Collection power aims to solve a few Splinterlands problems and allow the creation of real separate leagues with leaderboards. Currently, it only evaluates the power of the collection, it is not used in any way.
The current implementation of Collection Power is the sum of the burn values of all playable cards. There were other ideas, but right now this is the way to calculate it.
One of the biggest Splinterlands problems is the fact that low-level decks reach the high leagues and get disproportionate rewards in relation to the owned collection. The chart below seems to confirm this, half of the accounts reach the diamond league or higher. And over 2000 accounts reach the champion league.
Another problem is that the current leagues don't make much sense. For example, if you have a silver level account, you cannot compete for the title of the best silver player (excluding tournaments). Instead of fighting with the Silver players to see who is the best Silver player, you reach the Diamond or Gold league, while in the silver league there are almost only 1lvl decks. Collection power is the first step in creating real leagues that can significantly improve the Splinterlands ranking system.
Let's analyze Collection Power in all leagues, but first, take a look at top100 accounts with the highest Collection Power.
| . | PLAYER | Coll. power |
|---|---|---|
| 1 | endgegner | 31078406 |
| 2 | tendershepard | 19949492 |
| 3 | michealb | 18525226 |
| 4 | neoxian | 18026196 |
| 5 | digital-wisdom | 14371301 |
| 6 | cryptomancer | 14201395 |
| 7 | jaki01 | 14018971 |
| 8 | fighter4-freedom | 13622289 |
| 9 | glory7 | 13238270 |
| 10 | marianaemilia | 13233258 |
| 11 | wer-gewinnt | 12135524 |
| 12 | ula.ocean | 12088900 |
| 13 | angstgegner | 11992590 |
| 14 | jacekw | 11376873 |
| 15 | byzantinist | 11319460 |
| 16 | cyguy | 11216398 |
| 17 | gringotts | 10899192 |
| 18 | imperfect-one | 10559302 |
| 19 | bji1203 | 10528487 |
| 20 | goodasgold | 10433266 |
| 21 | clove71 | 10062256 |
| 22 | goldmatters | 9811750 |
| 23 | byzantinekitty | 9472930 |
| 24 | mellofello | 9183637 |
| 25 | matt-dub | 8977700 |
| 26 | nealmcspadden | 8263854 |
| 27 | coolbowser | 7756865 |
| 28 | raynie | 7702277 |
| 29 | jesterago | 7331129 |
| 30 | toocurious | 7148149 |
| 31 | noisme | 7037995 |
| 32 | thebadwitch | 6882429 |
| 33 | fw206 | 6748015 |
| 34 | gfriend96 | 6618163 |
| 35 | uwelang | 6567004 |
| 36 | conan123 | 6365310 |
| 37 | cryptkeeper17 | 6032483 |
| 38 | dhimmel | 5758712 |
| 39 | exyle | 5745834 |
| 40 | kick-ass | 5630987 |
| 41 | goldlegend | 5565620 |
| 42 | tradingideas | 5341681 |
| 43 | naturalkiller | 5205745 |
| 44 | bubke | 5178573 |
| 45 | steemgoat | 5128842 |
| 46 | mattclarke | 5125782 |
| 47 | ccoin | 4964996 |
| 48 | epicteam | 4903090 |
| 49 | valnamor | 4770601 |
| 50 | themightyvolcano | 4736534 |
| 51 | cjturtleman | 4714692 |
| 52 | codebull | 4612983 |
| 53 | foggybottom | 4565548 |
| 54 | zagorlord | 4390115 |
| 55 | warrentrx | 4290343 |
| 56 | rawutah | 4287382 |
| 57 | goldninja | 4079940 |
| 58 | kryptodenno | 4043701 |
| 59 | freedomfighter4 | 4004345 |
| 60 | davemccoy | 3971011 |
| 61 | tailcock | 3893246 |
| 62 | ts-warrior | 3886860 |
| 63 | redundancy4040 | 3882193 |
| 64 | newageinv | 3876247 |
| 65 | forever-crypto | 3864497 |
| 66 | victor-alexander | 3858515 |
| 67 | imperij4427 | 3835174 |
| 68 | transcript-junky | 3655377 |
| 69 | faiyazmahmud | 3599985 |
| 70 | myfriend | 3584957 |
| 71 | purplegoblin | 3579183 |
| 72 | dkid14 | 3345507 |
| 73 | twinner | 3337905 |
| 74 | masterthematrix | 3177894 |
| 75 | back-off | 3162000 |
| 76 | begonethot | 3125114 |
| 77 | steemitromney | 3101672 |
| 78 | spirits4you | 3063052 |
| 79 | ninjascott | 3025969 |
| 80 | vettev | 3007679 |
| 81 | ty2nicerva | 2971466 |
| 82 | nateaguila | 2960993 |
| 83 | mmckersie | 2941906 |
| 84 | notsoperfect | 2921460 |
| 85 | stiant | 2909293 |
| 86 | udow | 2877917 |
| 87 | agami-heron | 2869686 |
| 88 | demon402 | 2864595 |
| 89 | giker | 2836932 |
| 90 | vulcanator | 2803727 |
| 91 | swedishdragon76 | 2803230 |
| 92 | photohunt | 2747290 |
| 93 | truskawka | 2746213 |
| 94 | louis88 | 2742256 |
| 95 | schneegecko | 2719119 |
| 96 | maybenexttime | 2711743 |
| 97 | eatmyshorts | 2683535 |
| 98 | joshman | 2650803 |
| 99 | monstermother | 2649987 |
| 100 | sci-populi | 2637302 |
Champion 1
There are some outliers that make this histogram less readable. Let's just remove them (top 4 accounts). Now chart looks a bit better. (Outliers will be removed from all next histograms too).
| . | . |
|---|---|
| Number of accounts | 449 |
| Maximum | 31078406 |
| Average | 2159487 |
| Median | 1041015 |
Median is much lower than average, but just looking at the chart it's no wonder, most of the values are on the left side of the histogram.
Champion 2
| . | . |
|---|---|
| Number of accounts | 907 |
| Maximum | 9022859 |
| Average | 353971 |
| Median | 190460 |
Champion 3
| . | . |
|---|---|
| Number of accounts | 703 |
| Maximum | 7429605 |
| Average | 138494 |
| Median | 66740 |
Diamond 1
| . | . |
|---|---|
| Number of accounts | 381 |
| Maximum | 1575883 |
| Average | 90189 |
| Median | 37375 |
Diamond 2
| . | . |
|---|---|
| Number of accounts | 533 |
| Maximum | 1399732 |
| Average | 61409 |
| Median | 21605 |
Diamond 3
| . | . |
|---|---|
| Number of accounts | 942 |
| Maximum | 1055018 |
| Average | 31185 |
| Median | 12005 |
Gold 1
| . | . |
|---|---|
| Number of accounts | 1042 |
| Maximum | 2402000 |
| Average | 27583 |
| Median | 11557 |
Gold 2
| . | . |
|---|---|
| Number of accounts | 562 |
| Maximum | 1257730 |
| Average | 22766 |
| Median | 6585 |
Gold 3
| . | . |
|---|---|
| Number of accounts | 266 |
| Maximum | 513655 |
| Average | 29424 |
| Median | 8537 |
Silver 1
| . | . |
|---|---|
| Number of accounts | 172 |
| Maximum | 1188370 |
| Average | 50410 |
| Median | 10447 |
Silver 2
| . | . |
|---|---|
| Number of accounts | 132 |
| Maximum | 2262269 |
| Average | 51557 |
| Median | 9340 |
Silver 3
| . | . |
|---|---|
| Number of accounts | 96 |
| Maximum | 3023605 |
| Average | 108441 |
| Median | 9560 |
Bronze 1
| . | . |
|---|---|
| Number of accounts | 149 |
| Maximum | 1988755 |
| Average | 60147 |
| Median | 3115 |
Bronze 2
| . | . |
|---|---|
| Number of accounts | 465 |
| Maximum | 698138 |
| Average | 9123 |
| Median | 245 |
Bronze 3
| . | . |
|---|---|
| Number of accounts | 697 |
| Maximum | 1138375 |
| Average | 11627 |
| Median | 255 |
Novice
| . | . |
|---|---|
| Number of accounts | 204 |
| Maximum | 892780 |
| Average | 5292 |
| Median | 0 |
Novice league seems to be a combination of new accounts and inactive accounts (with some cards).
Bonus
Probably not very useful, but I was just curious if there is a visible correlation.
My opinion
I am a big fan of Collection Power and separate leagues. However, I am not convinced that the current implementation of Collection Power is optimal.
In my opinion, it would be better to calculate Collection Power this way:
- ignore foil (gold foil Lyanna 8lvl has the same power as regular Lyanna 8lvl, there is no difference in battle)
- treat Alpha / Beta the same way (8lvl Alpha Lyanna does not have an advantage in battle over 8lvl Beta Lyanna), maybe treat all editions the same way (Alpha / Beta / Promo)
- ignore duplicates (if you have 10 maxed out Gremlin Blasters it does not give more power to your collection than having single maxed out Gremlin Blaster ;-) )
- maybe ignore Reward cards (anti-bot, anti-farm way)
- maybe take into account only cards with cooldown (if you are not using a specific card at all it does not add any power to your collection)
Another idea is to calculate collection power based on some formula including bcx / rarity and not burn value.
In my opinion, Collection Power should be measure of the combat strength of the collection rather than the value ($) of the collection.
Very useful information and statistics, thanks! I want to take a minute to talk about how collection power is being calculated and why.
The goal of the collection power is very simple - to require players to have a certain amount or value of cards in order to be able to gain higher levels of rewards. It is very similar to a staking system - the more tokens you have, the more you can earn. With this in mind, it doesn't really matter if you have 1000 Rusty Androids, or 2 GFLs, or whatever - you have cards, you can earn rewards. Additionally, it makes sense that gold foil cards and alpha cards should be worth more as far as earning rewards is concerned.
After we first announced the plans for collection power, a number of people said things similar to what you are saying here - that the collection power should be calculated more as a strength of the collection in battles rather than just a simple value. This then caused about a 2 month delay in getting the collection power change released as we worked on trying to accomplish that.
We built three different versions for calculating collection power, very similar to what you proposed, and in all cases there were many situations where it didn't really reflect the "power" of the collection in battles and there were numerous other issues that arose from it. Most importantly, though, is that these methods of calculating the power are MUCH more complicated to calculate and the system needs to be able to quickly keep track of the collection power on the fly every time a card is created, transferred, delegated, or updated.
In the end, I decided that it doesn't work to try to make collection power solve two very different problems, so we should stick with the initial, simple goal and intention of collection power - to require players to have a certain amount or value of cards in order to be able to gain higher levels of rewards.
While the collection power will limit the max league that players can get into, the rest of the league leaderboard system can be handled separately, and I believe we have a pretty good initial plan for that should allow for skill to be the primary factor in the rankings at each league. That being said, we will always keep a close eye on how it goes and will listen to player feedback and make changes as necessary.
Leave it to you the experts to figure out what is best.
One thing i have noticed is i struggle each and every season to make it to the top 50.
Seems no matter how high my collection score gets - it still doesn't make me any smarter.😂
In the last season, there were 2 accounts in top10 with Collection Power below 2 million. So looks like it is enough for top competitive champion accounts. So what does it mean that account has Collection Power equal 5M or 10M? Does it make it any stronger? Well, no. The difference is only Gold Foil cards (and I think duplicates in the current implementation). So how to calculate thresholds for a given league if Collection Power does not clearly define the strength of the account? It's more about deck value (in $) right now, not about the power of collection. And I think the point is to make Silver decks competing in Silver league, Gold decks competing in Gold league. But if there would be enough low-level bots but with some Gold Foil cards, they would be able to play in higher leagues and win battles. So the problem won't be solved. (Yes, ofc it would be still much better than now).
Haha, you and me both!
lol
@jacekw Cool! I haven't been following the discussions about this on Discord, but from what I read, I agree with your conclusions, you must have already expressed your opinion to the developers, what did they answer? Thank you for the informations.
I was a bit disappointed when I find out that the current implementation is just based on burn value. Yes, it's a step it the right direction, but I don't think it's the optimal way as Collection Power does not say how strong deck is. There were 2 accounts in top10 with Collection Power below 2 million last season. And there are several accounts with Collection Power above 10 million, but that does not make them any stronger.
So I wanted to do some research and then share my feedback based on data. Also I want also to calculate Collection Power of all players but ignoring Foils / Edition to see what would be the difference.
It is not even burn value.
My account 'r1s2g3" has 282610 as DEC value and collection value of 249730
My account "saachi" has 255145 as DEC value and collection value of 261250.
So having higher burn value and higher level summoner than other account I have low collection score.
True, my other account does not appear on the top 100 list and was close to the top 20 of the season.
I completely agree with your proposal. It makes little sense to me that my (beta+Untamed/regular foil) deck is complete (except for PJ and Dricken) and maxed (except for maybe 2 cards and the new reward cards), yet still there are 98 people with a higher collection score....
Taking every playable card into consideration + giving alpha and gf a bonus doesn't provide the information that is important to break things up into leagues, imo...
Yes, that is exactly what I try to say!
The current implementation of Collection Power is more about monetary value of the collection and not deck strength in battles. And it does not lead to creating real separate leagues. Yeah, it solves some problems, but I don't think it's the optimal way.
If 2 million is enough to be in top10, what does it mean to have 10M or 15M? It does not make these decks any stronger. So how to build thresholds for leagues if Collection Power does not actually evaluate the strength of the deck. 1lvl deck with some gold foils (epics / legendaries) is not Silver/Gold level deck. It is still a 1lvl deck and shouldn't play in Silver / Gold league.
Exactly... I don't see how they'll be able to put up a fair system is they don't change the way collection power is calculated
I agree. One would think if a hypothetical deck had one of each card at the maximum level (without regard to alpha & gold), its collection power would set the top bar. This does not appear to be the case and it seems to be cumulative based upon the deck's monetary value. That offers no indication of the actual battle power of the deck. How many in the top 100 have ARCHMAGE ARIUS AND PRINCE JULIAN? There are only around 80 of each of them. I have both and 99% of the other cards in the game maxed out (without regard to alpha & gold), and I barely crack the top 100?
Yeah, I think thresholds should be based on X% of Collection Power for the full deck for a given league (some rare cards like Shin-Lo could be omitted, only a few people have them) with foils ignored. Right now it's more about monetary value not strength in battle like you said. Well, it still solves some problems, but I think it's just not the optimal way.
Collection power could potentially focus on different aspects. I think the priority should be 1 (and maybe 2, 3). And right now it's more like 4. But 4 does not lead to separate leagues (where Silver decks are competing in the Silver league to find out who is the best Silver player and so on).
(List borrowed from someone else)
Simplicity is important. Also the dopamine. We need to make sure pulling a gold foil stays exciting. A guy placing top 10 in silver league gets a couple of packs and flips a gold foil rare, quickly checks his updated collection power; holds his breath and clicks the 'Gold league for the rest of the season' button.
Yeah, I know that solution should be simple (not only for players but also for servers to not cause performance issues), but I think it's all about what specific problem Collection Power tries to solve?
It could be:
(List borrowed from Byz)
In my opinion, priority should be 1 (and maybe 2, 3). And right now it's more like 4. But 4 does not lead to separate leagues (where Silver decks are competing in the Silver league to find out who is the best Silver player and so on) and could be easily exploited just by renting some high DEC cards, that are cheap and not too useful.
If they're not useful,
I think it all depends how you frame it up.
"Cash out your winnings in cryptocurrency, or re-invest them to be competitive in a higher league", sounds much better than 'Pay to win'.
There are a lot of low levels decks that have a bad time in gold / diamond / champion league right now because of... oh wait, actually, they don't have a bad time because there are so many of them that they play against each other most of the time! :)
If I have Silver level deck I want to compete with other Silver level decks in Silver league, to find out who is the best Silver level player. I think it's better than the current situation when it's all about reaching as high league as possible. Except when you are a bot, then probably you want to farm as much as possible just by being in the high league and not having fun by competing with other Silver level players.
I'm a big fan of Collection Power, but I don't think the current implementation is the optimal way. My collection power is 11M, but almost 10M is just from gold cards. If gold foils are ignored it would be about 2.0M-2.5M total, but the strength of my deck would be the same.
but how can the little ones achieve much?
The most important thing is not the combat power of the deck, but the ability to make combinations of cards, which allows you to break through the ranking higher.
I think the goal is just to make real separate leagues when Silver decks are competing with other Silver decks to find out who is the best Silver level player. Right now it does not make sense, Silver level players reach Gold/Diamond/Champion and Silver league is full of 1lvl decks.
If some player can afford only Bronze level deck, he will be able to compete with other Bronze level decks. Right now the only thing he can do (except tournaments) is just trying to reach as high league as possible (but it's only possible because of bots / points inflation)
the distance between the leagues is too small, therefore, they reach higher leagues with small maps. If you change points for victory +2 for defeat -1, then just silver will compete with silver
there are also examples when a deck with 1 level of cards defeated a deck with 4 levels. which, in fact, just shows that small can achieve a lot.
although these are purely minor changes in the parameters of some cards between the first and fourth levels. but everyone forgot about it
What a great post on collection power!
Thanks for sharing! - @cornavirus
Wow those are really cool statistics that you got there jacekw! Hopefully the devs can use this as a guide to determining what the floor of each league can be.
Great offers. It is absolutely logical and expedient !!!
Great article! Sorry for my noobie question, but how do I figure out my Collection Power? Cheers!
Just go to the Collection tab: https://splinterlands.com/?p=collection&a=babalor
Not sure if it works on mobile though :)
Thanks buddy!
Super insightful! From the initial post @jacekw , all the replies, and @yabapmatt explanation. Seems beneficial to have TWO values: the existing Collection Power for quick calculation and ultimately reward ROI; and battle power which may need updating once a day or so to help ranked battles, perhaps tournaments, etc. This could be something Kiokizz, PeakMonsters, or MonsterMarket could probably do right now? At least to start tracking it. (don't intend to sign folks up, I just admittedly have no skills in this area)
I believe this discussion will keep coming up because a fun crypto/blockchain based game has two appeals: the gamer looking for a fun game, and the investor wanting ROI. The core player base seems to value both - and are being rewarded in both areas very well. However, mass adoption will require appealing to the former more than the latter, especially further out the ROI curve.