Theory about mf

[Fearox]

Theory about mf

I have this theory about which item types that can be dropped. If I do, say, a lot of pit runs and all items I see are exceptional or normal types of items (I'm not talking about magic, rare or uniques here - that's depending on mf) I restart the game. With "restart" I mean turn it off completely (exit to windows). Then I start it again and - voilá! Now I see different kinds of item types in The pit. If I still only get exceptional and normal type items I restart the game again, and so on and so on...
The thing is this little trick seems to work (just like the game had some sort of item generator which decides what can be dropped every time you start it). This method have given me lots of nice items (I used it back in 1.09 too). My friends use this method as well, and they are too satisfied with the results.

So, what do you think folks? Can this be real, or am I totally mistaking? Or maybe you already use the method?

 

[Ryshenron]

it has to be coincidence, the item a monster will drop is determined when that monster is spawned, and if ur playin on realms, the server randomnly decides that...and as u already know im sure, mf% only affects the item type(normal,magic,rare,set,unique:in that order(or backwards if you will)) the item will be when the monster actually bites the dust

 

[Knightmare]

Just like how we're destined to become either rich, poor, or scum.

 

[dop_obo]

I find that games spawn a little bit weird. If I create a mf game and high treasure class items drop in the pit, then I'll do the whole lvl 85 run. If no TC 85(?) items drop in the pit, I'll create a new game right away, and try again. It seems that if one area drops high, then the other areas will too, and vice versa, of course.

 

[Fearox]

it has to be coincidence, the item a monster will drop is determined when that monster is spawned

Yes, but somewhere it has to be decided what the monsters can drop from time to time (like with a dice). Sometimes you only see exceptional items in certain areas and sometimes you see a lot of elites. It seems the game has decided before the monsters even decide of spawning which kind of different items they can drop. And it seems the game has decided which treasure classes that can be dropped.
This can be totally wrong, but like someone else here wrote: I you find elite items in one area you often find them in other areas too. I believe this is true even for crap items.
This has nothing to do with the server. It is your game that spawns monsters and items and maps. The server saves the files.

 

[Qroft]

This has nothing to do with the server. It is your game that spawns monsters and items and maps. The server saves the files.

If that were true it would be a hacker field day.

...

Maybe you are on to something there.

Seriously though, items are generated by the server. I'm not sure about monsters and maps, but the items for sure.

 

[Fearox]

Seriously though, items are generated by the server. I'm not sure about monsters and maps, but the items for sure.

If the items are generated by the server my theory is false. How do you know that the items are generated that way?
On a second thought... maybe it's the log off that changes the sort of items that are dropped?
I'm not one of those "read game code like a newspaper"-guys. In fact I don't even know how to read the game code. Some help from someone who's into making mods and stuff would be appreciated on this one.

 

[Qroft]

Well I admit I don't have any intimate knowledge of the code, it's just kinda an assumption. If our computers generated items how long do you think it would be before the hackers came out with some sort of "unique hack" that would spawn a WF or SOJ or whatever every time you killed something?

So I'm pretty confident that is the case, but I don't really know for sure.

 

[hoopyfrood]

If the items are generated by the server my theory is false.

Not quite true - IF (<--- and that's a big if...) bliz wanted to, they could change the item generation code from server to server. Recall how people were tracking DC by which server's IP they were on - what applies to one of bliz's servers doesn't *necessarily* apply to all. Were that true, though, once you get a "good drop" game, you'd get a "good drop" game each and every time until you disconnected from the server.

That said, the probabilities of drops involved here lend themselves to one person not being able to obtain a statistically significant number of samples from which to draw a valid conclusion. And owing to human nature's need to attempt to find patterns in the completely random (astrology, numerology, etc), what develops is pseudoscience and voodoo. On top of that, I seriously doubt Blizzard would bother maintaining different code for any of their servers.

 

[festers50]

Luck is very much a factor in whatever drops, switching servors could be as effective as having a one nighter with an Island girl to change your luck.

While scientificly unproven, it does make "you" happy, which is all you need sometimes for a change of luck

 

[Shadow_247]

Item generation takes place when a monster is killed. This can of course be changed by modders and the like, but in those cases, you would not be able to connect to BNET, as everytime you connect it verifies your application version.

Everytime a monster is killed the game goes through a very long and lengthy process. The Item Generation Tutorial here on this site explains just what the game goes through to generate each item. The author(s) of the article have looked through the game code and the article discusses what they have arrived at. I realize the article is from 1.09, but much of the same discussion still speaks true.

The item generation process is mainly formulae. The formulae include things like calculating no drop, and calculating chance for certain quality items. Each time the game uses one of these formulae it makes a new roll, giving a new result each time. Now, the game goes through the formulae step by step. The game calculates nodrop - if the game decides to allow a drop it moves on to other calculations - such as rolling TCs or calculating which item quality your item will be of. Monster level and qLvl are other factors in the equation, meaning each drop should be random. I'm sure one of the other extremely in-depth individuals could explain it better, but basically, the point i'm trying to make is that your method is more than likely revolving around coincidence, and a hoop of luck. Now, if it works for you, then by all means do it - I myself keep my lucky rabbit foot ontop of my monitor whenever I MF.

 

[Vizier]

Thanks for summing that up Shadow. The only way the server-side code could start doing something non-random is if the seed used to create randomness was somehow not getting randomized itself somehow. Certain random number generators are not as random as good/complex ones but the better ones cost CPU time and may be skipped for performance reasons. I'm not sure how well Blizzard's code performs on its many servers, but just stop and think for a second how many times per second those caluclations that Shadow just mentioned are being run for each drop! RNG's are based on some sort of seed value that itself should be randomized. If it's not, then patterns could start showing up in the results and then it's not exactly random anymore. In my expereince in the Pit as the original author mentioned, I'm had a few nice things drop really close in time to each other and thought "This is going to be a good game!" when nothing but crap dropped the rest of the game. However, I've seen the reverse as well, so it makes me feel good that randomness is still alive and working fine. If it wasn't for those randomly sweet drops once in a great while, many would all be giving up on MFing areas like the Pit. Long live randomiety! (is that a new word?)

 

[Unholy_Carnage]

The reason you are seeing such thing is because monster drops are a scale, not a range. Meaning, it drops from tc nothing to the max TC they can drop, instead of dropping from TC50-TC85, for example( like they should, im tired of seeing normal sashes in hell). This allows them to drop MANY things, which is why sometimes you see only exceptionals and normals - because theyres so many that can be dropped.

I have never ever looked at the game code, so this is my opinon, not fact.

 

[Shadow_247]

The reason you are seeing such thing is because monster drops are a scale, not a range. Meaning, it drops from tc nothing to the max TC they can drop, instead of dropping from TC50-TC85, for example( like they should, im tired of seeing normal sashes in hell). This allows them to drop MANY things, which is why sometimes you see only exceptionals and normals - because theyres so many that can be dropped.

I have never ever looked at the game code, so this is my opinon, not fact.

You are quite correct in saying that. Level 85 areas, for example, roll down the TCs. First it makes an attempt at the highest TC, armor87 and weapon87, if it fails, it then makes it's way down to armor84 and weapon84, if that fails, then armor81 and weapon81. This continues until a succesful roll is made on a TC. Now that is very basic, there is actually alot more to it than just that, but for the sake of simplicity, let's leave it at that.

 

[Pent]

omg this theory is true!
I just mf'd windforce, tal rasha armor, and tyraels armor in one pit run!!!!!!!



Lets play a little game, it's called, see if you can spot the "I've been playing diablo2 for so long I'm started to hallucinate and make up crazy stories and thoerys" persons.

 

[Usufruct]

I only mf naked because I only ever find good stuff when I'm naked.

*scratches chin thoughtfully*

Maybe I only find things when I'm naked because I always mf naked... hmmm...

 

[FinalGodSlayer]

I kinda noticed this recently, and mentioned it on another thread, then they linked to this thread here. I agree with some of the earlier comments, and think that some replies have mis-interpreted them slightly. I'll be fairly specific about what I've noticed:

I'm playing Single Player 1.10, and I'm running the same map seed each time (NM Andy runs) I start the Diablo II application and then do about 10-20 Andy runs until my stash is full of items I want to mule. During this time I end the games and start new ones, but I do not exit the Diablo II application.

Then I exit the Diablo II application (exit to windows), and run ATMA to mule my items. I then exit ATMA, and start the Diablo II application and then do about 10-20 more Andy runs, and so on.

I've been doing this every day for a few months (looking for an elusive SOJ), which involves about 40-60 runs a day, and so have a very good "feeling" for how the game is dropping for this instance. I used to say to people that the game had "good days and bad days", meaning that some days my luck was in, and I got a lot of good drops, and some days it wasn't. Luck is just that - luck - after all.

However what I have noticed is that that the game isn't having "good days and bad days" at all. It is having "good application runs and bad application runs", which is quite different.

Often I'll think "The last 20 runs were pretty naff", but then when I restart the application (not just start a new game), I will get 20ish much better runs.

Now the averages of luck should dictate that there should be little in the way of cycle for this - you should get average luck, with peaks and troughs now and again. I think (and I have no factual proof) that (in Single Player) there is an additional Seed used in the drop system, generated once when the Diablo II application starts, which influences the drops. It doesn't really matter what the seed is (it could just be the startup clock time). If it's there then we should be able to detect it, and thus know when we've got a 'bad' one or a 'good' one...

So how to detect it? I need a method that's quick, 'cos Andy runs only take 1 minute + stash time, and as I only play for 1 hour a day at lunch, I don't want to slow down my kill rate.

I was thinking of "scoring" Andy's drops (The drop bug is affecting my game).

1 point for a rare item,
2 points for a set item,
5 points for a unique item.
nothing for blues or failures

This gives a score between 0 and 30 for a given run. I can have a pad of paper by my laptop and note that score down in a couple of seconds based upon the colours of the items she drops on the floor.

For a set of about 20 runs within a single instance of the Diablo II application running, I would get a score of 0-600, and by the laws of averages these *should* tend towards a common average over time.

If however there is indeed a seed involved, we should see individual sets of about 20 runs fluctuating wildly from the average.

I'd be interested in comments on this idea (including the suggestion that I'm actually insane) before I spend time doing it?

Maybe someone with better maths than me could devise an alternative scoring system that would give clearer results? It has to be quick though. I need to be able to scribble this down fast without using a calculator!