cRPG
cRPG => General Discussion => Topic started by: Banok on January 09, 2011, 04:10:27 am
-
EDIT: just saw this post from chadz so my maths must be a bit off
There was a minor update an hour ago on EU and NA, reducing the break chance for the winning team a bit.
4% break per min, minimum earned 50 gold per min, 5% repair cost of broken items.
okay so...
break 100/4 = 25
repair 100/5 = 20
(50goldx25)x20= 25000
25000 gold the minimum cost you can use without losing money assuming you lose 100% of the time and only get 50 per min.
Asumming you win half the time you multiplier would be 1.9 on average (thanks pukudo), which means 47.5k gear. I am talking in AVERAGES, sometimes you will have bad luck always have 20k or so spare gold for rainy days.
47500 is too much for an just average player imo. I think the system is too random, and often totally dependant on how your team plays not how you play. It would be nice if even if you lost you got more gold for doing well.
Amount of items matter?
so a 10000 gold item costs (10000/25)/20= 20 gold per minute to upkeep.
1x 25000 gold item is 50 gold per min
25 x 1000 gold item would be still 50 gold per min
so no number of items I dont think matters, just total cost.
-
whoops the quote and modify buttons are in the wrong places!
-
The second part of your math is wrong but i'm not going over that again, just so you know, the limit of sustainable equipment is in place for ALL players, there are no average, low or high players. All can afford the same. That's another problem with upkeep, removing character differentiation. You might think 37500 is too much, but some characters use more items (take a 1hsword/shield with trowing weapons, a average template for example), and once you make good setups for what the real upkeep limit is (48500) everyone will just use that. In native mode players will look more different.
-
try to back up claims with explanations, your post doesn't really make any sense. upkeep makes there much more differentiation since before everyone just went plate. how is that second maths part wrong? real upkeep is 48500, for what?
-
try to back up claims with explanations, your post doesn't really make any sense. upkeep makes there much more differentiation since before everyone just went plate. how is that second maths part wrong? real upkeep is 48500, for what?
everyone where? if only in my clan from 20 ppl only 3 where using plate LOL
-
Assuming a 50% chance to win.
The probably of your multiplier at any one point is as follows!
x1:50%
x2:25%
x3:12.5%
x4:6.25%
x5:6.25% (this is because as a cap it will loop back onto its own multiplier)
Therefore:
The Average multiplier is 1.9375
Watch this space for some pretty graphs.
-
With the recent patch (see the 0.201 announcement thread) things have changed a bit, losing teams get penalized a little and winning ones have a smaller chance, i have not seen the numbers behind those yet. From a few rounds played today things seemed to stay the same overall or maybe upkeep has increased a bit, now winning or losing matters more, i had even 6 items break on one round i lost. We need to correct the math to calculate the new sustainable equip value.
-
this "everyone has the same chance" bullshit is really getting on my nerves...
Have you guys played the game lately? For some reason it is usually 1 or 2 kills that decide the individual rounds and that shows how much impact a single player can have with a few kills.
Meaning that you can VERY MUCH influence the chance to win a particular round and it will have a big impact on your maximum upkeep costs if you play well , hardly die and kill a lot.
Not to mention the fact that winning will also reduce the chance of stuff breaking slightly.
I would claim that any average player can easily upkeep 30k of equip.
-
I've been unable to apply any Probability distribution that makes sense so far..
The standard deviation works out to be 1.183
Which would mean that 68.2% of people would have an average multiplier of 0.7545 -> 3.1205
Under the sustainable 25000 gold per x1 multiplier model. Which is a logical one.
That would mean 16% of people would be able to sustain over 78000 gold..
The issue is I think that you can't apply Standard deviation to this example.. as if projects people out of the Limits of the range 1->5
However with the Avg calculated. 1.94 * 25000 the Average sustainable amount seems to be 48500. This is really to high!
----edit---
I would claim that any average player can easily upkeep 30k of equip.
This is a mathematical discussion.
-
+1 48,500 is too high.
-
----edit---This is a mathematical discussion.
No its not, since you lack data to make it one. You are ignoring factors that are part of the equasion making it absolutly pointless.
It is also not a "mathematical discussion", because its fifth grade maths you are talking about here and calling it a mathematical discussion is really a bit much for that.
edit: just so noone cries about me not backing up my point:
You can't even start with 50g/min because it doesn't work out that way. The first gold "tick" is about 20 to 30 seconds into the round and the last tick is hard to pin down, because of the randomness of fights ending. It might be less then a minute overall or more --> we would need statistical data for that. If you now think this is an unimportant point then look closer and calculate how many gold ticks per round you will get and how long rounds take on average, especially taking the fact into the equasion that you have servers with fewer players, thus making the individual rounds shorter or more random. Then take siege mode.. where you will have different chances for defenders and attackers depending on the map you play on. Not to mention that you have this same problem with the battle maps. Then you can start on thinking about autobalance and joining ... people who only play for half an hour at a time have a much better chance to enter crap teams since when you initially join a server you will end up on the team that lost the last round in most cases ( also would need more data to really look at that factor)...... and so on... there are so many things manipulating your "mathematical diskussion" wich you just ignore, so your calculation is pointless.
The only thing that will solve this problem is not a mathematical equasion but collecting data and seeing how it will work out once the extremly rich and older players have used up all their horded gold. There are players with so much gold they can probably play a month as full tincan without running out of money.
If you just started playing the whole thing feels very balanced and not "to easy".
And let us also not forget that since we are talking about balance here it is not enough to calculate the possible upkeep. Things like people buying and selling equipment also have to be considered if you want to make a proper point. Loosing 1/3 of the value of any item bought when you sell it also counts a lot, not to mention people collecting different sets of gear, wich also takes gold form the "economy".
Sit back and wait and look at how the balance works out over time and stop trying to create arguments with pseudofancy maths.
-
Ramble Ramble Ramble.
Sit back and wait and look at how the balance works out over time and stop trying to create arguments with pseudofancy maths.
You don't need to plug in every absolute condition and please refer to my first post where I have assumed a 50% chance of victory.
First of all the chance of an item breaking from what has been told. is directly prepositional to the amount of resource ticks so the length of time becomes irrelevant!
Its ment to be a 4% chance per tick.
If you want to plug in ABSOLUTELY EVERY CONDITION then you will find that true randomness doesn't even EXIST EVER.
As autobalance is designed to work on a 50/50 chance it will be safe enough to asume it does its job well..
As I said Amusing a 50% of victory the average exp modifier is 1.94.. That is a MATHEMATICAL STATISTIC.
-
You are missing the point mate. You are trying to make something very comlicated into a simple equasion and that doesn't work.
As i said.. sit back and wait how it works out =)
I have not seen the code and the original calculations that determine the true values and neither have you afaik. We just do not have the complete picture unless chadz really posts all the formulas , wich he has not and probably wont.
-
+1 48,500 is too high.
No, go play native pls kthxbye
-
Assuming a 50% chance to win.
The probably of your multiplier at any one point is as follows!
x1:50%
x2:25%
x3:12.5%
x4:6.25%
x5:6.25% (this is because as a cap it will loop back onto its own multiplier)
Therefore:
The Average multiplier is 1.9375
Watch this space for some pretty graphs.
Awesome, I was thinking about trying to work this out myself but wasn't sure how. I dont understand what your trying to achieve with SD in your other post tho.
even though in theory anyone should be able to upkeep 47.5k with ease [(1.9x20)x25], it definitely does feel alot harder than that ingame. I was using <20k gear for a long time about 1 million exp's worth and only recently started to make any profit. and I see 5 items breaking at once fairly regulary, which should be very unlikely. so I have a feeling the numbers we have been given are wrong.
-
You are missing the point mate. You are trying to make something very comlicated into a simple equasion and that doesn't work.
As i said.. sit back and wait how it works out =)
I have not seen the code and the original calculations that determine the true values and neither have you afaik. We just do not have the complete picture unless chadz really posts all the formulas , wich he has not and probably wont.
Hey newtons laws are basic. and we design complex systems around them. Probability isn't and never was an exact science but rather shows you and overall projection when plugged in with lots of data.
Awesome, I was thinking about trying to work this out myself but wasn't sure how. I dont understand what your trying to achieve with SD in your other post tho.
even though in theory anyone should be able to upkeep 47.5k with ease [(1.9x20)x25], it definitely does feel alot harder than that ingame. I was using <20k gear for a long time about 1 million exp's worth and only recently started to make any profit. and I see 5 items breaking at once fairly regulary, which should be very unlikely. so I have a feeling the numbers we have been given are wrong.
Standard deviation is used to predict the variance of where people will sit below and above the mean. So you would be able to predict that the top 10% of players (do the random nature of luck and skill) would be able to maintain x Gold of equipment. ... It ahh just doesn't work in this example at all...
Now but a bigger issue? I'm not sure about this 4% thing? is this a 4% chance that a Random item will break? Because this could be abused by spawning with low end items then dropping them! Or is it rolled for every item?
visitors can't see pics , please register or login
The in right side of the graph is ment to have the same area as the left side of the graph.
Where as the x axis shows the Avg. Multiplier, The y Axis is ment to be an indication of the probablity. The area below the line shows (or is ment to if made properly) the amount of people..
-
I still believe you are working with the wrong numbers. It does seem that we lack the exact formulas still. I am quite familiar with mathematical theories and your basic ideas are alright , no worries about that. You just lack the correct formulas and not just the stuff i meantioned that you don't take into account.
I have played quite a lot and the numbers you end up on just don't feel right.
We will have to sit back and watch how it plays out or get the correct formulas from chadz or there is no point in calculating anything.
4% chance on an item breaking could mean that every item gets a 4% roll each time you hit the 1 minute mark OR 4% added to the chance of each item breaking (for example 20% for a 5 minute round) OR 4% that something breaks each minute displayed at the end OR 4% added to something breaking (wich is unlikely since so many different things seem to break in some rounds)
I think (and i just assume since i lack the data to be sure) that it is 4% added to each items break chance for every minute and that is calculated at the end of the round. Wich would mean that in a 5 minute round you have a 20% chance for each item to break and need repairs.. wich still seems to low a chance considering how many repairs are needed in some cases. But then again we perceive the negative a lot stronger then the positive in memory so it might just be the correct formula for item repairs.
4% chance that a random item breaks is extremly unlikely, since having to repair 5 or more items is not that rare and rounds don't last long enough to make that likely.
oh yea.. and you are still missing my point i think ....
-
problem with upkeep is that upkeep is stable of course
if you do 1000 round
not 50 or 100 or 200.
becouse on 50 or 200, you can have even a 150 lose streak
-
What I can say is that the changes from. .200 -> 0.201.
Amount to 1/3rd of the cost before hand.(theoretical)
From a factor of .006 to .002 By the cost.
I'm not supporting or condemning the first formulas.
But people shouldn't consider them to be absolute fact.
They don't seem that wrong to me. And it would be hard to tell anyway unless you compare your data to other peoples.
As for the formula... It's breakable.. People will just need to do some experimentation.
Get Person A to wear only 1 item.
and Person B to wear all of Them.
Record: Number of resource events, and number of items breaking.
Comparing data from Person A and person B over say 20-30 rounds. Would make it possible to able deduce if it rolls based on the amount of items.
Then using the number of events E and the number of items breaking included 0 events.(N)
It should be possible to work out a relationship.
However I'm unable to collect such data so I'll leave that to someone else...
The First Post Account for a scenario where all the items are rolled for individually.
I donno I'd give it a 70% chance.. Its hard to say because I have a problem with decay working at the moment.
-
The thing is, people don't do averages. They play a 10 or 20 round session and then look at the results.
I've had sessions where I wore 50k gear and almost nothing every broke, netting me a boost of gold even while losing.
I've also had sessions where I wore a more conservative 20k gearset and I lost a few thousand gold.
If everybody were to store and average their numbers, they'd probably be a lot less shocked and annoyed by the repair system.
Maybe chadz could implement a small statistic window in the crpg website that tracks this stuff so people don't have to come up with crazy statements like "I lost all my gold and sold all my plate and I can't even afford a pithfork anymore!".
-
There is no need to track this stuff. I think he should just post the formulas show the math. Everyone can then agree that it works. Then anyone who comes here whinging can be shunned off.
What is his current statement on the Average Gear Score?
Because given that he changed it from.
20% cost at 3% chance.
to
5% cost at 4% chance.
Thats a rather large factor change.
-
There is no need to track this stuff. I think he should just post the formulas show the math. Everyone can then agree that it works. Then anyone who comes here whinging can be shunned off.
What is his current statement on the Average Gear Score?
Because given that he changed it from.
20% cost at 3% chance.
to
5% cost at 4% chance.
Thats a rather large factor change.
You're missing that the 20/3 was for losing teams only while the 5/4 was for both teams, regardless of winning or losing and yesterday it got changed again.
-
factor is that you cant base upkeep on random shit factor
becouse random stuff become "trustable" only on big numbers 1000+
for a normal guy who play lets say 40 round (8 maps) /day can happen to upkeep 3-4 items for round on 30 losing round streak/40, and sink all his money, of course if he will play another 10 day 40 round day(400), he will be get balanced , (maybe) or he can get unlucky, give up the game for frustration and never play the other 600 round that will make him make money.
-
You're missing that the 20/3 was for losing teams only while the 5/4 was for both teams, regardless of winning or losing and yesterday it got changed again.
Oh yea good point!
-
48000 isn't too high if you're a cav. But since not every map is cavalry map and sometimes on the cav friendly map you want to play as infantry I'll say that 48k gives everyone an equal chance. You can be black armor knight if you want but without a horse and not every round (change to mail sometimes, if won't hurt you much if you're skilled a bit). You can ride a good horse in chanmail everytime or you can ride a crappy horse in plate. You can even ride an elephant in black armor but not every round.
I think that chadz balanced well this upkeep thingy, I personally wouldn't change anything at all.
-
I agree and said in the OP that the system is too random. although its not a major problem just get some spare cash.
@pukudo yeah I get SD but personally not that interested in predicting what % of player can earn what, just like to know the minimum you can upkeep to shut up whiners and so I can avoid losing gold on average. its assumed that better players will be able to upkeep more than the average.
-
Now i am not very good at math now, its been almost 10 years since i finished college. But i came to the similar numbers, even though i made calculations before i even opened this thread and used a different "method"...
Here is my, practical example of the math theory:
I bought my gear before i could wear it ( used most of my skillpoints on AGI after the patch ), so i saved some gold, and could afford to lose it even with the worst odds. So... My total gear cost was around 37000, and after 20+ hours in-game, i have gained 5.000 gold. Basicaly its breaking even for me.
37k is far from 48k, but i guess, if one has 200-300k to "gamble" with - it will last for a good amount of time... If you only have 10.000 left when you buy your top 48k gear - you might get unlucky, and be naked in no time.
48.000 is not that much. I have a heirloomed 1h ( 6.5k ), a heavy board shield, heraldic chainmail, chainmail boots, chainmail gloves and a plated 50+ armor helmet ( + a great maul as my back-up weapon ). As you can see - its a very medium armor set. Both of my weapons and shield - are pricey for what they are.
I would have to drop something, if i want to get a good horse. Or a siege crossbow with high end bolts... Etc.
48k is actually not that stupid. I still hate the patch, but it is starting to make sense...