|
Heres my tip.
There is a 'socket rate', which is a fixed value on the server. This value is likely to be very small, probably 0.05 or something.
The server calculates a random value each time you upgrade. This value could be seeded by anything (time, tickcount, previous result, #of players on server etc. It is unlikely they would use something that can be predicted to seed their random algorithm). Anyway, there is the possibilty that a random alrogithm can produce common extreme values at certain times, or after a certain number of calls to that random algorithm, but neither are predictable, particularly with our limited statistics on socket times, and biased opinion towards xx:33 where more mets are tested. The random algorithm will produce a value between 0 and 100 as an example.
Now, if our random value is between zero and our socket rate, it makes a socket. There is a 0.05% chance of this happening.
Based on mathermatical probability, if we try 2 mets, we have another 0.05% chance of making a socket. That means, 2 mets gives a 0.10% chance of a socket.
Evidently, there are 2000 0.05s in 100, meaning after you have used 2000 mets, you theoretically have a 100% chance of making a socket.
Those values are just random values I made up for demonstrative purposes, don't take them for real. I am just trying to show that socketing is based off mathermatical probability, and nothing else. The more mets you use, the more likely you are to receive a socket.
If you really wanna argue that times work, or would like a more accurate guess than my 0.05% probability, it is going to take more than 1 or 2 sockets to show that.
You need a broad range of statistical data, as many sockets as you can, including the time, accurate to milliseconds, the number of mets used, and even then, you probably won't find any trick to socketing.
There is also the factor of upgrading quality produces a much greater socket chance. Take my example of 0.05% for mets, and multiply it by 20 perhaps. a 1% chance of socket for dbs for example.
The trick to socketing is try as much as you possibly can, discarding any theory about time, co-ordinates, itemid, or anything else. Sure, these values might seed a random algorithm, but even if they do, the result is unpredictable.
|
