Statistically speaking,
64 had eight starters and four unlockables. Therefore 50% of the roster was unlockable.
Melee had fourteen starters and eleven unlockables. 44% of the roster was unlockable.
Now let's count ZSS, Sheik, and PT's transformations as roster spots (These characters took just as much time to program as any others). Let's also assume that every character on the Dojo with the exclusion of Sonic and Snake is a starter. Finally, let's assume that no more starters exist. This puts the starting roster at 23, and the total minimum roster at 25.
Let's put the maximum roster size at 46 (given that it is unlikely that there will be more unlockables then starters).
3
So far, we have established:
Mean empirical proportion of unlockables to starters=47% ([50+44]/2)
Standard Deviation=4.24 characters (sqrt{[50-47]^2+[44-47]^2})
Minimum Roster Size=25 (Currently revealed)
Maximum Roster Size=46 (Double current assumed starter size)
So, when we apply this to a normal curve we see that the most likely final roster size is 34 (25*1.47) with a 68% chance that the final size will be between 30 and 38.
Now let's take into consideration the fact that Brawl has been in development for roughly 30 months. The development time for Melee was around 10 months.
Also, we must take into consideration that the Brawl team is much larger then the Melee team, with these facts, quantitative and qualitative alike, let's put the likely final roster three standard deviations above the mean.
So lets apply this to a normal curve.
25 --------- 29.76 ---------- 34 ---------- 38.24 --------- 42.48 ---------- 46.72
(minimum) (-1 stdev) (mean) (+1 stdev) (+2 stdev) (+3 stdev)
Therefore, I would postulate that based on the time that Brawl has been in development, plus the sheer amount of staff that Sakurai has on hand, that the final roster will be 47-48 characters.
Keep in mind that since I only had two samples available, I couldn't construct a suitable confidence interval. A census of 2 is hardly enough.
64 had eight starters and four unlockables. Therefore 50% of the roster was unlockable.
Melee had fourteen starters and eleven unlockables. 44% of the roster was unlockable.
Now let's count ZSS, Sheik, and PT's transformations as roster spots (These characters took just as much time to program as any others). Let's also assume that every character on the Dojo with the exclusion of Sonic and Snake is a starter. Finally, let's assume that no more starters exist. This puts the starting roster at 23, and the total minimum roster at 25.
Let's put the maximum roster size at 46 (given that it is unlikely that there will be more unlockables then starters).
3
So far, we have established:
Mean empirical proportion of unlockables to starters=47% ([50+44]/2)
Standard Deviation=4.24 characters (sqrt{[50-47]^2+[44-47]^2})
Minimum Roster Size=25 (Currently revealed)
Maximum Roster Size=46 (Double current assumed starter size)
So, when we apply this to a normal curve we see that the most likely final roster size is 34 (25*1.47) with a 68% chance that the final size will be between 30 and 38.
Now let's take into consideration the fact that Brawl has been in development for roughly 30 months. The development time for Melee was around 10 months.
Also, we must take into consideration that the Brawl team is much larger then the Melee team, with these facts, quantitative and qualitative alike, let's put the likely final roster three standard deviations above the mean.
So lets apply this to a normal curve.
25 --------- 29.76 ---------- 34 ---------- 38.24 --------- 42.48 ---------- 46.72
(minimum) (-1 stdev) (mean) (+1 stdev) (+2 stdev) (+3 stdev)
Therefore, I would postulate that based on the time that Brawl has been in development, plus the sheer amount of staff that Sakurai has on hand, that the final roster will be 47-48 characters.
Keep in mind that since I only had two samples available, I couldn't construct a suitable confidence interval. A census of 2 is hardly enough.
