250 Battles, 245 against bots.
Prologue
I took the Survival challenge. Last season I thought about it, but didn't have the time as I the though came too late. Different this time! So, here's my resume of Survival Bracket 4, Gold level. 50 players in that group - but it seems that my gaming time does not coincide with theirs at all.
Of the last 100 battles, only 3 were against humans. And in total I think it was 5 or 6, so barely 2,5%. I know, the chances are low when it's around the world, and also the leaderboard is not very concentrated. 2,000 Rating Points are between me (currently sitting on no.1) and the number 10.
The Bots and me
Cards
Which leads to my main critique - the grind. Yes, the bots were able to kick my but around 20% of the time. But one could feel that it was very random. Not because of particular skill in combining the cards, just by what cards they had available. Sometimes they suddenly had all the bangers, and then they'd barely have reward cards.
My own deck is quite concentrated on the cards that I use most, and the combinations that I find useful. It's rare that I buy a new one, like recently the Darkguard Ironwing or the Conjurer's Hubris. From there, it's a matter of skill and combinations. Sometimes I face a ruleset and mana supply that I just can't deal with. Sometimes.
For the bots, it seems all the time. Which leads to a win-rate of 80% in Gold League, what my deck is built for. Which takes the excitement out of playing against Bots in Survival. When they win, I can't really learn something from it. It's more coincidence, luck.
Tactics
This is were the bots fail a lot. Even when they have the cards, they just throw in some really weird combinations. Like this one:
Giving the cat repair instead of Armored Strike and Piercing made them lose. Since the cat has True Strike, it would've put up quite the battle against my Commander.
But even weirder this one - not only because of the choice of summoner regarding the ruleset, but because they managed to win despite it. Very poor team by me.
The good part here - I was able to experiment a lot, which is visible in the fight. The bad part - I lost focused after winning to many times in a row, just grinding, not paying attention but wanting to get the 250 battles full.
Positioning was also questionable from time to time, like in this one where Mantaroth is placed in the last position, without any need for that.
If the bot had used the Warden or another Void Armor in first position, okay, then the Repair and tank in the rear with Shield would've made sense. But not like this.
The Settings
In the beginning, time-out for cards was reduced by 99%. Which is already pretty cool, as it takes a couple of hours for my lower level cards to be back. So, I did have to be careful with them, use defensive tactics, or blow out the cards that needed to die when they had the perfect setting (Arrowbait).
But then it was changed to -99.99% reduction, basically seconds. The the whole idea of the Survival mode was gone completely. I just played as I would in modern Gold League, and that was boring against bots.
The Brackets
That was what I loved most. I could finally play all the strengths of my deck, without having to hope that Dragon is available (Akane is my only max level Summoner). And that did help me training.
Lucky me, right now a Gold League Torunament started, and it seems like training with bots did pay off:
I was very outspoken for Brackets for a while now, as they're a lot of fun and reduce the pay2win, but enhance the skill needed to make it to the Leaderboard. Right now, I'm playing my deck between Diamond and Champions. I would probably try to keep it in Diamond, as I usually have more fun there. In Champions, I always get kicked around.
Conclusion
For Survival, it's a big no-no as it is. The bots are too erratic. I might try the same in Diamond next time, though I fear that I will get my but kicked there as my cards aren't good enough, and the bots will be a lot better.
But as test for brackets - gold star. If the idea is transferred to Modern, I'm all for it, after fine-tuning the bots. I read somewhere that the bots in Champions are cannon fodder for the big players. Anyway, I'm still liking the bracket idea, even more than before, having it tested.
Oh, what I did like is that most of the bots were marked as such, by "Scholar: Archmage.app" or something like that.
EDIT: @davemccoy forgot to tag you, now I did, as requested :-D
That's it! Thank you all for your attention! Any questions are welcome!
2% of the rewards go to @saydie for free Dividers - Click here. Always grateful for all of that! Remember to support those who give you cool stuff for free 😉
Former Calculation Posts
Calculations around Land
Land - Setting it up.
Calculating Land part 1 - How to get your ROI
Calculating Land part 2 - Are pools worth it?
Calculating Land part 3 - Organic Growth
Land Plot Efficiency for non-whales
Card-Production-Coefficient - Combining LPE and ROI
Renting for Land - at what cost?
What do refined resources cost?
Calculating Totem and Title boosts
Calculating Plot Rarity Boost
Setting up a Rare Stone Plot
My perfect plot - Conclusion of my calculations
Thank you very much for reading, I hope you enjoyed it at least as much as I enjoyed writing :-D If you have any comments or feedback - please let me know! And don't forget to leave your own posts for me to curate. Thank you very much!
If you don't play the game yet, you can use my referral: Click here for referral :-)
Dividers by Saydie - Click here. Always grateful for those!
Diamond bots are not easy once I got into Champion 3
Well I guess then they're Champion Bots :-D I don't play against bots, but keep it to humans. More interesting for me. Just did this for those who do, to keep it fair - I wouldn't like there to be an advantage by playing against bots because they're weaker than humans. Or the other way around. Balance is my credo! :-D
So what I think you are saying is that we shouldn't move forward with brackets in Ranked play because the bots make it boring for you?
I loved that you took the time for the analysis BZ, and I'd rather hear this and not start the Brackets than start the Brackets and then people complain because the bots are boring.
Thanks for the input, you are one of the few that did this.
ps... shared with @clayboyn for visibility
No, in modern it will be totally different, because there are more human players and I can turn the bots off for more fun. Which I'm currently doing, I'm not playing against bots in modern at all, turned that off. I'd adjust the difficulty though, working on better combinations.
I'm guessing here, and I hope I'm not trying to re-invent the wheel, but bots usually copy the data they get from other players, from matches all throughout SPL, the winning teams. They try to get the best combination with the cards available. So, the fun would be in the combination of cards that the bots can play. Not frontier summoners, but something balanced. For example, in Gold Tier it could be:
all common summoners available for the bot to play
all rare summoners available
(we don't have epic summoners in the moment I believe)
50% legendary summoners available, different for each bot, maybe randomized
Same for the monsters. All common, all rare, 75% of the epics, 50% of the legends, also randomized. With that set, the bot builds the best team possible based on the data they have.
That way, the bots won't be overpowered, but it would still be interesting to play them as they will always choose the best team possible. It could also be divided still in Gold 1, 2, 3, with the bots getting more and better cards in each.
For example:
Gold 1: 100% common, 75% rare, 50% epic, 25% Legends available for the bot to play
Gold 2: 100/100/75/50
Gold 3: 100/100/100/75
This way, even in Bronze it would be challenging to play in the third Tier. Players won't need the cards at high level, but they need a wide range of cards or the best cards. When they then decide to step up to Silver, they'd have the same. Also, it would still be possible to play a great and broad Bronze deck in Silver - but you probably won't get higher than Tier 2. And so on.
Thanks again for the feedback BZ
View more
Just to be clear, in case it wasn't... the entire point of this current exercise is to test the liquidity bots. You pretty much shouldn't be seeing a human when testing in survival modern manual brackets right now. I get not everyone wants to play against bot, and in modern brackets you could always turn off liquidity bots like some people do now, but the literal entire point of the current exercise is to test and dial in the liquidity bots. Just making sure to set expectations. You shouldn't be looking for or expecting matches against human players in the survival test brackets right now.
Oooh okay, that's something that I didn't understand well, sorry. How are the bots currently working? I commented on @davemccoy's comment an idea on how they could work better, and not as erratic (here's the link to the comment). @saydie stated the same, bots using foundation summoners in Gold League is not adequate.
In modern, the bots won't appear as often as more humans are playing. But they have to be adequately equipped with cards and strategy to not make them "easy" targets, and hence to farm SPS.
As I said, I'm a fan of brackets, and I'm very sure that the bots can be fine-tuned towards a better performance that will allow that, to a balance of 50/50 or 60/40 chance of winning against a bot. But not 80/20 anymore, nor 20/80 :-D
It's a bit more complicated than just aiming for a win percentage. We can build bots you'd never beat that hold a 50% win rating. We actually did this early on. The bots are much more dynamic now and we can keep tweaking and adjusting them over time. Originally we had nothing between "free win" and "instant loss" and now there's much more granularity. There's some difficulty in catering to players with drastically different skill levels as well, so after 3700 you're basically always playing the hardest bots, we put some restrictions so you aren't just playing heloise bots every time, but it's ultimately a balancing act. What's easy to you may be hard to someone else and if you can beat a bot every time and I lose to it every time, that's still a 50% win rate. All in all, it's more complicated and complex than just aiming for a number. Bots should be adapting over time to become harder or softer based on rating. Just submit the feedback and we'll compare results/notes and make adjustments where we can.
View more
Using Foundation Archons and units can work, so long that the team tweaked their level and stats to be adequate for the bracket or rating that they were being played by the bots. The team can do it in their end since they were working with ghost cards and have actually done it before since there used to be a time where modern liquidity bots can use level 6 or 8 foundation Archons.
I'm currently in 3rd place with my alt. My main concern about the liquidity bots is that they keep on using Foundation cards even at Diamond/Champ rating, which makes it almost an easy win. If the bot will keep on using Foundation cards, at least try to increase their stats and level to match the bracket they were being played. There used to be a time when modern liquidity bots were using max-level Foundation cards, so it is certainly possible.
Another thing that I think is making balancing liquidity bots more difficult is the card balancing for both the Rebellion and Conclave and Conclave Arcana sets, with select lineups that I can continuously spam with my eyes closed, such as Bella, Hanv, and Mancer teams. I don't think that will be the case if we are still in CL era.
This post has been manually curated by @bhattg from Indiaunited community. Join us on our Discord Server.
Do you know that you can earn a passive income by delegating to @indiaunited. We share more than 100 % of the curation rewards with the delegators in the form of IUC tokens.
Here are some handy links for delegations: 100HP, 250HP, 500HP, 1000HP.
100% of the rewards from this comment goes to the curator for their manual curation efforts. Please encourage the curator @bhattg by upvoting this comment and support the community by voting the posts made by @indiaunited.
Thank you!