Odds in boxing

[ghjkl]

removed

continue to talk about random staking stuff if you want

delete this thread if you want

[Hameigh]

What it will mean is people will start only fighting if they have a better than 50% chance of winning so then people will have a much harder time finding matches.

[Grimy_Bunyip]

Why did you post your botting script here?

idiocy aside here, congratulations on developing the new code.

So from my understanding it works my running mass simulations as opposed to a markov chain?

So it is able to simulate, the eating of food for example? or healing from blood spells that a markov chain would not be able to account for efficiently?

[ghjkl]

Why did you post your botting script here?

idiocy aside here, congratulations on developing the new code.

So from my understanding it works my running mass simulations as opposed to a markov chain?

So it is able to simulate, the eating of food for example? or healing from blood spells? and soulsplit?

yes is this just mass simulations

it is possible to with the right amount of work

but i simply chose boxing as it is much similar

[Grimy_Bunyip]

Why did you post your botting script here?

idiocy aside here, congratulations on developing the new code.

So from my understanding it works my running mass simulations as opposed to a markov chain?

So it is able to simulate, the eating of food for example? or healing from blood spells? and soulsplit?

yes is this just mass simulations

it is possible to with the right amount of work

but i simply chose boxing as it is much similar

if by "just boxing" you mean just exchanging blows back and forth, then markov chains are the same.

A markov chain is essentially a calculation that runs calculates every single possible simulations, using matrices to optimize the calculation time.

The limiting factor of markov chains is the inability to factor in complex, decision based inputs such as eating food should one fall below X amount of health.

 

as to your question about overkill loss a long long long time ago.

It has occurred to me that while it is very time consuming to calculate overkill losses, once calculated, you never need to calculate it again.

IE once i make a spreadsheet, of a matrix of:

 

Max hit vs Overkill loss

Once that spreadsheet is filled in with the computed values, it does not need to be calculated once again.

Sorry I didn't think of that earlier.

[ghjkl]

Why did you post your botting script here?

idiocy aside here, congratulations on developing the new code.

So from my understanding it works my running mass simulations as opposed to a markov chain?

So it is able to simulate, the eating of food for example? or healing from blood spells? and soulsplit?

yes is this just mass simulations

it is possible to with the right amount of work

but i simply chose boxing as it is much similar

if by "just boxing" you mean just exchanging blows back and forth, then markov chains are the same.

A markov chain is essentially a calculation that runs calculates every single possible simulations, using matrices to optimize the calculation time.

The limiting factor of markov chains is the inability to factor in complex, decision based inputs such as eating food should one fall below X amount of health.

 

as to your question about overkill loss a long long long time ago.

It has occurred to me that while it is very time consuming to calculate overkill losses, once calculated, you never need to calculate it again.

IE once i make a spreadsheet, of a matrix of:

 

Max hit vs Overkill loss

Once that spreadsheet is filled in with the computed values, it does not need to be calculated once again.

Sorry I didn't think of that earlier.

stuff like eating and speccing can be simulated through conditional statements based on the hp of either player

I remember i did something like that a year ago but i lost it :(

[Grimy_Bunyip]

stuff like eating and speccing can be simulated through conditional statements based on the hp of either player

I remember i did something like that a year ago but i lost it :(

im sure you can add a new eating condition in there fairly easily.

You totally should. The limitation of the markov chain is that, although it gives you an exact result in a ridiculously efficient number of calculations, it makes several key simplifications over brute force simulation.

These simplifications do not work if you add in conditional statements such as eating of food, venging, etc.

[Defy]

I lost 17 boxes in a row, and we were both maxed. where as other times i win 10 in a row. its all just luck i dont think there is 50/50 chance. because losing 17 times in a row is like winning the lottery

[curtis95112]

I lost 17 boxes in a row, and we were both maxed. where as other times i win 10 in a row. its all just luck i dont think there is 50/50 chance. because losing 17 times in a row is like winning the lottery

 

I don't think you understand probabilities.

[Assume Nothing]

I lost 17 boxes in a row, and we were both maxed. where as other times i win 10 in a row. its all just luck i dont think there is 50/50 chance. because losing 17 times in a row is like winning the lottery

 

I don't think you understand probabilities.

 

Yep. A particular duel's outcome is not determined by the previous duel's outcomes. Every duel is just as likely to win as any other duel, given that all other factors remain equal.

[n64jive]

Yep, staking someone with equal stats as you is the same as flipping a coin. There really is no way to gain an advantage. High defense seems to be the key to winning non-balance stakes, however you're results prove different, mainly because you don't include Jagex's formula into you're code.

 

If you knew this algorithm, than the answer would be yes, you could simulate duels, however they would be just that, simulations. Plus many people find ways to gain an advantage, mainly via cheating. The key to staking is that it takes capital to earn capital. So even with a max account, there is possiblity of a bad run, which is why so many people get cleaned.

 

End note, staking is not a good way to earn gp. The only way to get ahead is to get lucky, and the best way to get lucky is to do one big stake against someone with equal stats, and then go back to whatever it is that you are doing.

 

Watch NightmareRH's youtube video, he did it right. He was even at a disadvantage(the other player had better stats), and still manage to get lucky. I'm sure he hasn't tried staking since.

[ghjkl]

I did include jagex's formula.

Max hit/attack bonus with no str boost is just the level +5

[MightyMuddy]

I lost 17 boxes in a row, and we were both maxed. where as other times i win 10 in a row. its all just luck i dont think there is 50/50 chance. because losing 17 times in a row is like winning the lottery

 

I don't think you understand probabilities.

 

Yep. A particular duel's outcome is not determined by the previous duel's outcomes. Every duel is just as likely to win as any other duel, given that all other factors remain equal.

Yeah but still given a 50/50 chance of winning/losing, it would be something like this. 0.5x0.5 = 0.25 which is the chance of losing both times, because there are 4 outcomes that are all equally possible. (WW WL LW and LL, with a 25% chance of each option.)

So by this logic losing 17 times in a row, given its a 50/50 chance is something like 0.000763% chance, because theres something like 131072 options. (WWWWWWWWWWWWWWWWW WWWWWWWWWWWWWWWL... ect.) Although it is still true that the next duel has a 50/50 chance of being lost/won.

[Defy]

I lost 17 boxes in a row, and we were both maxed. where as other times i win 10 in a row. its all just luck i dont think there is 50/50 chance. because losing 17 times in a row is like winning the lottery

 

I don't think you understand probabilities.

 

Yep. A particular duel's outcome is not determined by the previous duel's outcomes. Every duel is just as likely to win as any other duel, given that all other factors remain equal.

Yeah but still given a 50/50 chance of winning/losing, it would be something like this. 0.5x0.5 = 0.25 which is the chance of losing both times, because there are 4 outcomes that are all equally possible. (WW WL LW and LL, with a 25% chance of each option.)

So by this logic losing 17 times in a row, given its a 50/50 chance is something like 0.000763% chance, because theres something like 131072 options. (WWWWWWWWWWWWWWWWW WWWWWWWWWWWWWWWL... ect.) Although it is still true that the next duel has a 50/50 chance of being lost/won.

 

 

What he said, i do understand probabilities, im just saying. i could of won the lottery with that percentage. :P

[sees_all1]

Someone that flashes their prayers would do better than someone who didn't, all other things equal.

[n64jive]

I lost 17 boxes in a row, and we were both maxed. where as other times i win 10 in a row. its all just luck i dont think there is 50/50 chance. because losing 17 times in a row is like winning the lottery

 

I don't think you understand probabilities.

 

Yep. A particular duel's outcome is not determined by the previous duel's outcomes. Every duel is just as likely to win as any other duel, given that all other factors remain equal.

Yeah but still given a 50/50 chance of winning/losing, it would be something like this. 0.5x0.5 = 0.25 which is the chance of losing both times, because there are 4 outcomes that are all equally possible. (WW WL LW and LL, with a 25% chance of each option.)

So by this logic losing 17 times in a row, given its a 50/50 chance is something like 0.000763% chance, because theres something like 131072 options. (WWWWWWWWWWWWWWWWW WWWWWWWWWWWWWWWL... ect.) Although it is still true that the next duel has a 50/50 chance of being lost/won.

 

 

What he said, i do understand probabilities, im just saying. i could of won the lottery with that percentage. :P

 

chance of winning the lottery = 5.26*10^-9

chance of losing 17 times in a row, by your odds = 0.00000762939453125 or 7.63*10^-6

 

ie, you are 1000x less likely to win the lottery, or rather the odds of winning the lottery are equivalent to flipping a coin 34 times and having them all land on heads.

 

However, as said, each event is independent of eachother. Losing 17x in a row has the same odds as winning 17x in a row and the same as losing 12x then winning 4 times then losing again...

 

That was a long post. I think you get it. I'm not trying to say you don't understand this, I believe you do. I just figured the true math may actually baffle you.

[pulli23]

one can profit: if one has infinite coins to start with and infinite time/willingness of the opponent to follow:

 

first stake 1 gp..

 

If lost: stake 2.

 

if lose keep doubling the stake amount..

 

Now if you win a game, you'll end up with 1 more gp than you have brought into it. (or actually the amount you start off with).

 

Then go back to the start amount.

 

just requires infinite money + time.

[Jimmyw3000]

one can profit: if one has infinite coins to start with and infinite time/willingness of the opponent to follow:

 

first stake 1 gp..

 

If lost: stake 2.

 

if lose keep doubling the stake amount..

 

Now if you win a game, you'll end up with 1 more gp than you have brought into it. (or actually the amount you start off with).

 

Then go back to the start amount.

 

just requires infinite money + time.

http://en.wikipedia.org/wiki/Martingale_%28betting_system%29

[Golvellius]

You don't account for draws and you don't account for one player being about to attack first.

Both of those are significant reasons why dueling isn't like a 50/50 coin flip.

[Stev]

one can profit: if one has infinite coins to start with and infinite time/willingness of the opponent to follow:

 

first stake 1 gp..

 

If lost: stake 2.

 

if lose keep doubling the stake amount..

 

Now if you win a game, you'll end up with 1 more gp than you have brought into it. (or actually the amount you start off with).

 

Then go back to the start amount.

 

just requires infinite money + time.

If I'm on a relaxed day (More doing it for the fun) I do similar to that.

 

I make sure that every stake, win or lose, earns me an extra 10M. So I do this (If losing):

10M (10M profit if won) - 30M (20M profit if won) - 70M (30M profit if won) - 150M (40M profit if won) - 310M (50M profit if won) - 630M (60M profit if won) - Etc.

 

Very very very fast cash. Odds of getting cleaned... Very very very slim. Highest I've had to go was the 1270M mark and only happened once.

[xSxqPowerx]

Someone already posted a link... but long story short, betting like that does not work.

You have a very good chance of profiting, yes, but eventually one of those times you will simply keep losing until you can no longer afford to double your stake, and get epically cleaned.

[Assume Nothing]

one can profit: if one has infinite coins to start with and infinite time/willingness of the opponent to follow:

 

first stake 1 gp..

 

If lost: stake 2.

 

if lose keep doubling the stake amount..

 

Now if you win a game, you'll end up with 1 more gp than you have brought into it. (or actually the amount you start off with).

 

Then go back to the start amount.

 

just requires infinite money + time.

If I'm on a relaxed day (More doing it for the fun) I do similar to that.

 

I make sure that every stake, win or lose, earns me an extra 10M. So I do this (If losing):

10M (10M profit if won) - 30M (20M profit if won) - 70M (30M profit if won) - 150M (40M profit if won) - 310M (50M profit if won) - 630M (60M profit if won) - Etc.

 

Very very very fast cash. Odds of getting cleaned... Very very very slim. Highest I've had to go was the 1270M mark and only happened once.

 

You only need to lose once, before you're cleaned. Even if you lose about 10 times, you're staking 1024 times more gp than you originally lost. Odds of losing is slim, odds of winning an amount that could pay for potential losses before actually losing is slimmer.

[Stev]

one can profit: if one has infinite coins to start with and infinite time/willingness of the opponent to follow:

 

first stake 1 gp..

 

If lost: stake 2.

 

if lose keep doubling the stake amount..

 

Now if you win a game, you'll end up with 1 more gp than you have brought into it. (or actually the amount you start off with).

 

Then go back to the start amount.

 

just requires infinite money + time.

If I'm on a relaxed day (More doing it for the fun) I do similar to that.

 

I make sure that every stake, win or lose, earns me an extra 10M. So I do this (If losing):

10M (10M profit if won) - 30M (20M profit if won) - 70M (30M profit if won) - 150M (40M profit if won) - 310M (50M profit if won) - 630M (60M profit if won) - Etc.

 

Very very very fast cash. Odds of getting cleaned... Very very very slim. Highest I've had to go was the 1270M mark and only happened once.

 

You only need to lose once, before you're cleaned. Even if you lose about 10 times, you're staking 1024 times more gp than your originally lost. Odds of losing is slim, odds of winning an amount that could pay for potential losses before actually losing is slimmer.

With the system I've been using, I only have to win once to make back my losses and a huge amount of profit. I've not started putting up my phats yet, but if it got that high, I wouldn't hesitate.

 

With the above method (10M-30M-70M-Etc), as soon as I win one, I start back off at 10M. My GF's been cleaned a few times, but I give her some money for some bigger ones and it pays off and we come out on top by far. :P. Fun to see her do 250M+ (Which is never by her choice... Lol). Shakes like a leaf. <3:.

[hfbrads]

Good luck with that. If you continue like that you're going to get cleaned out one day.

 

Staking is an absolute waste of time. Expected value = 0. Doing anything, even picking flax is better in the long run. (Provided you don't have a gimmick)

[Stev]

Yet to lose any money from it. :).

 

But meh, I fly out to Australia on the 20th of April and after that I won't be seen on RuneScape again. So unless it's going to happen in the next 10 days, I think I'm safe. ;). Plus, April 19th I go "all-out" and see if I'll win/lose it all; A chance at a double or nothing of the past 8+ years which will just end up be given to friends. :).