New Way to Randomize Numbers: Opinions

[Deleted]

Well, the whole time I've used rpgm2 I've always been suspicious of they're number randomization method (hearing that it's the remainder of the # F on a map divided into a hundred, and personal experience makes it seem wierd), and recently I was playing Fire Emblem and went to the gamefaqs forum for help. Anyways, one thing I learned there was quite interesting.

Fire Emblem uses a (no one knows how the numbers are generated) of a set of numbers that seems to go on forever, and everytime it needs a random number (1-100) it draws the next number in the sequence. I thought about it, and ideally, having every number from 1-100 in the list appear one time, then having it repeat to the top at the end, will truly give the exact chances for success to the player.

Anyways, I'm wondering what you guys think of this idea. I'm probably gonna do it anyways, but I thought it was a pretty cool idea. I'll just do something like take pieces of paper out of a hat to get the order.

[Jugem]

Have you played Tetris? Did you catch on at all to my randomization method? It works like a charm to spit out "true" random numbers to decide what block should be the next one. When entering the map, I just wait until the player presses the square button before getting on with the action. Since rpgm2's random number generator depends on the amount of time spent on the map, this method throws it a curveball (since the time spent on the map will be different each time because of the initial delay always differing). The only downside, obviously, is that you need to press square everytime you enter a map. Or, I suppose you could do this once and just save a bunch of random numbers to some variables.

[Doan the Nado]

Is it the time spent outside of content scripts, or all time, total, on the map. That is, if a message window is open, is the time still counting?

As far as your idea, Will, while that will give truly even success, it will certainly not give random success. The chances of all 100 numbers appearing once when 100 are chosen is astronomical (100!/100^100), which comes to about 1 in 1 dozillion times. Yeah, that's a word I made up for million, billion, trillion... (=1, 2, 3...), dozillion (=12)! Try choosing 100 random numbers that many times, and the situation you describe will happen once.

So in my opinion, no, what you describe is not very random at all, although I will certainly agree that it is "fair". One of the points of random numbers, however, is not knowing what's coming next... if the chance of failing is 50% and you've managed to fail 10 times in a row, the chance of failing is still 50%. Your system eliminates this, and if the player has failed 10 x in a row, that means that there are 50 numbers left above 50 and just 40 below, and their chance of success is now 55%.

Anyways, just some thoughts on statistics theory that I thought I'd like to share.

[Rodak]

The only problem I had with RPGM2's randomizing was the repetition.

I posted a way to keep random numbers from repeating for at least 4 (or any other number) times. It's here somewhere...

I'll find it and link to it if anyone cares.

Good idea though.

But would it not be the same 100 numbers in the same order each time?

With no chance of repetition in the list it actually becomes less random (mathematically speaking).

You may want to make the list more like 150 numbers with repeats on "key" numbers just to give the game flare.

I dunno.

Peace.

[Deleted]

It must be the total time for both, otherwise everything in the DBS would be rigged, which they seem to not be.

What I never really liked was how waiting a half-second to input your battle commands, or the text speed's setting, and what not effect the number randomizing. From my experiences that's been a problem. I remember one time I kept trying to flee (I'd made my flee a flat 40% chance), failed about 10 times in a row because (I guess) the time alotted in between those times was keeping the random numbers the same, waited about five seconds, entered to flee and succeeded (the 11th time). They've always seemed off to me (especially if it's that formula above).

I do see how this undoes the randomness, but it improves the accuracy to being perfect.