blocks are the enemy

[imagination]

im woopin this guy zarrack24 when all the sudden everyblock gos his way... im taking about every peice side blocking ... i was surprised the cleric didnt block. it got down to a knight a 50 vrs my knight at 8 after he got like 7 blocks in a row... why cant the blocks get more evened up lol so unfair i lost like 20 some stats SOOOOOOO UNFAIR!!!

[Inuyasha2102001]

i think blocks should just be token off i was facing stick boy i hade like 5 to his 2 and he won because ever hit i did he blocked i was so pi$$ed.:mad:

[DeadFishGuy]

I take it you define "unfair" as "didn't go my way".

You won't be taken seriously until you make a genuine complaint about winning due to blocking.

[Duffman]

There is no genuine complaint when it comes to blocking. Ther is nothing wrong with the blocking.

Statistically, blocking is possible. Why do people complain when something that is possible actaully occurs. Improbable is not a cause for complaint. When units with 0% blocking begin to deflect your attacks *then* you may squeal your head off.

[DeadFishGuy]

I agree.

People seem to only be able to grasp certainty. They can only see things as guarranteed to happen, or never ever going to happen.

[Walrus]

but there are certain degrees of probability you take to be acceptable. you are willing to hit a witch from the front, because it is very unlikely to block it. you are willing to let the scout shoot your knight from the front, because you somewhat assume he will block it.


obviously, you cannot be certain, thats the whole point of probability. but when does this begin to seem like more than just improbable? well...to answer this question i have conducted a hypothesis test of a witch blocking from the front.

H0 (original hypothesis) : p= 0.2 (20% chance of blocking)
H1 (my hypothesis) : p > 0.2 (it is higher than 20%)

The test is conducted at a confidence level of 5% (i doubt youll understand what that means, but it is used later in the formulae)

So, i play my games as usual, and take note of all the blocks and failed blocks (from the front) witches do in my games, the results are as follows:

F,F,B,F,B,F,F,F,B,B,F,B,F,F,B,F,B,F,B

B = block
F = failure

Therefore the results are:

Out of 20 attacks, the witch blocked 8 and failed 12.

In my AS level formula booklet, the chance of that happening is : 0.0100

or 1%


this is where my knowledge of statistics 1 mathematics begins to fade, but im pretty sure that since the result is less than my confidence level of 5%, my hypothesis is correct and it is past improbable that such a series of results should occur.

[Duffman]

1%.... Whoa thats huge. Seriously. Think about it. How many games are played everyday?... I did the maths on how many games are played and it's a frikkin lot and you are talking about something have 1/100 chance of happening. FOr the dummies that happens every 100 games. in peak times there are over 100 games being played at the same time.

But really you have just proved my point. 1% chance is not impossible. Improbable yes. But with the amount of times you hit something with a chance of blocking per game and the amount of games played each day you are bound to get people having something crazy happening. I for one know the opponent is bound to get at LEAST one unlikely block to their favour and i will get one too. It's because people voice their stories of bad blocking that everyone gets the impression there is a problem.

People Move along... there is nothing wrong here. You all have perception disorders.

[Walrus]

well you see, the whole point of a hypothesis test is that is draws the line between merely improbable, and downright stupid.

let me emphasise that, the exact point of a hypothesis test is to draw the line. my hypothesis test showed that from my results, that it is more than merely improbable that my witch should logically block so much.

[Duffman]

Fine don't call it improbable. It doesn't matter. What matters is it isn't impossible. And if it isn't impossible then it CAN happen.

If something the CAN happen happens there *isn't* a problem with the system.

If you want to debate the actual use of having blocks in the game then you can all you want. But it won't make a lick of difference. IMO the enjoyment of the game comes from highs and lows and the thrill of not knowing the outcome to everything. I love blocks. they keep you on your toes. Makes you have to account for more things when you battle. You may lose because of them sometimes but big whoop.

[Walrus]

thats not what i said at all.


obviously everything that happens on TAO *can* happen. how would it otherwise?

thats like me saying "i can type on a keyboard"

the point is when it becomes TOO improbable which blocks happen, that it begins to make you doubt whether the percentages are correct.

your argument assumes that the current percentages are correct, if they were found to be incorrect then all that has happened, and what i said at the start, would be irrelevent.

[Duffman]

I'm sorry you seem to also have been struck with this nasty perception disorder thats going around.

Really just because because something that has a 1% chance happens doesn't mean the system is wrong. If something with a 0% chance happens then there is something wrong.

I can't see how you don't understand something that has 1% chance is very likely to occur with the amount of games played. Just because that exact unlikely combination of blocks occured to you once doesn't mean that any of the next 99 games after that have the same blocking pattern.

Maybe i am under the assumption that the program is correct but IMO blocking percent isn't that hard to code and i see no reason for them to purposely botch it up.

Also how is it like you saying you can type on the keyboard? I don't think that has anything to do with percentages...

[Walrus]

yet by saying how many hundreds of games are going on all the time, you screw up your own argument, as it shows that games with stupid blocking are happening even more.

the 1% is the exact chance that in a game where there are 20 attacks on a witch, she will block 8 of them. (side attacks not included). in theory, she *should* block 4 of them. if, based on the results i took, you look at the chance that she would block exactly 4 attacks out of 20, you will probably find that to be around 10%. i did it roughly and got 9.78%.

if in less than 1 in 10 games, a witch is blocking the way she *should* then there is almost definitely *something* thats going wrong.

it isnt about perception, its about where you draw the line of improbable and unfeasbile.

[xyxaxyz2]

Exactly.

There is such a thing as statistical impossiblity.

It is more likely that something is wrong with the system than that things just happen to work this way.

[imagination]

grrrr my luck sux lol oh well ill probably get the blocks to be even next game

[Frank]

Walrus- it appears our friend Duffman has had some shortcomings in the math department, don't hold it against him ;). As for Duffman, if you're still confused all he's saying is that either a) the stats are listed wrong or b) seed made a little error when coding for blocks. Not that the game is out to get people or that it's rigged, like friends of mine with ADD get pissed off at my Xbox when I pwn them in Halo.

[Clark Gable]

I think Walrus makes a good point. Of course a witch shouldn't block 8 out of 20, that makes the percentages off. But I think what else it depends on, is what is the 20% based on? If Walrus played 5 more games, what would the percentages be then? Would it even out towards 20%? Maybe I'm not making any sense, but what I'm thinking is they should just throw out the percentages all together. Either have them all block the same, or just don't tell us how much they block, because then there is nothing people can get pissed about. I've had knights miss frontal blocks 3 times in a row, and witches block from the side in the same game. The percentages tell me that isn't feasibly possible. Well, this is my first post, and will forever be, my first post.

[DeadFishGuy]

The mistake most people make is thinking that witches are supposed to block 20% of a single game. That's not what it's supposed to be. It's 20% of all the games that everyone plays. If you recorded all the results of every frontal blockable attack on witches in ever game ever I think you'd find it comes to about 20%.

[Clark Gable]

I still like my idea of not telling us any percentages of any of the units, :) .

[DeadFishGuy]

That won't work because we already know them.

[Walrus]

obviously, if you look at more games overall you get different results. and perhaphs there are some games where a witch would block nothing outta 20, to "even out" for my game.

but again, its a question of where to draw the line. look at the game over a 50 year period and chances are (even if by pure fluke) the two wrongs will be making a right.


nonetheless, i for one would rather a witch blocked 4/20 consistently than 8/20 then 0/20. imo the two wrongs dont make a right. in the long term, any block above 4, and any below 4, will work against each other and cancel each other out.

and i really cant be bothered with taking down masses and masses of results just to prove a point that cannot be proven, since we are talking about probability.

[MorbidIllusion]

I've always felt that luck was to big a factor in a STRATEGY game. When you think about it, no matter what setup you chose, no matter what moves you make. You have no control over whether a unit will block or not, with the exception of hitting from the back and hitting units without blocking. Therefor not making it much of a strategy game. (Though I understand the ultimate object is to get behind your oppenenet.)

When playing chess (which I think this ressembles) I don't think you'll ever here the phrase "Call it in the air" used. Do you? When playing chess you must think carefully about each move you make much like in TAO. The only thing I don't like about TAO is that you're well devised plan is easily foiled by chance.

Here's an example, I had a cleric, a LW, a Knight(50), and a Scout(40). My oppenent had the same thing. We had assembled around our respective LWs. I made the choice to attack first but however attack his scout head on from across the board (my only choice). A hit. He caught on and attacked mine. A Block. To make a long story short, my scout blocked 3 times, his didn't three times. He then surrendered knowing it was only a matter of time before I widdled down his knight and Cleric. Not very fair is it?

I feel that blocking should be a little more logically based thus giving knights a 100% chance to block in the front (he's a tank for crying out loud). I also feel that pyros and DMWs should not be able to catch arrows, and deflect steel swords (which btw should cleave their sticks in half). However prehaps a small damage reduction can be done. And with amored units, an armor deduction as well? That steel shield don't last forever.


Tada - My two cents and a nickle

[Walrus]

i think any blocking below around 30% should really just be taken out. it doesnt help the game at all. over 30% and you begin to take into account that the enemy may block it, below that, you assume it will hit. this randomness factor just detracts from the strategic elements of the game.

[MorbidIllusion]

How bout 33, b/c pyros block everything

[Duffman]

Originally posted by Walrus
obviously, if you look at more games overall you get different results. and perhaphs there are some games where a witch would block nothing outta 20, to "even out" for my game.

but again, its a question of where to draw the line. look at the game over a 50 year period and chances are (even if by pure fluke) the two wrongs will be making a right.


nonetheless, i for one would rather a witch blocked 4/20 consistently than 8/20 then 0/20. imo the two wrongs dont make a right. in the long term, any block above 4, and any below 4, will work against each other and cancel each other out.

and i really cant be bothered with taking down masses and masses of results just to prove a point that cannot be proven, since we are talking about probability.

So you are saying that there is nothing wrong with the overal game it is just the games you play that are whacked...

What you are suggesting is recording the individual blocking history of each unit, so say a witch has blocked once the next 4 times it will get hit... I'd like to say that could work but it wouldn't. Not only would keeping the history suck up so much space but people would theoretically be able to know if it was going to hit or not and it wouldn't be random.

I'll admit... what happened to you is just crazy but it doesn't prove that the system is wrong. It may prove to you that it should be changed but it doesn't prove it wrong.

Frank - Don't patronize me tubby. There are no shortcomings in my maths department i stand by what i said... That if it has a 1% chance it's going to happen once in every 100 games... there are more than 100 games being played at any given time so theoretically it is happening right now... I don't see anything wrong with the maths there.

And i know what he is saying but when it comes down to it there is no proof that a) the stats are listed wrong or b) seed made a little error when coding for blocks.

[Frank]

You called me tubby!!!!!!!























Bump...

[dark samurai]

Originally posted by Walrus
but there are certain degrees of probability you take to be acceptable. you are willing to hit a witch from the front, because it is very unlikely to block it. you are willing to let the scout shoot your knight from the front, because you somewhat assume he will block it.


obviously, you cannot be certain, thats the whole point of probability. but when does this begin to seem like more than just improbable? well...to answer this question i have conducted a hypothesis test of a witch blocking from the front.

H0 (original hypothesis) : p= 0.2 (20% chance of blocking)
H1 (my hypothesis) : p > 0.2 (it is higher than 20%)

The test is conducted at a confidence level of 5% (i doubt youll understand what that means, but it is used later in the formulae)

So, i play my games as usual, and take note of all the blocks and failed blocks (from the front) witches do in my games, the results are as follows:

F,F,B,F,B,F,F,F,B,B,F,B,F,F,B,F,B,F,B

B = block
F = failure

Therefore the results are:

Out of 20 attacks, the witch blocked 8 and failed 12.

In my AS level formula booklet, the chance of that happening is : 0.0100

or 1%


this is where my knowledge of statistics 1 mathematics begins to fade, but im pretty sure that since the result is less than my confidence level of 5%, my hypothesis is correct and it is past improbable that such a series of results should occur.

Dude are you like a genius?

[Walrus]

i like to think so, but no. that is just applied maths. i remember the very lesson we were taught hypothesis testing, coz i put my hand up and said "why are we EVER gonna want to know this apart from at the exam?" guess i was owned there.
turned out i got quite a high A grade in AS-level maths though, and im planning on retaining that for my A2 course :D


my underlying point is that : if a witch blocks 5/5 times one game, and 0/20 times in another game (making a total of 5/25 times = 20%) that it is NOT a true representative of 20% blocking. the system should be more tuned to individual battles than the database as a whole.

[Thargor]

i understand your point walrus...we are humans playing this game and feel wronged when the bad side of blocking hits us.

Since this is yet another thing that falls into the yawning pit of "Hey Seed can you fix this" i might as well throw my bit into the pit as well.

So heres my idea, the actual percentages start each game as normal. Then, as each hit or miss occurrs, you alter the % chance for the next hit toward the theoretical percentage. a couple examples:

The witch gets a front block. Therefore the chance for the witch to block from the front again might be 15% instead of 20% Likewise if the witch is hit from the front, the chance to block the next one might be 22%.

[KBHoleN1]

ok guys, this is my first post in a while, so bare with me:

Thargor, I'll start with your comment: while the idea seems good, the basis of the blocking idea is to have odds, just like betting. if you change the odds after each round, you violate serious laws of statistics, and might (i say this b/c i'm too lazy to actually think about it hard enough to come to a conclusion) end up making the blocking more unfair by concentrating all of the combined percentages in one area - i.e. - almost guaranteeing some blocks to players and making others almost impossible (let's say a knight missed 3 blocks in a row, the opposing player should not try the next hit b/c there is almost no chance of hitting). the idea of blocking is 'always equal', meaning each block has the same chance of hitting and takes such timing issues out of the game.

Next: on the subject of blocking in general: i personally dislike the entire idea - i have had my share of bad blocking games, some almost unreal, over my long experience of this game (I was in the original KWOzmaj to give you some perspective) - but i have also had games against people where i almost (notice the qualifier here) wanted to surrender out of pity to them. i feel bad when i win b/c of blocking when i would have liked to play it fairly and see the outcome, or when my good blocking just humiliates someone who's already losing. i consider it my personal bad luck, not my blocking luck, that my good blocking games come against weaker players that i would rather have bad luck against, and that my bad luck emerges when i'm in battle with a worthy opponent - does that make any sense?

Finally: but as much as we bad-mouth the blocking, it is a very good system, one which everyone has been hard-pressed to replace and one which tries its hardest to be fair - and the laws of probability are laws for a reason: Law of Large Numbers says that over many, many repeated trials the probability of an outcome will equal its assigned percentage. so the system is sound, it is only the particular instances of improbability as someone said.

WALRUS: i aced my AP Statistics class my senior year of high school (last year), and understand your confidence test perfectly, but refer back to the Law stated above - 20 trials is hardly a large enough sample size, and taking 20 trials in a row is a sample of convenience, not a simple random sample as would be required for a confidence test:

np>15 and n(1-p)>15 [n=number of trials, p=probability of occurrence]

so (.2)(p)>15, so p>75, and it must be a simple random sample-despite the valiant efforts of your math skills, we cannot trust the results of your test, a larger sample size attained randomly over the course of gameplay would provide evidence for the null hypothesis: that p=.2

sorry if i pissed anyone off or bored anybody, but these are my thoughts and calculations, check them on a statistics website or in a book if you don't believe me - i breezed through the AP test and made a 5 out of 5

peace guys, Kiel (KBHoleN1)

[Thargor]

I personally am fine with the blocking system as it is, and see no reason to question its integrity.

However i know that a great number of non-statistical minded people are greatly frustrated by it, and some have even stoped playing because of it. I'm certain it could be modified in a way similar to what I've suggested (not exactly like that, it was a quick short post) without mucking up the integrity of the system. A front or side block would never be 100% or 0% certain, of course, but it is possible to reduce the wide variance of blocking in games.

Anyway, like I said, I am fine with blocking as it is. The above is a compramise idea just for discussion really, as of course Seed would need to be around and interested in the idea and have the time to do it.....

Oh, and since eveyone else is throwing around their credentials I might as well mention that I am a Statistical (Water Quality) Analyst by profession.

[Walrus]

Originally posted by KBHoleN1

WALRUS: i aced my AP Statistics class my senior year of high school (last year), and understand your confidence test perfectly, but refer back to the Law stated above - 20 trials is hardly a large enough sample size, and taking 20 trials in a row is a sample of convenience, not a simple random sample as would be required for a confidence test:

np>15 and n(1-p)>15 [n=number of trials, p=probability of occurrence]

so (.2)(p)>15, so p>75, and it must be a simple random sample-despite the valiant efforts of your math skills, we cannot trust the results of your test, a larger sample size attained randomly over the course of gameplay would provide evidence for the null hypothesis: that p=.2

sorry if i pissed anyone off or bored anybody, but these are my thoughts and calculations, check them on a statistics website or in a book if you don't believe me - i breezed through the AP test and made a 5 out of 5

peace guys, Kiel (KBHoleN1)


most of that is down to the fact that i cant be bothered to spend half a night or whatever collecting data about witch blocks.

and why would that be a sample of convenience? there is always a 20% chance of a block, hence it doesnt matter how spread out the samples are.

[RikuinKH]

*looks at Duffman's posts* 1000! W00T! good job lol

[KBHoleN1]

Walrus - I understand that it is hard to get that many samples, I didn't say that you should. But to make the study valid, you would need that many. A sample of convenience is polling, say, the first twenty people that walk out of a convenience store and using that to represent all the customers that day. The probability that one customer responds with 'yes' is the same for each customer, but customers must be chosen at random from the total customer base that day. This could be done with a random number generator, like on a graphing calculator. By choosing at random 20 numbers out of 100, lets say, and only polling the customer that corresponds to that number, you obtain a simple random sample. Samples MUST be chosen at random, not in any order or pattern to make the study valid, it is simply a principle of Statistical Analysis.

[RikuinKH]

If blocks were gone then this game would be too easy.



Take a second.....................





Think about it..............





Pretty fuked up game huh? :D

[Cruel_Karma]

i think blocks should just be token off i was facing stick boy i hade like 5 to his 2 and he won because ever hit i did he blocked i was so pi$$ed.:mad:

Blocks are NEEDED without them it would be to easy to win :mad: Just be more "tactics" :D

[KBHoleN1]

And you brought the thread back to share that with us? How wonderful. You can repeat what others say, congratulations. I'm in a bad mood right now, can anyone tell? :mad:

[omnipotent_lord]

There are a number of points I would like to raise about this issue, although most of them are probably repetitive because this argument is almost as old as TAO itself. Firstly the blocking system will almost certainly be unfair in some games, because that is the very nature of chance. However overall you should find that if you play enough games (and consequently as the binomial distribution approaches a normal distribution) you should find that it will even out. So while some games will be wildly biased your way, and others your opponents way overall it will balance. As Walrus pointed out "look at the game over a 50 year period and chances are (even if by pure fluke) the two wrongs will be making a right." That is exactly the nature of these sorts of statistics (symmetrical) and consequently if you play enough games it will be fair.

As for Walrus' sample, "Out of 20 attacks, the witch blocked 8 and failed 12", according to the binomial distribution (p=.2) this should occur approximately 2.216% of the time (close to 1%, but not 1%). Now I conducted two random samples of 20 on my own, and interestingly (well probably not to you, but to me) the results were as follows: 18 Fails 2 Blocks and 18 Fails 2 Blocks.
I got these results by jumping greys attacking my and their dark magic witches from the front (to their bewilderment or amusement). Now KBHoleN1 brought up the very valid point that these samples are far too small to draw conclusions using the null hypothesis (I believe it is because the binomial sample size is too small to approximate a normal distribution), although the numbers he chose were fairly arbitrary for adequate sample sizes. However it is perfectly valid to use 95% confidence intervals for the proportion independant of sample size as they make no such approximations (I think):
Walrus' data: (p^=8/20=.4) 0.185=<p=<0.615
My data: (p^=2/20=.1) -0.031<=p<=0.231 (note the absurdity of a negative probability)
Basically the results show: nothing. We are 95% confident that the witches may possibly have 20% chance of blocking.

Blocking adds an extra element of strategy, as you try to attack from the sides or back for creatures with blocking, with the element of chance. I agree it would be more strategic if, say, the blocking reduced a precise amount of damage from the front and sides. However they chose not to do it that way, and consequently we should just simply accept and enjoy that it's this way, or go and find a chess board.

Finally I would like to mention the idea of changing the blocking probabilities as creatures blocked/failed to block. While it is an interesting proposal, while it would make individual games fairer, I believe it would make the overall system more based on luck (as we slide from a binomial distribution to something similar to a hypergeometric distribution, which I have absolutely no claim of understanding). Consequently it would be easier to, in a large number of games, get a high rating based on luck.

Basically to summarise: all this discussion proves is we can spend too much time on the forums arguing about things such as blocking, which aren't perfect, or we can accept them and get back to the game. I'm gonna do the latter.

[Duffman]

What is it... nCr.p^r.q^(n-r). I used it here too after all this had died down. But I figured it would just start up the argument again.

But i agree with you. I just hope someone doesn't come in here, disregard the facts, and start whinging about blocking again.

[WhiteGhost]

wow that really sucks (the battle)

[Prince Duran]

Yes I agree it can be very annoying. But you cannot tell me that the blocking was never in your advantage. Taking a risks to hit a knight in front is also strategy.

They will never change the blocking system.

[Kyir]

ahg this is MY avatar MINE, MINE, MINE, MINE, MINE, MINE