invested

First, we need to find the amount of effort current gladiators have invested on their titles. We define effort as the gameplay time, but not all gameplay time is the same. The gameplay time for a high skilled player is worth more than of a low skilled one. But how are we going to find that time if all we have is the total glad points each one has earned? We must aproximate the time using in-game data.

Monitoring the winning streaks on ra and ta: The data required must come from monitoring the arenas and extracting information to create the following table (one row each time a player looses or leaves a battle):
Example:

Arena	Glad Rank	Consecutive wins on the run
1	3	4
2	2	7
so on...	so on...	so on...

On the arena column: 1 could mean ra and 2 could mean ta.
The range in time should be a week to consider every type of player, since most casual players enter weekends and a lot of hardcore players dont play on weekend.
All the data obtained must be compacted into 24 relative frequency tables, 12 for TA (one for each glad rank) and another 12 for RA, like the following:
Ra - rank 0 frequency example table (Example)

Streaks	Relative frequency	Notation
0	0.3	Pra(w0|s0)
1	0.2	Pra(w1|s0)
so on...	so on...	so on...

Streaks should range from 0 to 10 for ra, and 0 to 50+ for ta.
Relative frequency is the aproximation of the probability a player has to obtain certain win streaks on a run. It is calculated from the all the data by the total times a player with that rank in that arena obtained that number of streaks divided by the total number of runs players with that rank in that arena did.
Notation, since we are going to use that numbers of relative frequency later, a compact notation is in hand.
That way Pra(w#B|s#C) reads: Probability in RA that a skill level #C player does a #B winning streaks on a run.
and Pta(w#B|s#C) reads the same but for TA.
Since RA and TA most probably have different values of probabilities and we dont know where does each player has obtained its rank, we must correlate them to obtain a single usable value. Using the same in-game data we generate the following table:

Rank population between Ra - Ta (Example)

Rank	Matches in Ra (MRA)	Matches in Ta (MTA)
0	4532	1389
1	3781	1873
so on...	so on...	so on...
11	not sure if	they already exist

The matches are the sum of all (win streaks + 1) because there is a lose at the last of each run. So, the value of interest for each rank is calculated by the relationship:
P(w#B|s#C)=(Pra(w#B|s#C)*MRA(#C) + Pta(w#B|s#C)*MTA(#C))/(MRA(#C)+MTA(#C))

Lastly, for determinating the total amount of effort a player has invested we assume the player has a skill level equal to its gladiator rank (s#=rank). The formulae to aproximate the number of matches a player has been involved in is:

NM = CGP/P(w10|s#) * (P(w10|s#)*10 + P(w9|s#)*9 + ... + P(w1|s#)*1).
NM = Number of Matches played.
CGP = Current Gladiator points.

This way, now we can aproximate how many matches each player has been. The NM can be used to aproximate the gameplay time if we have the mid-time value for a battle. T=NM*BT (Gameplay time = Number of Matches * Battle time)

Posible Arena mechanics changes

Remove the translation from RA to TA

bla, bla, bla. (this should be done, independently of everything else) Some issues to be considered.

RA with Random teams every battle

I want to make a summary for this option:
In some topics of the User talk:Ryan Scott page some people suggested to create random teams every battle for ra. That means even if you win you randomly change your team also. I want to clarify the advantages:

• This suggestion solves the problem of synch.
• This suggestion solves leavers problem completely and not in a forcible way (like implementing punishments).
• Since every battle the team changes it depends a lot more on you rather of your partners to achieve win streaks. That way it differenciates more a skilled player from no-skilled ones.

A possible advantage:

• People might enjoy more because every battle you have to adapt to your companions. "That would be truely random and exciting" (User:Vallen Frostweaver)

A possible disadvantage:

• People might enjoy less because they don't feel integrity with any group. " I always liked the fact, that there grew some sort of relationship between the team members, while progressing to 10 wins (User:TeleTeddy)"

An evident consecuence: Obtaining winning streaks will be a lot more difficult.

Issues that people might find as unresolvable problems (all follow a logic, but all sentences are flawed):

• Since obtaining winning streaks will be a lot more difficult the title needs to start over. (There is no need to start over, because you can reflect the effort the players invested on the new system.)
• Since obtaining winning streaks will be a lot more difficult, there is no way we could advance the gladiator title. (You can advance if a new adapted system is implemeted.)
• If we change the gladiator title so you can advance on it in ra, then it will be cheapen because it will be too easy to advance it on ta. (You can assign different rewards for win streaks so two equal skilled players who prefer different arena receive a similar reward per time invested.)
• So no matter how you change gladiator title it will be incompatible for ra and ta at the same time. the only possible solution is to split the title, one for ra and one for ta. (see the next)
• Since Anet doesn't want to split the title, then there is no way you can think of implementing the random every battle for ra. (No need to split the title, if two properly adapted systems are found one for ra and one for ta.)

Conclusion: Is Random teams every battle proposal the way to go? If we can obtain an adapted new system for gladiator title, and after a week test the community express to enjoy more this total random system (this could be measured by a general polls on fansites). Then I think this should the way to go. If the community enjoys more the previous system then the gladiator points assigned by number of wins every 10 matches mechanic should be considered instead.

A note: The way to unlock TA should be updated.

Gladiator points assigned by number of wins every 10 matches

bla bla bla. bye bye leavers.

Increase the required win streaks as rank increases

bla bla bla. (interesting, but fails to reproduce a smooth progress at high ranks, completely unfair to ra)