DUNCANS'S PROFILE

It was fun ^^. One thing I would add is a credits section at the end - or did you do everything yourself?

Oh wow :S thanks for the feedback! I removed a bunch of BGM tracks to shrink the upload size - forgot that I was using that one TT...

The 8 steps for random encounters is, I agree, too often, but I felt that it would be better to get the grindy aspect of the game out of the way along the way rather than having to backtrack if you have a hard time on the boss. I guess maybe it needs some more balancing ^^;

Agreed that sideview is slow, but I like having more variety in battles - so much time in RPGs is spent in random encounters that I like them to look polished even if it slows the pace a bit.

Edit: A new version with the missing track is being uploaded now ;n; sorry

Edit 2: Updated the download once again to include an encounter rate modifier so you can decide how often to have encounters!

author=GreatRedSpirit
* The worst part of pokemon is that you can get in a random encounter without taking a single step (changing your facing direction counts as a step as far as encounters go). It is infuriating and I hate it.

^ This.

I agree, that looks like a good formula :3. Dynamic form if anyone cares :P

author=GreatRedSpirit
The average has little to do with consistency. Looking to the mode is better. There's no curve to flatten either with equal weightings. You can reduce the range to introduce consistency but you'll converge on

The average has to do with perceived consistency. By saying that there is in fact a mode, there must be a bell-curve - the values that are more represented are the average values in a random system. With 2 6-sided dice, you can view the bell-curve as a fractional representation of probability, ranging from 1/36, the odds of rolling a 2 or a 12, to 1/6, the odds of rolling a 7. The weighting on individual random rolls is indeed equal, but the increased number of rolls skews the weighting toward these mean values.

With n/2 + rand(n) + 1 you increase consistency by reducing randomness and flattening the bell-curve. By adding more dice, or more rand() values to the equation, the bell-curve spikes and consistency decreases.

author=GreatRedSpirit
With the bell curve of multiple dice you can have the consistency of a target number of steps that encounters are most likely to occur at and have a wider range for the occasional early/late encounter to keep it from getting weird.

This is actually pretty much the exact current system. Did you mean something like 4.times { val += rand(n/2) }; val += 1? The issue there is you then have M = (57+1)/2 = 29, which begins to skew the average (the mode in a random system) downward due to the 0<=n<max bound on rand().

author=Craze
...Of course, you could make it even more consistent with something like rand(n/2).

Wouldn't that change the range to 16-30 steps (rand returns n where 0<=n<max if I'm not mistaken)? Not really an average of 30 at that point :S. n/2 + rand(n) is probably fine at 15-44 steps, no? Or set up a dynamic encounter rate with something like ((n/2 + rand(n)) / ((rate+1).to_f / 100)).to_i :P