Printable View
A couple of well known forumers have theorised this possibilty now for quite some time with varying degrees of confidence. Im going to go ahead now and say your gold weight is no longer your defensive war weight. 97% sure.
What are your guys findings?
I still use it to calculate weights and it is accurate. It is quicker then trying to input all of the other weights.
Example I posted in Mismatch thread:
Clan ID: #QP2YLPPY
Clan name: Star Brawlers
War Start Date: 06-14-18
War Start Time: 9:35 PM GMTNotes: Star Brawlers won 29 - 0 in this mismatch. Their clan had a TH9 that had two Level 3 Xbows and we could only get 2 stars. For some reason opponent disbanded the clan they could have got a perfect to win. CopIt lists their number one as engineered thus we defeated another engineered clan due to no particpation. US Rank TH Level BK Level AQ Level GW Level Weight Gold 1 Gold 2 Gold 3 Gold 4 Town Hall 1 8 10 0 0 55 13750 13750 13750 0 13750 2 8 10 0 0 55 13750 13750 13750 0 13750 3 8 10 0 0 55 13750 13750 13750 0 13750 4 8 10 0 0 54 13500 13500 13500 0 13500 5 8 8 0 0 52 13000 13000 13000 0 13000 6 8 7 0 0 49 12250 12250 12250 0 12250 7 7 5 0 0 40 13332 13334 0 0 13334 8 7 1 0 0 38 12666 12667 0 0 12667 9 7 3 0 0 36 12000 12000 0 0 12000 10 5 0 0 0 19 6332 6334 0 0 6334 Totals 7.4 6.4 0 0 453 124330 124335 80000 0 124335 THEM Rank TH Level BK Level AQ Level GW Level Weight Gold 1 Gold 2 Gold 3 Gold 4 Town Hall 1 9 12 10 0 64 12800 12800 12800 12800 12800 2 8 10 0 0 52 13000 13000 13000 0 13000 3 8 10 0 0 51 12750 12750 12750 0 12750 4 8 6 0 0 49 12250 12250 12250 0 12250 5 8 6 0 0 46 15332 15334 0 0 15334 6 8 3 0 0 33 11000 11000 0 0 11000 7 7 4 0 0 39 13000 13000 0 0 13000 8 7 2 2 0 38 12666 12667 0 0 12667 9 7 2 0 0 37 12332 12334 0 0 12334 10 7 1 0 0 35 11666 11667 0 0 11667 Totals 7.7 5.6 1.2 0 444 126796 126802 50800 12800 126802
How do you know Its accurate?
Look at the data. If you compare their number one to our number one see who has the most weight. Also in war it gives you those values and when you upgrade your base they go up. In a degree of difficulty their number 1 has two level 3 xbows and is a town hall 9 hence it has the most weight for both clans. While our number 10 has the least weight. I also been using this for several wars it is accurate.Quote:
Originally Posted by Vikingchief
How do you know the place on the list is the actual defence war weight value used in the match making algorythym? Of course the gold weight matches the map placements, as those are based on gold weight, but how do you know that those values are the defence weight values used to match your clans bases?Quote:
Originally Posted by KingAroth
Their number 1 has stronger defenses than our number 1. It has maxed TH8 defenses and two Level 3 Xbows. If you want to calculate based on actual defensive weights you can do that and I did do both for a time. I used the gold method and defensive weight method and they still shown which clan had the most weight. As for the war I listed we do have better defenses we have 5 ranks with 50 point weight and in my clan I have another spreadsheet with defensive weights the order of the bases match what is in the gold method. How your members are listed in war is determined by defenses with a tie breaker using experience points. I have three maxed TH8s and the one with the most experience is always first. So the member with the highest weight is listed first and the lowest last with any ties using the highest experience. So this method is still valid in calculating weights in wars.Quote:
Originally Posted by Vikingchief
IMHO it really does not matter any longer. Why bother tracking defensive weight if offensive is of equal or greater value in matchmaking and invisible?
My opinion has also changed on gold weight. I feel that it is simply a bonus awarded for a "win" on the base, adjusted simply by the level of difficulty. It can still be used as a rough guide, but is unlikely to be a true weight. Of course, elixir value is the same, so either can be used as determinates. I also agree with Doom. Even if it were accurate, there are too many other determining factors involved, for it to be very useful.Quote:
Originally Posted by Vikingchief
Yes, the gold weight is still tied to defensive units and yes it will go up when you buy new defensive units or upgrade. The question, though, is whether those gold weights you see actually mean anything at all. For example, if the inferno weighs A in the gold weight but for matching purposes SC considers it to weigh B and just hasn’t told anyone of that fact, the gold weights end up being meaningless.Quote:
Originally Posted by KingAroth
Highly agree with you. And non only offensive weight, there are so many other factors that impact on matching. It is meaningless to discuss about gold storage weights anymore.Quote:
Originally Posted by DOOM2U
Yeah, the war we just had showed this. Base 2 had higher gold weight than base 1.
You shouldn't be running that many 8s without a 9 up top this will just keep happening to you.Quote:
Originally Posted by KingAroth
They learned their lesson now it's your turn, and you'll pair against 8s.
And I doubt you'll let them win ;)
We’d want a picture to be sure you weren’t mistaken, as while there’s doubts the defensive war weight correlates to gold weight, so far nobody has been suggesting gold weight doesn’t correspond to map lineup order.Quote:
Originally Posted by TheWarDictator
i dont think this thread matters because obviously our positions on the map are determined by the gold storage numbers which are pretty much directly influenced by what defenses we have dropped...
what we don't know is how much unplaced buildings are affecting the actual weight which would be impossible to see because well, they aren't dropped so we have no physical way to account for them. we would have to rely on a trend of what defense gets matched.
for example if a clan has 3 players that bought the th11 pack, never placed eagle but match eagles anyway then perhaps we could make the assumption that the eagles that aren't placed are still accounted for but even that is still a vague comparison.
for what it's worth though here are the gold total from our last 40 v 40.
our clan had 4,209,000 gold total vs the enemy 4,075,000 gold total. they had an extra th12 but we had 2 extra th11s and 4 extra eagles. the war was 32 eagles vs 28 eagles and 37 inferno bases vs 35 inferno bases
No one can say for sure other than super cell. But I track increases on a spreadsheet for my clan after each war and the increases in gold correspond with the defensive upgrades. This has been the case through my last ware that ended yesterday. However, what the gold weight likely does not capture is weight of unplaced defenses, which are now counted in matchmaking. I assume that weight is hidden along with offensive weights. Once the unplaced item is placed, however, then it will likely impact viewable gold war weight on war map.Quote:
Originally Posted by Vikingchief
Wow I had no idea that this was actually about gold storage's this is absolutely absurd. I can't imagine the MM ever functioned like this.Quote:
Originally Posted by RepublicanFarm
You need a heavier anchor regardless, lets say a clan ran 60% MAX 11s their going to pull 12s sometimes if not always.
SC has stated many times they are not going to match TH levels.
It's just its much safer to run a war with a stronger TH up top, and I guarantee you that war is not and has not been based on gold for as long as I've played.
Also with the recent buffs to drags, cc troops, etc. the best you could hope for is a draw with a 6 TH8 anchors.
Just push to 9.
That's a excellent point but we don't know that amount of gold is meaningless either.Quote:
Originally Posted by 2222
As long gold storages increases as the weight increases, I'm going to consider that gold weight is the best way to get weight.
The reason I think that gold weights are the actual weights because you're still ordered in the war respective to your gold weights. The higher the number in war map, the difficult your base is. And the gold shows how tough you are. That's my theory.
I'll leave the thread I just posted as my comment on this - I'd been working on that for a couple of days.Quote:
Originally Posted by Vikingchief
Well, it was probably possible to analyse offence weights by looking at recommended targets. But this relies on the gold in the targets being meaningful.Quote:
Originally Posted by DOOM2U
Also, it's possible to get an advantage on defence by optimising defences. If you could replace a max TH9 with another 70k base that's physically impossible for a TH9 to clear, and not certain for a mid-large TH10, then that's a significant advantage of itself. And that wasn't hard to achieve using gold weights.
That's always been possible but rare. The map order is fixed when the war search starts. But the gold loot is recalculated at the end of prep day. So if any base upgrades (or adds guns) during prep, they'll sometimes have a gold loot which is greater than the base above.Quote:
Originally Posted by TheWarDictator
You have been on this train of thought for a while, and your one of the voices one should pay attention too, skrags made some pretty comvincing posts too, the matchmaking seams to be in line with your guys thinking very clearly, the only reason I have been struggling to accept it was that two others of the forums war weight weight superpowers (tank and sservis) seam to believe that gold still is the number to go by.Quote:
Originally Posted by OnyxDS
The in game experience is making it very hard to believe gold weight is still the gold standard.
We are talking specifics here mate. Not how the roster should break down, for optimum war effect, but as a means of monitoring the defensive upgrade path of the actual account. Using gold storage is, indeed, a pretty good way of achieving this. Only question is.. How accurate are the totals.Quote:
Originally Posted by BlindSyphon
Sure, you can record the weights, and prove they increase with every defensive upgrade that is done. That makes it an effective tool in plotting your optimum defensive upgrade path.Quote:
Originally Posted by RepublicanFarm
It also goes a long way towards proving, beyond reasonable doubt, that defence is linked to gold weight, or vice versa. What it doesn't do tho, is prove in any way that the totals are accurate. They could simply be a bonus awarded for attacking a harder target, and graded accordingly.
I will add tho, that at one time, I had 3 accounts at th9, maxed except for walls and heroes, and all had the same gold value. This could be one of 2 things.. Either they had the same defensive weight, meaning walls, especially, carry very little extra weight going from level 9 to 10, or they simply offer the same reward, since a dozen or 2 wall pieces doesn't make a base any harder, when putting things into perspective.
I feel it is still a good way of keeping check, but have doubts on accuracy. A big thing that casts doubt for me, is that they announced even defences owned but not dropped yet, are still included in the weighting system. I cant buy into the phantom weight ideas, it doesn't make sense for me, that they would do this. I feel it would be easier for them to just add it, whether its dropped or not, and include it in the gold weight, if gold weight were an accurate guide.
Yep, but SC could be using a totally separate set of weights for matching. They very easily could decide infernos weigh too much in gold weight and change it up for matchmaking, but not touch the gold weight.Quote:
Originally Posted by ExpertSAQ
Wasn't aware of this, can someone provide link to where it was announced please?Quote:
Originally Posted by joshsgrandad
https://forum.supercell.com/showthre...ll-Patch-NotesQuote:
Originally Posted by DOOM2U
under heading new town hall requirements.
https://forum.supercell.com/showthre...ll-Patch-NotesQuote:
Originally Posted by DOOM2U
New Town Hall Upgrade Requirements
Introduced in this update is a new requirement system in order to upgrade to the next Town Hall level. All Buildings & Defenses from the previous Town Hall level(s) must be placed in order to upgrade to the next Town Hall level. You will not be able to upgrade to the next Town Hall level until all available Buildings and Defenses have been placed in the village. Any buildings purchased in an upgrade pack count as already placed, with regards to War Weight, thus not placing a purchased building holds no advantage in War.
As to storage weights being used in MM, it's going to come down to your null hypothesis. Here's my case as to why storages weights being the
I this logic is ironclad, especially when you consider how much more SC cares about the casual masses. Make rec target/roster sort, a feature useful for the masses, less accurate to hose a tiny number of engineers? I don't see it.Quote:
Originally Posted by TankSinatra
I also think def weighting is less than half of the whole MM system now, so good luck pulling proof out from the noise either way.
Im not sure it is iron clad, clearly supercell has labelled engineering as a issue they intend to fix, and they have been aware of it for sometime. Surely the smartest way to play the situation is to mislead people to prevent sucessul engineering efforts? You still require a weight for loot, but why wouldnt you Iether have a server side weight system (or modifcation code which is in effect a seperate weight system) as this would significantly hinder engineering efforts, while still encouraging the gem the troops and heroes incentive If people believe the key defences carry huge weight, and they need to gem troops to carry their war weight?Quote:
Originally Posted by TankSinatra
So hiding the weight encourages gemming and mitigates engineering, in addition supercell have said they are tweaking weights with no visible effect that I am aware of on gold weights. And again a series of multiplication factors or modifiers is in effect a server side weight system.
For me the null hypothesis at this point has to be a server side weight system, I know you, sam, and sservis all disagree strongly which I admit gives me pause. Wonder what humbob would say.
This is true, I will concede this may be a compelling argument as the base of weight but it would be more of a constant that doesn't affect the matching systems final outcome. Trying to run 50% or more anchors of near the same weight when their are still higher THs is unrealistic without an engineered roster to a large degree.Quote:
Originally Posted by joshsgrandad
This could be a compelling argument for max THs but to think you can run 5 maxed lower town halls and not have a higher anchor match based on this just isn't realistic.
Engineers always max storage's, and casual players do as well and with the forced storage upgrades at most TH's esp 12 to upgrade anything it just doesn't seem to be something a war roster should be based on.
MM is different in each size of war 5v MM has a world of difference from 10 or 20/25v and 25 has a world of difference from 30v and so on.
Their might be on or two MMs that this has a larger effect on but across the board matching varies based on war size.
The larger the war roster the more anchors matter and perhaps the less a gold standard comes into play.
All I know for sure is you can't fight the MM to do what you want so if I had to choose between having the advantage half the time and balanced wars half the time I would take that one over it being balanced half the time and at a disadvantage half the time.
But the rec target system is needed, why make it deceptive in a way that hurts the masses that can use it? Makes sense only for a "penalty", whether weight tax or unplaced buildings, where SC wants it to affect the match but it shouldn't affect the rec (because the building isn't there). We may have a single example of this now, the unplaced eagle. But I'm not sure, not seeing it in my matches.Quote:
Originally Posted by Vikingchief
I think we all agree on server side weights by the way, storage weights are provided server-side. That's why Spanser never found useful weights, he wasn't digging into the server. The question is, do they have a super secret one in addition to the storage weight system? It's possible, but can't be proven or falsified. Don't see any evidence for it either.
Again its going to come down to your null hypothesis. I've put down how I arrived at mine, its very logical though maybe I shouldn't rely on logic in a game where you upgrade a bomb by hitting it with a hammer.
So lets flip the question, any evidence that storage weights aren't going into the matchmaker? I think the unplaced eagle issue is all we've got, and it appears to be extremely isolated.
Apologies if I'm not following, but I think you're way off track. When we talk about storage weight we're talking about gold shown in storages, on war day. Also visible during a scout. Nothing to do with storage upgrades.Quote:
Originally Posted by BlindSyphon
Thats the only actual evidence I have. Agree Its possibly an exception case.Quote:
Originally Posted by TankSinatra
Excellent thank you for the clarification.Quote:
Originally Posted by TankSinatra
With wall prices reductions it will be really rare for max players to ever have gold in their storage's though i'd think.
I might do a 15 or 20v war with less than 500k gold in all my storage's just to test this, anything else to consider?
Yes, have someone do a friendly challenge. Scout (don't attack), note how much gold or elixir is in a storage.Quote:
Originally Posted by BlindSyphon
Then go visit the farm base, you'll see a different amount. When we talk about storage weights, it has nothing to do with actual gold.
But they didn't make target/roster sort less accurate. It was already like that; no work required at all. They needed a sensible list of weights to make the MM work. To make the server run the calculation twice using the 2 different files might only be a couple of lines of code - 5 minutes work. Much less effort, probably, than making it count bought but unplaced buildings.Quote:
Originally Posted by TankSinatra
Well they both showed one tick above empty on the friendly challenge. TH12 lvl3 and TH11 max with lvl1 eagle.Quote:
Originally Posted by TankSinatra
I do remember wondering though in war why the storage's sometimes appear fuller during match day though.
I'll be looking for this now, thx.
Somethings not right. Try checking your war log, look at the war map for PAST wars (won't work for current map in prep day). Check storages - that's a weight indication, has nothing to do with what they've raided.Quote:
Originally Posted by BlindSyphon
As it has been pointed out, also, the storage weight in war equals defensive weight only. This has been tested and tested a gazzilion times. Upgrade any offense and don't touch any defense and the weight will remain the same.Quote:
Originally Posted by TankSinatra
This is a long time myth. Measuring storage weight only tells you how "heavy" the base is defensive wise.
And leave the nonsensible ones in place on the roster... see what I mean? You're assuming SC puts a high value on slightly misleading the small number of engineers that really crunch the numbers, and slight value on providing the masses with some kind of accurate indication of strength through roster position and rec target.Quote:
Originally Posted by OnyxDS
I'm sure we'll still be stuck at our respective null hypotheses. I've put down how I arrived at mine, how did you arrive at yours?
Yep. We know there's more at play than just weights as well. Possibilities include win record, lopsidedness, loose TH check, prioritizing equivalent match at the highest TH. As to the specifics we're just guessing.Quote:
Originally Posted by iBork
Will do, can't right now though as we've all been farming hard since the release and I can't view the past wars anymore.Quote:
Originally Posted by TankSinatra
I'll definitely check this after our next war though.
Thx for the clarification again.
Did you forget flags?Quote:
Originally Posted by TankSinatra
If you send clan name or tag someone Im sure would pop over to explain how to find it on a scout for friendly challenge.Quote:
Originally Posted by BlindSyphon
Quite possible. Base 2 had a 2 star th 12 while base 1 had a 1 star th 12.Quote:
Originally Posted by OnyxDS
I'm wondering whether we're a little hung up on the storage idea because it seemed to work in the past?
It looks like historically (before SC started tweaking with the MM algorithm) the gold storage weight was a useful measure for the effect of certain defences on matchmaking. We used it to try and manipulate the MM, and it seemed to work. So we've assumed that that's what its purpose was...
Could it be that the storage weight was always just there as part of the war loot calculation, and it just happened to be useful for the other? That SC always had a different weighting system for MM, and that it just seemed to match with storages?
And now that SC has started tweaking the MM algorithm more, the storage weight no longer matches? Not that they've changed things to be sneaky, but that's just how it always was... and we didn't know?
Quite possible N1NJ4, but we are still guessing.Quote:
Originally Posted by N1NJ4smurf
Lets not forget that before March 2016, a significant proportion of the community believed that Gold storage was your War Weight total (including offence), including prominent Youtubers, and numerous (but probably not all) prominent engineers.:)
Gold weights are accurate and a factor within war, you can check it during war day or through scouting a friendly challenge.
They also account for your offense weights, which you can see as your recommended target before any bases have been attacked on war day.
Server side the MM breaks the lineup into lanes and tries to match the lanes together with weights of offense and defense (1-5, 6-10, 11-15 etc). This in itself is not proven but was an idea floated around and has stuck for me.
Why it may not make sense to you that the gold weight of each clan does not match up is that the matchmaker weighs offense quite heavily, especially the warden.
I would hazard that although the MM does not match town halls directly, it will work to arrive at a close approximation.
Our current match we have 2 TH12 vs 1, and 11 TH11 vs 8 but they have 3 TH10s where we have none, and 1 more TH9 than us in a 25 man war.
Gold weight is solely for defense structure, not including offense at all, if any change can happen on hero lvl increase, because hero has both def and off weight, so at certainlvl heroes weight increase becoming visible in gold weight because of heroes defensive weight, this is easily seen in defless acct, in normal acct, maybe not as visible as thatQuote:
Originally Posted by Skrags
Quote:
Originally Posted by MajorJohnson
Take note N1NJ4 mate. If these 2 guys say its relevant, then chances are it is.Quote:
Originally Posted by samratulangi
On the chance that they aren't true weights, it can still be used as a guide, in order to plot your defensive upgrades. If you like to optimise, which applies to most of the more serious war players, then you need a tool with which to work with. Gold storage and rec target are as good as anything out there. What I would be more wary of, is the 3rd party sites out there, that try to sell the idea of having perfect scoring, for every level of every piece.
Of course, once you know that point, air, and hidden, give good protection for low weight, splash is heavier, and the marquee level defences are very heavy, then you have a pre ordained upgrade path that you can use anyway.
Dependant on your clans needs, simply do air/point, then onto the next. So far as offence goes, concentrate only on troops you war with. Heroes are likely to weigh quite heavy, as they score on both sides, so for optimisation, have them at a level they can do their job. If you need a bigger impact, upgrade them, if they are working, they are big enough. Optimising within game is about gaining best outcome for the lowest weight, and with the game changing shape, then so will optimum, but the basic key ingredients will remain constant until they directly change them.
Interesting responses, thanks... I honestly wouldn't have a clue, but the thought occurred to me so thought I'd see what you guys think.
I've always tended to go point def first, splash later, IT/EA later again, but not to the point of truly being optimised. Wouldn't make much difference in my clan cos we're all over the place when it comes to balance, and I'm not serious enough about war to move on to a hardcore war clan.
Do we know whether the weight of IT/EA (and even xbow) drops off as they get upgraded, compared to the original weight of building it? Or do the additional levels add as much weight as the first?
most def has highest weight when dropped, and subsequential lvl weight increase are lower, but we dont know for th12, since it is new, and lvl16 point def increase the dps much more then few previous lvlQuote:
Originally Posted by N1NJ4smurf
Yes, Sam. I realise this, thanks!Quote:
Originally Posted by samratulangi
The point about gold weight was that BEFORE March 2016, a lot of players didn't.:)
But if gold weights are an accurate measure of what the server thinks the defences are, why do I see, in wars where one clan is much more optimised than the other, that the gold weights differ - always with the more optimised clan having a lower gold weight. But, often, with the actual usefulness of the defences looking rather closer.Quote:
Originally Posted by MajorJohnson
So, for example, a clan that's missing lots of eagles and infernos, seems to have lower gold numbers than their opponent. And they have less eagles and infernos, of course. But more of the other defences.
If the gold was the measure of defence weight, you'd expect it to match, on average, yes?
The actual number is not important, what's important is the relative defensive weight of each base. And I have not seen a single instance in which a stronger base had less gold than a weaker base, whereas in every single instance I've seen weaker bases either have the same amount of gold (if they're close in strength) or less gold more often. It looks the exact same to me as it has in the past.Quote:
Originally Posted by Vikingchief
I agree that it looks the same as the past, but SC does not measure relative weight. Optimizers measure and build according to relative weight. SC ranks on absolute weight. This is why 9.11 with max point rank lower than th10’s, but defend better.Quote:
Originally Posted by MossackFonseca
The rare times I calc it out (because I suspect a weights are way off & want to understand the mismatch) I find overall defense weights close.Quote:
Originally Posted by OnyxDS
Which is funny because I expect lighter weight on the optimizer's side, and it doesn't surprise me that you're seeing it. If matching is primarily based on storage/rec target weights, plus loose TH matching, then the optimizer gives the MM a problem - a match that looks good to the MM both on weights AND THs at the same time might be rare. Too much daylight between the two (your TH makeup and your weights, think defenseless or th8 weight th11s) and there will be no match possible, spin on Blombardo. In more moderate cases the MM stretches weight and TH tolerances to allow the match - so the optimizer gets a bit of a TH edge, maxer clan heavier defense.
I don't understand what you're getting at here. The hypothesis is that the map rank still uses the same calculation as the gold weight. But the war MM uses sensible weights for guns. Therefore map rank and gold will correlate. The question is whether either of them reveal anything at all about how the MM is working (nowadays)Quote:
Originally Posted by MossackFonseca
If by "stronger" you mean actually stronger in practice, just look at any rushed base with an eagle. Ranked well above any TH10 but weaker. Or any max 9.11 (non-inferno) - ranked a middling TH10 IIRC, but easily stronger than a max TH10 in practice
Our current war we have the same amount of eagles and infernos as the other side as well as better defense and heroes at the top so we would have a defense, offense and town hall advantage.Quote:
Originally Posted by OnyxDS
You would not expect it to match defense weight on average because you are taking into account offense weight too, across the entire roster. One clan can have a bit higher gold weight than the other where the other clan have a little higher offense weight than the other.
Clan A could have large heroes, offense and defense in key places, and lower elsewhere, where as Clan B could have low medium heroes, offense, and defense across more of the roster. Clan A could still have lower gold weight and offense weight than clan B.
I rarely bother calculating all our offense and defense weights except when creating our lineups and that is where I am envisioning which bases will be taking what part of the enemy roster within a degree and who would be defending.
the big question for me all this aside, and i agree the somewhat hidden offence makes it hard to prove gold weight isnt the actual war weight, and now comes the question,Quote:
Originally Posted by MajorJohnson
-as we now know unplaced building weight doesnt effect war gold weight but has been confirmed as part of defence weight (so is server side) isnt it logical to assume there would be further weights done server side?
I've found there to be pretty strict townhall matching as of late maybe that's why searches taking so much longer. Were always within 1-2 on each townhall count
But the question is, and will remain hypothetical, as there is no way of knowing for certain, without confirmation from SC. A case could be put forward, that by buying a pack, check for any gold increase, then placing it, and checking again, to confirm a rise (gain) in gold storage weight. This can be argued away, as simply an increase in win bonus, due to the base being harder to hit (although I think if the increase becomes evident even before it is placed, it would appear to be a more likely true weight). On the other hand, trying to prove SC use a different, hidden weight, is almost impossible, as there are so many unknowns built into the MMA, that pin pointing anything specifically to a hidden defensive weight, will always be pure speculation.
I think the best way to test this would be to look thru some samples.Quote:
Originally Posted by Vikingchief
Anyone having:
a) an "equal" eyeball match but gold math is significantly off
Or
b) the gold weights in a match where a clan used th packs but never placed buildings.
Matches seem to be improved over past few months for my clan - with closer th matching as other pointed out. I would even presume gold totals are close between clans. But we do run X.5s, so I will take a look at recent matches.
Until I see data, I will not jump to conclusions...TBF my clan is more rushed than optimized so I don't particularly care anymore. But I do enjoy these posts/discussions.
Why would there be a need to show what ratio the weights are handled in? Or like what parts of the lineup should be heavier than others. We all assume you can't match only TH12s and TH3s, which means the roster itself is weighted, and there is no reason to show that either.Quote:
Originally Posted by Vikingchief
RT and gold weight are only a part of the equation and it does show us that so we can builder build our offense and defenses.
I don't do the calculator stuff either any more. But reading the counter arguments has whetted my appetite.
I feel it is impossible to prove hidden weights, so in order to try for proof, we can only either prove or disprove the weights they show.
Heaviest base is always top, and is affected by anything on the map. I tend to feel, unless I am missing something, that if something is purchased but not placed, and the shown gold weight increases anyway, that there must be more to the "war map gold" than purely win bonus, as it would hold no value until it is placed, thus making the base harder to hit. If there is no increase until it is placed, then it is purely a bonus payment. I have reservations about SC doing anything to simply throw optimisers off the scent. I doubt if they would go that far, they seem to be more blunt with their actions, rather than a stealth move.
Has anyone actually seen a case where an unplaced eagle upgrades to TH12, and has the weight added? Not that I don’t trust SC, but for 2+ years they haven’t been able to fix weights, and now suddenly they anticipate a relatively seldom used work around, while taking the hammer approach for all other TH’s. Easier to say “oh we thought of that, and storage is not linked, but trust us, we have fixed this”.Quote:
Originally Posted by joshsgrandad
I think it would have been better to just require people to place them before starting TH12 than to make it add weight even without it placed because there is no official notice of that change other than a couple posts buried in the masses. Lot's of people just have them, not to use it to gain some advantage and unknowingly keep them unplaced to their detriment.Quote:
Originally Posted by joshsgrandad
The amount of people that would have the TH12 pack would have been pretty minimal anyways though I do understand why they would try to close it.
I agree with you both. They should be placed before moving on. Better still, anything purchased should automatically be placed on the map. And I get where you are coming from ripple. They are never subtle. tend to stomp about in a pair of big boots, tripping over their laces. Stealth and forethought have never been big on their agenda. They certainly are making hard work of the latest fix. Seems as tho the MMA needs loosening a little, to get people back into play. People at both ends of the scale are suffering at the moment.
What's the meaning of gold weights then? Misleading?Quote:
Originally Posted by 2222
I don't think it should be automatically placed on the map but require it to be placed before going to the next town hall level, or atleast TH12. It doesn't make much difference if you are dropping any level of eagle during prep day, it is more likely you would win. If you drop an eagle early, or have it placed on the map for you, then you are at a disadvantage until the rest of your base catches up.Quote:
Originally Posted by joshsgrandad
The work around would end up applying to new TH11s, where they would buy it, and later on go to TH12 without the eagle. It would not be immediately noticeable but early TH11s would have a clear advantage by not having to drop eagle to go to TH12 later on and could become the norm of progression that existing 11s don't have access to.Quote:
Originally Posted by Ripple1972
there wouldn't be a way to know...Quote:
Originally Posted by Ripple1972
if you didn't place it then it wouldn't be on your map to add to your storage totals and therefore there is no physical difference between having it or not having it.
the only way would be to estimate if your opponents are consistently having extra eagles which could be happening anyway just because of higher heroes or weight made up elsewhere.
This is what I'm doing and its likely to be inconclusive. We've got one 9.11 with unplaced but purchased eagle. War #1 two eagles on the other side - not common but not unheard of. War #2 no eagles other side - again not common but not unheard of. War #3 2 eagles... Kinda sorta suspect the phantom weight is there but don't expect to see results that rise to the level of 'proof'. Especially since we're likely to see big population changes affecting the pool, bunch of defenses dropping on heavy engie bases for players that want to see the end content.Quote:
Originally Posted by AFrostyBlaze
Tank must be aware when were the eagle dropped, 2 wars ago my opponent dropped 2 eagles on their top 2 bases in prep day, in was not there in the first hour we got we matchQuote:
Originally Posted by TankSinatra
It once was the weight used for defense in the matchmaker. It still has some meaning as it tends to show the defensive strength of the bases in order, although as has been discussed above, because of its shortcomings it doesn’t do a good job of that with some types of bases (such as a 9.11 ranking lower in gold weight but defending better than bases above it). The question is whether SC still uses those gold weights or now has a different set of weights and just hasn’t told anyone. I don’t think SC should tell us what the weights are, but I do think SC should tell us if the gold weights are no longer being used by the matchmaker.Quote:
Originally Posted by ExpertSAQ
Well darian did tell us and therefore imply they arent using it by stating that the unplaced buildings are accounted for in the matchmaking.Quote:
Originally Posted by 2222
They wont show up in the gold total because without placing them the difficulty of the base will not increase but they are "allegedly" being accounted for.
The gold storage is useful for showing the defensive strength of the base but not for the amount of war weight a base contributes to a given war
Interesting bits straight from the source, which I don't think anyone's been looking at. Updated 90 days ago - https://supercell.helpshift.com/a/cl...p-in-clan-wars
"Town Hall levels aren't a deciding factor in matchmaking either." There goes my loose TH matching, unless they're being a little loose with the language and its a 'contributing' rather than 'deciding' factor. Direct contradiction to Darian's statement a while back about the MM trying hard to match THs.
"Matchmaking aims to match Clans with similar ratios of offensive and defensive power against each other." Very interesting. Lopsided vs lopsided, even vs even? Not sure that's what we're seeing, though it could be more of a tendency than a rule. Explains the Blombardo effect though (clan so heavily lopsided no match could be found).
ETA something similar has been hinted at/theorized, lopsided penalty etc., but don't think I've seen "ratio" as the key metric before.
"Finally, matchmaking will factor in Win/Loss records." I take that as confirmation its not specifically 'win streak' as a criteria, though I'm sure not everyone will see it that way.
ive always thought the appearance of townhall matching was done by a relevant proxy like presence of warden, presence of 5 spells, presence of archer Queen etc. Makes more sense and would provide a better foundation and still provide the appearance of loose townhall matching which helps reduce the appearance of mismatches.Quote:
Originally Posted by TankSinatra
agree however deciding factor/ contributing factor language provides uncertainty.
They didn’t draft it very well. On the one hand, the entire context of it suggests hall level isn’t a factor at all. However, that one sentence right after saying trophies don’t matter they say hall level isn’t a “deciding factor.” If hall level wasn’t a factor at all, you would just say hall level isn’t a factor. The addition of the qualifier “deciding” suggests hall level actually IS a factor, but isn’t an absolute deciding requirement.Quote:
Originally Posted by TankSinatra
My age old hypothesis is rescued by 2222 and Viking! Though perhaps on life support. Oh well whether true or not, its a significant part of our streak. Which we'll probably lose shortly. But not this day....Quote:
Originally Posted by 2222
Good catch Tank. I hadn't noticed that. As stated above, it's strangely worded.:straightface:Quote:
Originally Posted by TankSinatra
I think that "deciding" could mean anything or nothing. Indeed if TH played no part at all they could have omitted that word. However they might have been aware that TH12 was coming, where it is a sort of factor as it's a defence. They might have considered that TH correlates to strength indirectly, so many wars do sort of match TH to TH, because things like spell and camp upgrades push up offence scores. Or they might just have been waffling. So I think it's hard to conclude that TH does play some direct part in matching from that. All it really rules out is that TH is an overwhelming factor - which is obvious anyway from the matches where the THs are rather different each side.Quote:
Originally Posted by 2222
Seems to suggest they mainly look at defence and offence weights, and try and match them. In so doing, loose hall matching would be a natural by product anyway, since both clans would need to be within the given tolerance.
The win/loss record strikes me as odd. I feel that is their attempt at recent war performance, but to say win/loss doesn't fit in with what we see. I feel some of it has been lost in translation, but if they do, indeed, try to match both off and def, as separate entities, and tightened the screw on both, then it would explain a lot of the non matching.
Yes, all of that plays into why it isn’t drafted well. In the US one of the rules of interpreting laws, statutes, etc. is to try to give meaning to every word. I realize this is not a statute, but I’m just saying in this case they added in a word “deciding” that has no need to be there unless it has meaning. If they wanted it to say hall level isn’t a factor at all, they could have said it like that or even just said it isn’t a factor. To say it isn’t a deciding factor suggests the word “deciding” has some importance. I agree it is hard to conclude what it means, hence my note that it wasn’t drafted well.Quote:
Originally Posted by OnyxDS
War #4 2 eagles. Once is happenstance, twice is coincidence, three times is enemy action. Big Steve's going to 12, I'll probably follow with an account or two though it'll take a long time.Quote:
Originally Posted by TankSinatra
Streak's going to be impossible to keep but I've been expecting it to end any day for a while now.
you can do it, just gotta believe!Quote:
Originally Posted by TankSinatra
and maybe hope the other clan stinks a bit
To confirm have you got any eagles your side? If you weren't facing eagles before then it's a fair shout Darian wasn't joking about the unplaced eagle weight.Quote:
Originally Posted by TankSinatra
One unplaced eagle from a bundle, none on the map. Prior to the update we'd frequently see one eagle on the other side, and on rare occasions two. Seems we draw two now.Quote:
Originally Posted by TheGame1
This isn't proof of course but increases the likelihood in my mind that the phantom weight scenario is in play, and it isn't confined to th12.
They usually do but the luck will run out sooner or later. Frankly it'll be a bit of a relief.Quote:
Originally Posted by AFrostyBlaze
SC has said that unplaced eagle counts in the weight for the war matchmaker. They haven’t said unpurchased/built defenses have a phantom weight, but they specifically said the one you own but haven’t placed is now counting in the weight.Quote:
Originally Posted by TankSinatra
However I believe a few players have confirmed the unplaced defence isn't showing up in gold weights at all. Hence the "phantom weight" phrase I think - unless SC are lying (I really doubt it) or just mistaken (possible, but unlikely) it means the unplaced gun is counting to some other weight, not related to gold stores at all.Quote:
Originally Posted by 2222
C’mon, are we to believe that an organization that couldn’t fix their mm algorithm for 2 plus years, designed their stopping engineering by mandating a build (not adjusting weights), foresaw the buy but don’t place eagle th11 path and changed weights. That is a stretch to me. Maybe I can interest SC in a bridge I own In Brooklyn, they can pay me for it, and I’ll place it for them.Quote:
Originally Posted by OnyxDS
Initially I used to think that it's our offence vs their defence kind of matching but with the recent wars played in the last couple of weeks I too feel there is something like phantom weight ruining our matches.
Here is our roster:
1. Maxed th8
2. Th8 without pekka and bomb tower, rest of the def buildings and war troops are maxed.
3. Th8 with th7 def and off ( no dragon or hog, only level 4 loon and rage for war)
4. Th6 (healing spell unlicked, only the first 3 troops unlocked, maxed def)
5. Th6 ( no spell/ level 3 archers and giants for war)
6. Same as 5th base
7. Th6 ( only point def placed/ level 2 giants and level 3 archers for war)
8. Same as 7th base
9. Th6 ( only point def placed/ 20 barbarians for scouting in war)
10. Th5 ( point def/ giants and barch for war)
Currently we are at a win streak of 45 ( could have been 60+ if I hadn't forgotten to attack!). We usually match with 2 th8s and rest high to mid th6s with a couple of th5s or/and 4s down below. With maxed cc troops it's easy for me to score 30 but I find it strange those clans who score 30 again and again to match us fail to score even 10 against us! If I attack late they manage to score as high as 20 stars.
Well after the xp event ended I took a break so cannot comment on the recent changes. But our last 7-8 matches were pretty tedious on preparation day. I was almost at the same situation as Tank, I pretty much wanted the streak to be broken but as it didn't happen, I am taking a break for a while and gonna concentrate on my main bases in the mean time.
It doesn’t surprise me at all that a clan full of th5-6s score only 10, often you are likely facing little kids. I’m not saying that is always the case, but it often is. Even when they aren’t little kids, they are brand new. Not many veteran players want to play a war that is easy to 3 star everything like that, which it would be for anyone who spent some time playing this game.Quote:
Originally Posted by Bishalb1
My feeling is If there is 1 server side weight we know about, and spanser even hinted long ago their were server side adjustments, and skrags point that supercell said they are tweaking weights without seeing gold values change, and your points in your thread, and the fact that it makes sense to hide the weights from engineers (technical and business), Its getting very hard to believe gold weight is anything more than a way to assign loot in wars.Quote:
Originally Posted by OnyxDS
We had another match last night with quite a big difference in the "heavy" stuff: 0 vs 2 TH12s; 2 vs 6 eagles. They also have something like 6 extra wardens. We probably have better BK/AQ, although they do have some good ones. And although their #6 is rushed (I mean REALLY rushed) it will still outweight our #6 in gold I imagine. So we're likely to be something like 50k down again on the top few bases. Yet the defences are probably reasonably equal
Incidentally, it's also a good example of recent war performance, which was mentioned in the other thread. This enemy has a public log. It's 14 streak to us, 0 to them. 47-3 current log to us, 29-21 to them. And given that the match on strength could politely be termed "loose", it doesn't look like the server is trying very hard to match recent war performance. (Unless it's something odd like average percentage destruction)
It could be giving them that extra weight to make up for being down in the recent war performance factor.Quote:
Originally Posted by OnyxDS
Actual SC comments indicate the win/loss record is an attempt to match like to like (better clans vs better clans, etc) rather than a weight penalty. That jibes with my experience, I don't see any kind of persistent weight penalty against me when on a streak.Quote:
Originally Posted by 2222
PS does everyone have apologies drafted for clashkings? Discrepancy between storage weight and their reported weight was the key argument in telling them they're full of it.
not yet but Ive flagged this item as important on my to do list.Quote:
Originally Posted by TankSinatra
No I think it was their claims about flags, claims of super-secret information SC was giving them etc that was the main reason for saying they're full of it.Quote:
Originally Posted by TankSinatra
It seems SC has fiddled with the spAnser numbers in the client file - the ones whose meaning is unknown: https://forum.supercell.com/showthre...nses-Decreased
And looking at CK calculator, they've faithfully replicated these into their calculator. So it now has a max DE store at more defence weight than a max eagle AND max giga tesla combined. Closely followed by DE drills weighing the same as inferno towers. :saywhat:
It's the gift that just keeps giving...
Just came across a war where a base weighing 114k is placed below a base weighing 109k
https://s22.postimg.cc/9nhqzkmc1/image.png
https://s22.postimg.cc/d73opermp/image.png
https://s22.postimg.cc/ygrb0bksx/image.png
https://s22.postimg.cc/wc6xz8ylt/101.png
That can happen if someone upgrades during prep. My opponent have one of those once my war starts: #1-#4 were all max TH11s when it was spun. Within minutes of it matching #3 had booked/gemmed his TH to 12. So he'll be higher weight than 1 or 2 on battle day.Quote:
Originally Posted by Rak2180
Incidentally, did anyone notice Darian, in the AMA, confirmed that specifically weights have changed:
(and not just other parameters). And yet there's been no change in the gold weight of individual bases.Quote:
Yes we have made a few changes to the weights in the matchmaking system
The prosection rests, m'lud.