Asturias
Destrier
Posts: 65
Karma: 78
Pilot name: ᴀsτυяιαs ẕωєι ⚙
Platform: iOS
|
Post by Asturias on Feb 24, 2017 21:21:34 GMT -5
UpdatePlease see Strayed 's post here for a quick debunking of my theory. While this refutes the "given to player with fewer wins" theory, I still suspect that the "not random" part might still hold water, especially with the uncanny 50-streak. - Please continue to send in debunking evidence against the original hypothesis, I'm looking for trends.
- Does the "fewer wins" hypothesis still hold when you consider only L30 v L30 comparisons?
- If we were to assume that it is indeed not random, is there any testable claim which can explain my observations?
- Another of my pet theories was "account which was registered later gets the 5Au", is this testable? I think it's quite obvious how this could easily be correlated with number of wins, and the win count is simply a confounding factor...
My ClaimBeacon Ties are NOT assigned randomly. The 5Au is given to the player with fewer wins. What you can doAll I need is one Black Swan to disprove the hypothesis, so if you find a match where the player with more wins gets the 5Au, please screenshot - The Scoreboard
- Beacon Tie Winner's Profile
- Beacon Tie Loser's Profile
This will help firstly to disprove the theory, and secondly to give other clues if there is another non-random factor in play.
MethodologyHaving this suspicion, I observed separate instances whereby there was a beacon tie, until I obtained 50 samples of 2-way Beacon Ties. I took note of the number of wins on the profiles of the winner and the loser(s). ResultsThere were no 4-way or 5-way ties across the duration of my experiment. There were (3) 3-way ties. The 5Au went to the player with the lowest number of wins 3 out of 3 times. There were (50) 2-way ties. The 5Au went to the player with the lowest number of wins 50 out of 50 times. Wolfram Alpha Calculation: prob x < 1 for x binomial with n=50 and p=.5
gives
8.88176 x 10^-16 If Au were truly assigned randomly with a 50% chance in the case of beacon ties, the possibility of the player with higher wins winning the tie 0 out of 50 times is a 1-in-a-quadrillion chance. This leads me to conclude that beacon ties are not random. I am slightly less confident that the factor in play is definitely the number of wins, and this might simply be a confounding factor correlated to the actual factor, but I am almost certain that it is not random.(Unlike my previous experiment, I apologise that I did not take screenshots for this due to the sample size of 50. In some sense, you will have to take my word for it, but I invite you all to test the theory out for yourself by taking your own observations, and please send in the data here if you have contradictory evidence.)
|
|
|
Post by EatStinkyTofu on Feb 24, 2017 21:26:54 GMT -5
Off the top of my head, your p value is definitely less than .05
|
|
|
Post by amoebastudios on Feb 24, 2017 21:28:39 GMT -5
I personally feel there are "partial" beacon captures that gets counted. I'll do some counting.
|
|
|
Post by Ⅎ₹ѺC₭₩ELDEℲ₹ on Feb 24, 2017 21:29:35 GMT -5
I cant believe you tested this... Talk about drilling down on something... Random? This one is pretty damn deliberate...
|
|
|
Post by ł⸰§ĦȺĐ◎ŴƧŦḀɌ on Feb 24, 2017 21:47:16 GMT -5
Are you a wizard?
I'll have to take your word for it... o.0
|
|
|
Post by launchpad on Feb 24, 2017 22:02:19 GMT -5
Carefull dude, going against mainstream opinions probably will get you banned. They working on he rules.....
|
|
|
Post by ewing411 on Feb 24, 2017 22:25:52 GMT -5
Let the likes be his shield. Paper thin... like medical paper gown thin.
|
|
|
Post by zeus on Feb 24, 2017 22:33:39 GMT -5
I like that we have a standing theory on this with some proof to back it up. It might not be true, but it's a start. I will take notice of beacon ties from now on, be it mine or others. Not that hard to tap in and check their profiles.
Personally, I still think it has something more to do with flipping reds counting more. And being the first to flip the last red that results in an eventual tie will get you the win. I can't really back it up though, just observations. But yours is a fresh angle I have yet to consider in my games.
|
|
|
Post by Strayed on Feb 24, 2017 23:15:08 GMT -5
Hm, interesting, will bring this up with Pixo.
I don't get why they didn't just tell us this in the first place. I would vote for the old system where they player with the higher damage won the tie.
|
|
|
Post by Strayed on Feb 24, 2017 23:41:26 GMT -5
|
|
|
Post by whatttupG on Feb 24, 2017 23:50:22 GMT -5
scientific/engineer/data/facts language instantly detected, so much so it broke my eGuage, I wanna award agreement now without doing any research as requested... but... I'm gonna try anyways with the expectation that I cannot live up to this OP
stellar post, it should get an award
|
|
|
Post by blastronaut on Feb 25, 2017 1:22:49 GMT -5
Strayed just debunked it.
I always got the feeling the tie breaker went to the last player to cap a beacon before the match ended, but I could be totally wrong.
|
|
Asturias
Destrier
Posts: 65
Karma: 78
Pilot name: ᴀsτυяιαs ẕωєι ⚙
Platform: iOS
|
Post by Asturias on Feb 25, 2017 1:35:05 GMT -5
Thanks Strayed, this is exactly what I was hoping to find. This is why you need other people reproducing the experiment to see if it holds in all circumstances. Clearly, it does not.This is definitive evidence that "5Au goes to least wins" is wrong, or at least does not apply to all cases. My immediate gut reaction is to point at the L27 and 28; could that be a factor? It's not something I can test within my games because I see L30 almost all the time; and anyone with lower than 30 is definitely going to have less wins. My streak of 50 observations with everything going one way still makes me believe strongly that it is not random, even though my explanation for the mechanism was wrong. Anyone have other ideas? One of my other pet theories was "the one with the newer account gets the 5Au", which would in most cases (but not all) also be correlated with less wins... Very hard to test though.
|
|
|
Post by truechill on Feb 25, 2017 1:35:32 GMT -5
Well that theory was short lived.
|
|
|
Post by nocluevok on Feb 25, 2017 3:53:27 GMT -5
My ClaimBeacon Ties are NOT assigned randomly. The 5Au is given to the player with fewer wins. What you can doAll I need is one Black Swan to disprove the hypothesis, so if you find a match where the player with more wins gets the 5Au, please screenshot - The Scoreboard
- Beacon Tie Winner's Profile
- Beacon Tie Loser's Profile
This will help firstly to disprove the theory, and secondly to give other clues if there is another non-random factor in play.
MethodologyHaving this suspicion, I observed separate instances whereby there was a beacon tie, until I obtained 50 samples of 2-way Beacon Ties. I took note of the number of wins on the profiles of the winner and the loser(s). ResultsThere were no 4-way or 5-way ties across the duration of my experiment. There were (3) 3-way ties. The 5Au went to the player with the lowest number of wins 3 out of 3 times. There were (50) 2-way ties. The 5Au went to the player with the lowest number of wins 50 out of 50 times. Wolfram Alpha Calculation: prob x < 1 for x binomial with n=50 and p=.5
gives
8.88176 x 10^-16 If Au were truly assigned randomly with a 50% chance in the case of beacon ties, the possibility of the player with higher wins winning the tie 0 out of 50 times is a 1-in-a-quadrillion chance. This leads me to conclude that beacon ties are not random. I am slightly less confident that the factor in play is definitely the number of wins, and this might simply be a confounding factor correlated to the actual factor, but I am almost certain that it is not random. (Unlike my previous experiment, I apologise that I did not take screenshots for this due to the sample size of 50. In some sense, you will have to take my word for it, but I invite you all to test the theory out for yourself by taking your own observations, and please send in the data here if you have contradictory evidence.) I think it is much simpler. You capture 1 beacon and liberate 1 beacon. I liberate 2 beacons...I win the tie. If we tie in captured and liberated beacon count, whomever did it first wins the tie. Just my observation without all the data mining.
|
|
Asturias
Destrier
Posts: 65
Karma: 78
Pilot name: ᴀsτυяιαs ẕωєι ⚙
Platform: iOS
|
Post by Asturias on Feb 25, 2017 6:18:13 GMT -5
You capture 1 beacon and liberate 1 beacon. I liberate 2 beacons...I win the tie. If we tie in captured and liberated beacon count, whomever did it first wins the tie. Just my observation without all the data mining. I do wish it were this simple and this fair. However I've lost ties to Stalker pilots who cap two or three (usually Yama) beacons at the start of the game and then eject leaving us to play 5v6. There's no way anyone can liberate less beacons than them. I seldom get annoyed at this game, but those cases make me hate seeing Stalkers on Blue and really stick out in my memory. It's unlikely.
|
|
|
Post by sochilli (Saltesers) on Feb 25, 2017 7:06:05 GMT -5
|
|
mr7q
Destrier
Posts: 71
Karma: 48
Pilot name: mr7q
Platform: Android
|
Post by mr7q on Feb 25, 2017 13:04:43 GMT -5
There is no way it's random... I've been keeping track for a month or so, and I'm ~8% of getting gold when tied. I'd love to know what the actual logic here is.
|
|
|
Post by whatttupG on Feb 25, 2017 13:20:31 GMT -5
Been playing with an eye on this every match now, have a couple screen grabs but we don't need all that I'd guess...
Saw a few ties but the 5 gold does not always go to low win guy. Twice last night I tied someone for beacons and still took the 5 gold and both times I had more wins so as far as the win count theory goes, you might be close but this isn't it.
|
|
|
Post by FRΞΞDØM☆F1GHTΞR on Feb 25, 2017 13:57:53 GMT -5
Another possible theory:
Partial caps weigh less than full caps.
If a teammate starts the cap and you jump on to speed it up, you each get credited for a cap. However, since the teammate initiated the cap, they receive the gold.
Theory is completely based on speculation with no proof whatsoever.
|
|
|
Post by nocluevok on Feb 25, 2017 14:13:18 GMT -5
I think any theory is hard to prove on this one, given that you would have to see every beacon capture in the match to have hard evidence. Don't know about you, but I don't have enough eyes to keep up with what all 12 bots are doing AND concentrate on winning.
|
|
Asturias
Destrier
Posts: 65
Karma: 78
Pilot name: ᴀsτυяιαs ẕωєι ⚙
Platform: iOS
|
Post by Asturias on Feb 25, 2017 18:59:19 GMT -5
There is no way it's random... I've been keeping track for a month or so, and I'm ~8% of getting gold when tied. I'd love to know what the actual logic here is. Yeah that's what started me thinking and taking observations. I have a clanmate ( Fishin_Chip) who never wins beacon ties ever since he moved to the daughter clan, whereas he reports that he used to do so often when originally in the parent clan and sometimes does so in squads with experienced players. Although anecdotal, his streak suggested that there's some factor in play which pushes the beacon tie in the favour of weaker / less experienced players, and so I set out to test that. I'm sorry to hear that you have the same issue! Is there any information you could provide about your account / playstyle which might help us figure this out? Been playing with an eye on this every match now, have a couple screen grabs but we don't need all that I'd guess... Saw a few ties but the 5 gold does not always go to low win guy. Twice last night I tied someone for beacons and still took the 5 gold and both times I had more wins so as far as the win count theory goes, you might be close but this isn't it. Thanks for making the effort to take observations! Quick question, is everyone involved L30? Thus far all Black Swan examples have a non-L30 involved. Would love to see your screenies in either case. I'm extremely puzzled why I can observe a streak of 50 all supporting my hypothesis (seemed like very strong data to me) but then others can come up with refuting evidence so quickly; there's obviously something in the kind of matches / tier I'm playing where it works, but it doesn't work everywhere. Can't decide whether I'm frustrated or extremely excited by the potential of this.
|
|
kjester
Destrier
Posts: 31
Karma: 16
|
Post by kjester on Feb 25, 2017 19:10:54 GMT -5
Crap. See what you did? Now I am going to be paying attention to this. I was happy just grumbling about how stupid it was that it was random and now you go and prove it is likely even more stupid because they are probably giving the beacons to someone who definitely deserves it less. I would bet. Now I am going to start keeping track of a bunch of different stats just because I gotta know!
|
|
|
Post by whatttupG on Feb 26, 2017 13:29:03 GMT -5
There is no way it's random... I've been keeping track for a month or so, and I'm ~8% of getting gold when tied. I'd love to know what the actual logic here is. Yeah that's what started me thinking and taking observations. I have a clanmate ( Fishin_Chip) who never wins beacon ties ever since he moved to the daughter clan, whereas he reports that he used to do so often when originally in the parent clan and sometimes does so in squads with experienced players. Although anecdotal, his streak suggested that there's some factor in play which pushes the beacon tie in the favour of weaker / less experienced players, and so I set out to test that. I'm sorry to hear that you have the same issue! Is there any information you could provide about your account / playstyle which might help us figure this out? Been playing with an eye on this every match now, have a couple screen grabs but we don't need all that I'd guess... Saw a few ties but the 5 gold does not always go to low win guy. Twice last night I tied someone for beacons and still took the 5 gold and both times I had more wins so as far as the win count theory goes, you might be close but this isn't it. Thanks for making the effort to take observations! Quick question, is everyone involved L30? Thus far all Black Swan examples have a non-L30 involved. Would love to see your screenies in either case. I'm extremely puzzled why I can observe a streak of 50 all supporting my hypothesis (seemed like very strong data to me) but then others can come up with refuting evidence so quickly; there's obviously something in the kind of matches / tier I'm playing where it works, but it doesn't work everywhere. Can't decide whether I'm frustrated or extremely excited by the potential of this. Well the data was so strong it had me curious so no problem.. I agree as well, a 50/50 streak is beyond compelling alone as a case and would pass any statistics based analysis. So much that I expected to lose ties and was surprised here. I had two, one was a two way, one was a three way where I won both... I hit the wrong button before getting the third screen cap so anyways... based on what I did get, I'm at 30 and whoever I beat out at 28 or so. This must mean it's not just levels either. So the guy with more wins or higher level is not automatically anything. I saw somewhere, don't recall now, but looking into this a while back it said the center beacon is worth a bit more than the others. Not in beacon cap counts, but weighted heavier in calculations behind the scenes. So it may be important to say that my approach for beacons includes both a run for most in numbers and I always try for center beacon first, Even in shenzen. Usually it costs me a bot and is a pretty high price to pay just for one beacon but, my tact usually gets me center first, usually costs me that bot and usually gets my team to at least 3-2 to open. This is good cuz: Advantage me in beacons, advantage team in beacons, and maybe is a more productive demise for an opening capper that I'm gonna play til dead anyways... dunno there... but I think this may be relevant because I seem to win a lot of ties so which one you grab might be the factor in this mix, as well as, how long that grab then lasts. I don't have seen caps on this so a theory I'd offer up is when conditions limit me to 2 beacons in the finally tally, it seems I never win. When 3 I win quite a bit, ties too, as 3 seems to be a common count. 4 or more usually closes the door outright and is a winner most everytime. Personally, my approach is to try and manage 3 or more and will almost always include center if I can. Off this topic a hair, going 3-2 to open.. which sometimes is 3-0 on screen, seems to lead to alot of 4-1 and 5-0 starts which are devastating to the red bar. This is a more rare start but when I've seen it, it seem we win everytime. My record on 5-0 is a minute ten and was in squad with 4/6 clan and it just played out that we went blitzkrieg from hell and mowed our way right thru the center of shenzen. Who does that LOL.
|
|
|
Post by amidf on Feb 26, 2017 16:29:51 GMT -5
|
|
|
Post by whatttupG on Feb 26, 2017 16:43:18 GMT -5
Oh snap... viewing/comprehending/posting accurately by phone just isn't possible in threads like this is it?
Ouch......
|
|
|
Post by whatttupG on Feb 26, 2017 16:59:42 GMT -5
1. Please continue to send in debunking evidence against the original hypothesis, I'm looking for trends.
I'll skip my screen cap postings, this theory horse has been beaten pretty well already.
2. Does the "fewer wins" hypothesis still hold when you consider only L30 v L30 comparisons?
None of my examples held this facet, so I'll continue to look at ties with this new criteria in mind.
3. If we were to assume that it is indeed not random, is there any testable claim which can explain my observations?
Yes, however you would need data not available to users I think. You would need to account for things like which beacons are credited to a final tally, does this whole 2 guys both score in a near tie thing impact anything, does the time a beacon held impact anything, is there a which one times how long factor exist and impact, is there a difference with when the beacon was captured first/middle/last, its variable hell to me.
4. Another of my pet theories was "account which was registered later gets the 5Au", is this testable? I think it's quite obvious how this could easily be correlated with number of wins, and the win count is simply a confounding factor...
This would maybe be easy to test. If the pilot ID is sequential like license plates in most places, it's very easy to guess the vintage of a rig and if it was registered before or after you. For instance my Jeep plate begins with a 3 and so does the wifes car. Both are from the same year, both are pretty old if you are talking average. Today (in CA I guess) new cars have plates beginning with either 6 or 7 as.. all the 3's are gone, all the 4's are gone, all the 5's are gone, most the 6's too and now we're in the 7's.
Anyways before spinning off topic, I my pilot ID begins with ZGZ and is about 4 months old,and I would bet old timers (so to speak) have ID's way earlier like Y or X, and maybe new ID's are now numbers or something because there isn't a letter after Z to roll to...
|
|
|
Post by amidf on Feb 26, 2017 17:14:28 GMT -5
I have been playing about 2 months and my tag starts with C. But I like your thinking. If someone were to register multiple new accounts within seconds of each other, we might learn something.
-Amid
|
|
|
Post by miatahead on Feb 26, 2017 19:52:19 GMT -5
I have a theory. I'm 99% sure it is *not* random and when I'm running beacons, I can usually guess if I get the 5au or not on ties. But I ain't saying anything. It's my trade secret.
|
|
yhondeh
Destrier
Posts: 122
Karma: 94
|
Post by yhondeh on Feb 26, 2017 23:07:14 GMT -5
|
|