Games museum (2012/Jun/16)

KBA3u [2012-Jun-16 10:49:46]
3rd - shotgun round

---

kukumber [2012-Jun-18 16:05:23]
Well I've studied as an aircraft engines disigner, precision of calculations was always of a great importance for me. "Airplane probably will fly" is not what would work for me
But since worms are not going to fly in 100% cases we can use any model we like. Just my point of view. As I said it's still interesting for me

---

Dario [2012-Jun-18 14:25:31]
I know what you meant kuku, that's why I encouraged you to analyze something you might find meaningful ;). Sorry there but saying you can't analize a variable because it's interaction with other variables is too complex is saying that you don't believe in modern medicine because we mostly use treatment-overall effect tests instead of accurate modeling of the interaction between the drug and the physiology. Same for pretty much any science except math and theoreticall physics.

By individualizing the too many things you can get an aproximation to which of those too many things have a bigger chances of having an effect on the outcome of a match. That's how we can see that the sarting turn is one of those important variables (even though it interacts with too many other things) and that's how we could discard or prove that a double kill increases the chances of winning, with a bigger sample and more players.

You seem to be confused about what was the goal here. It was not modeling the interaction between the double-kill and all the other variables.

---

kukumber [2012-Jun-18 07:08:50]
Yep, someone messed up with names
And actually I just meant your calculations do not make any sense. Too many other things involved and they all affect the outcome of the game. You can never tell if first turn double kill was a decisive factor. And there is no way to make those variables become constant somehow/ Picking two particular players and comparing the results with winning ratio of first turn moves makes its work but it's just not enough.
Though always interesting to see some statistics

---

Kayz [2012-Jun-17 18:30:24]
"Kayz"? I am not kukumber

---

Dario [2012-Jun-17 16:08:00]
Yea, I'd mind calculating that Kayz, because the hypothesis that whishing your opponent gl alters the game result makes no sense to me ;). But you can do it if you really think it does!.

---

Khamski [2012-Jun-17 15:58:56]
Makes sense.
If we get a turn as a 50 dmg then starter has +50.
Plus he can always kill one worm and sometimes more.
So started automatically has +100 dmg average.
Plus he prevent opponent from making a hit back.
So by hiding his worms he may get +200.
Means he can go 4 turns of his opponent.
Average.
Can be more. Can be less.

But 4 turns is a factor.

---

kukumber [2012-Jun-17 15:45:07]
Mind calculating the influence of how fast you type hf? And do you actually have less chances to win if you also wish your opponent luck?

---

Dario [2012-Jun-17 15:30:34]
Yeap, many many variables, what's interesting here is that among the rounds where you start, getting a double kill or not doesn't make a big difference, maybe no difference at all, maybe makes a difference but not big enough to outstand among other random factors. That's why I am comparing it to the effect of the starting turn alone (because that is a factor powerful enough to make a difference).

---

Kayz [2012-Jun-17 15:19:27]
I once had a 3-2 with Antares, when I won when he started the round, and he won when I started the rounds. Extremely funny. I guess it also depends a lot on how many piles you have, how many ws you must use to save them, and blocked worms etc. There are too many factors in the game. But yes, the starting turn changes a lot, and you must also see, Dario, that you mostly get a double kill WHEN you start, otherwise the opponent hides the most of them.

---

Dario [2012-Jun-17 15:16:48]
Cool Khamski, you sure about the 73-40? because stats say 69-44 (almost the same, easy to make mistakes when counting so many rounds)
Kayz average winning ratio: 38,93%
Kayz average winning ratio when not starting: 20%
Kayz average winning ratio when starting: 52,83% (geez he is good when he starts, I have to work on my skills as second player)
Kayz average winning ratio when starting with a double kill: 61,53% (and only 18 rounds here, so this 61% could easily be a sample deviation from a 52% real ratio).

Big difference between going first and going second, no big difference between going first and going first with a double kill. Some variables not blocked here (like what happens when it's more than a double kill, but I checked all rounds and there were very few starts with more than 2 starting kills).

I don't feel like doing the proper statistical tests for this, but these numbers are already interesting. I was expecting the double kill on start to have a bigger overall effect! Maybe in some particular situations (like only in 3-2 games) it's effect is bigger.

Cool numbers, shows me I don't have to worry about starting with -2 worms more than I should worry about not having the first turn.

---

Kayz [2012-Jun-17 15:16:33]
Anyway, nice statistics :)

---

Kayz [2012-Jun-17 15:16:07]
something must be wrong there, khams. (69-44) it says, not (73-40)

---

Khamski [2012-Jun-17 14:29:41]
Dario - Kayz

Total (no Draws):
113 rounds

Won:
Dario - 73
Kayz - 40

1st moves:
Dario - 60
Kayz - 53

Won 1st moves rounds:
Dario - 48
Kayz - 28

---

kukumber [2012-Jun-17 14:24:59]
Eh seriously?

---

Dario [2012-Jun-17 13:39:51]
No Kayz mate, we can compare, because we can see how much the starting double-kill affects the win/loss ratio. If your winning ratio in rounds were you started and had a double kill is not really different than the winning ratio in rounds you started with less than a double kill, then we can say that there is not enough evidence to prove that a starting double kill is something that affects your overall chances of winning against me.

But I like the idea of using another extra block: 3-2 games, assuming that in a 3-2 game the players were having a similar performance. Since when players have a similar performance luck affects more the result than when one is playing good and the other one is playing bad. And I say extra block and not only one because if you only use 3-2 games then you have very few rounds between 2 specific players, and I really think that these variables (starting turn, double kills) act very differently depending on which players you are talking about regardless their performance.

Maybe nicky can write a small software to analize the whole database in a short time ;). Just say so and we can use some time to define what variables we need to have some good conclusion.

---

Kayz [2012-Jun-17 13:33:13]
And what about rounds where you kill 2 100 hp worms in the middle of the round, like with cows ;P that compensates a lot either.

---

Kayz [2012-Jun-17 13:03:49]
You should not compare Dario with me, he won much more games vs me than I vs him, so the whole statistic is nonsense, you can only make statistics, if you have equality.
Why we just don't compare all 3-2 games of all top 8 players vs other top 8 players, and check about double kills and starting turns there?

---

Khamski [2012-Jun-17 09:17:31]
lol... okay okay rounds
well i am on it then

---

LittleBiatch [2012-Jun-17 09:11:54]
khamski: xD

---

Khamski [2012-Jun-17 08:59:03]
Only 28 games here:
http://www.normalnonoobs.com/replays?listB%5B%5D=Dario&listB%5B%5D=Kayz&from=2011-06-18+09%3A06%3A08&until=2012-06-16+22%3A54%3A32&amount=3&byclan=-1
No idea where to get 113 games between you two.

---

Jellenio [2012-Jun-17 08:17:44]
Ok, but it wasnt a double kill, it was a QUADRA KILL

---

Dario [2012-Jun-17 04:43:59]
just some raw data:
Kayz vs Dario
113 rounds played. Dario 69 : 44 Kayz (tied rounds are not counted here)
18 rounds started with a double kill for someone (some other rounds started with more than a double kill, not counted here. Counted double kill as 2 worms being dead, tolerated some damage to other worms). 12 of them were won by the starter, 3 of them were tied, 3 of them were won by the second player.
13 rounds started with a successful double kill for Kayz. Won 8, lost 3 and tied 2.
5 rounds started with a successful double kill for Dario. Won 4 and tied 1.

I can't make a decent estimation of the advantage given since I have only 13 rounds for kayz and 5 for me and the variable "starting turn" isn't blocked here.

Preliminary conclusions:
Even though starting with a double kill seems to give some advantage, we have to compare the "starting with a double kill" subgroup with the "starting" group and see if the double kill itself makes a difference. Remember that just starting a round gives you some advantage.
Anyone volunteers to do some work? just have to open the 113 rounds and check who started and who won each round.

---

Dario [2012-Jun-17 03:20:06]
Define: double kill (if 200 hp loss counts as double kill or has to be 2 worms)
Define: pairs of equal opponents, pairs of unequal opponents (player names)
Define: "not that good positions"
Define: rarely (as maximum chance of something happening to be considered rare)
Go ahead and give us the stats, we have quite an extense database of replays to check. I am sure it influences the chance of victory, but I have no idea of how much.

---

nappy [2012-Jun-17 02:03:59]
>So wrong. :)

Against equal or better opponent? Oh please. The only chance you basically have after -2 / -3 is when opponent's positions are not that good either.

---

Kayz [2012-Jun-17 01:14:35]
[00:05:05.88] [nappy] you rarely win when you have -2 from start

So wrong. :)

But you are right with that bat, impossible kill.

---

nappy [2012-Jun-17 00:28:51]
Btw, I checked that lg bat in the end of round 5, the worm wouldn't reach the water anyhow, it's too far.

---

Jellenio [2012-Jun-17 00:01:08]
Free 4 kills in secound round :( Hate that not fair game

---