User Pass
Home Sign Up Contact Log In
Forum > Mod Only > RNG (Random Number Generator)
Page:
 
Deathblade
offline
Link
 
Originally posted by Bort
What are the constants through a player's career?

Height/weight/player_id/user_id

Is that it?

Ht/Wt go into speed. We know that much. player_id and user_id shouldn't go into anything unless something really odd is going on.

People started noticing speed differentials in ~season 6 or 7 when I messed with ht/wt factoring more.

Hm, I have an odd theory that just sprang into mind. I have been dealing with data loss "correction" for some time between my sim servers and the database servers. It springs up in different places, probably due to some packet loss or something, and the code sometimes has to re-try connections or updates. What if I am dealing with database corruption, or something of the like? Maybe I'm not getting the data that I think I am? Seems like weirder things might happen at times, but Perl will interpret strings as numbers without a hiccup or warning. If I'm getting garbled data for some players or something, maybe it could affect things (and cause some of those unreproduceable bugs that come up from time to time)?

Thinking outside the box here.


Could the corruption or whatever be reading some players as weighing 0 pounds, which makes their speed ridiculous?
 
Bort
Admin
offline
Link
 
Originally posted by Deathblade
Could the corruption or whatever be reading some players as weighing 0 pounds, which makes their speed ridiculous?


A garbled string of ascii crap would be interpreted as zero most of the time, yes.

Wonder if it sometimes would get zero strength or zero agility, though, too? Seems like other things would get messed up. Weirder things have happened, though. I consistently have run into a random time-to-time issue with updating a particular column in the pbp table, but no other column has the problem, and it works fine from other servers. Frickin' baffling that one.
 
Bort
Admin
offline
Link
 
Wonder if these players play on different servers?
 
Deathblade
offline
Link
 
Originally posted by Bort
Wonder if these players play on different servers?


Not sure...
http://goallineblitz.com/game/team.pl?team_id=609
http://goallineblitz.com/game/team.pl?team_id=1262

There's the teams.
 
Bort
Admin
offline
Link
 
Originally posted by Deathblade
Not sure...
http://goallineblitz.com/game/team.pl?team_id=609
http://goallineblitz.com/game/team.pl?team_id=1262

There's the teams.


Hm, definitely do. Team 2 is on server 2, while team 1 is on server 8.
 
Bort
Admin
offline
Link
 
So...if there is a database issue I would have no clue how to fix it (I have tried all kinds of stuff) other than writing in a workaround or reload data command. Of course, if I always get the same crappy data that wouldn't work.
 
Deathblade
offline
Link
 
Would it be possible for you to run a "stealth game" on each server using the exact players, or not?

To make sure that for instance, one of the servers isn't getting screwy data for one player.
Edited by Deathblade on Sep 1, 2009 16:21:10
 
Bort
Admin
offline
Link
 
Originally posted by Deathblade
Would it be possible for you to run a "stealth game" on each server using the exact players, or not?


No...it would show up in their game log, etc.
 
Bort
Admin
offline
Link
 
I could set up a script to just load the player data over and over to look for oddities.
 
Jed
offline
Link
 
Originally posted by Bort
Wonder if it sometimes would get zero strength or zero agility, though, too? Seems like other things would get messed up.


It's more common to see people getting very high Speed than Strength or Agility, which means there's more chances for that to happen to someone with Speed than Strength or Agility.

Also, it's much easier to see the effects of Speed than Strength or Agility in replays, which makes it even more likely.
 
Bort
Admin
offline
Link
 
Originally posted by Jed
It's more common to see people getting very high Speed than Strength or Agility, which means there's more chances for that to happen to someone with Speed than Strength or Agility.

Also, it's much easier to see the effects of Speed than Strength or Agility in replays, which makes it even more likely.


So, you're saying that if there were corrupt data for other attributes it wouldn't be as obvious?
 
Deathblade
offline
Link
 
I don't know

2+2 just doesn't add up to 4.

You could try a test game, and put a player with 120 speed 80 agility at 0 pounds, and see if he is ridiculously fast.
 
Deathblade
offline
Link
 
Originally posted by Bort
So, you're saying that if there were corrupt data for other attributes it wouldn't be as obvious?


Probably not.

I mean, there's a TON of RNG's going on with like pancakes and such, plus you'd have to compare it to another player and such. With speed, there's basically no other player (when running routes), and it doesn't seem to have as many, or any RNGs.

With that said, I've seen players with damn near identical builds have different results, rotating 50/50 at a position.
 
Bort
Admin
offline
Link
 
Originally posted by Deathblade
I don't know

2+2 just doesn't add up to 4.

You could try a test game, and put a player with 120 speed 80 agility at 0 pounds, and see if he is ridiculously fast.


Or I could just do the math with a calculator and tell you that would be true. Both acceleration and max velocity would be above the normal maximum allowances.
 
Jed
offline
Link
 
Originally posted by Bort
So, you're saying that if there were corrupt data for other attributes it wouldn't be as obvious?


Well, it wouldn't be as obvious to the end user. Deathblade can make a script to figure out that 2 guys are moving at different speeds. The same might be said for Agility (measure how fast it takes for someone to get up to top speed), but other than that, how would we even tell from the user interface?

Strength? Pancakes and Broken Tackles are the 2 best indicators and both have more major things involved than Speed does with top Speed.

Vision? Maybe you could measure how quickly someone reacts to a play on defense, but that would be very tough.

I don't see a single other stat which, if it dropped down to 0 for 1 play, you could see the obvious difference on the replays.
 
Page:
 


You are not logged in. Please log in if you want to post a reply.