The thing is … I do see why they made their decisions. I completely understand them.
But they need to say 1 thing when they throw these decisions here. And they haven’t.
GW hasn’t been a matter of statistics throughout a 2 months time. We have compared data daily. In which they failed this very update.
I mean … at the end of the week, we might find out who is fit to be the paragon… But daily wins are absolutely out of place.
1st of all …
… Winning guild wars … Fast … Feels wrong to even read. I don’t believe anyone has ever said that in their entire life who has been playing this game.
So in case no one has ever said this to developers:
GW are not the same as PvP battles where all the criteria they presented work (which I called TDS-The Dragon Soul data before). They have always been about careful strategy, team-building and board responsiveness.
… I have never thought I would have been thrown in a decisive battle with words:
Be bullheaded and try to break through the enemy defenses! If you stay alive, you won!
I pretty much had the same scores as the day before, avg around 1,4xx. With one battle where I lost one troop scoring 1,35x.
Team used a very similar exploding strategy as the day before, so comparative game times/amount of turns.
Team was:
Gorgotha
Sylvanimora
Emperor Khorvash
Elemaugrim
Yep, that was my feelings too when I dropped a troop. Speed and kamikaze GW teams are upon us!
EDIT: Troop lost was Sylvanimora (almost full HP) to a Kerberos devour. Not sure if thats why I lost so little points because I had to then kill the HP back off the Kerberos.
It seems that today I’m having a 1200 points per battle day 
No matter what I do I get something around that amount of points. 2 troops of color, lost one? 1200. 3 troops of color, lost noone. 1200. Fast game. 1200. Slow game - yeah, 1200 
Why not go full open random mode with those points? After each battle we should play one armed bandit with as many goes, as many troops survived a fight xD Equally random and much more transparent system than the one we have right now 
feels like armed bandit would be more fair indeed,
also it seems today i rolled a 13xx points day 
Did you also base around exploding?
Thanks, this proved incredibly useful. Time for a deeper number dive. Please note that this is reading tea leaves, looking for patterns that might acutally not exist.
All the measurements above were reported to have been perfect wins with full color bonus.
Measurement 4 likely has neither MANA_BONUS nor DAMAGE_BONUS, due to the division by zero issue. It should also have full COLOR_BONUS and full SURVIVAL_BONUS. Removing the BASE_SCORE added to every result we get
BASE_SCORE * (FULL_COLOR_BONUS + FULL_SURVIVAL_BONUS + PARTIAL_SPEED_BONUS) = 1014.
Measurement 3 likely has a full MANA_BONUS and full DAMAGE_BONUS, due to only granting a single turn. It should also have full COLOR_BONUS and full SURVIVAL_BONUS. Removing the BASE_SCORE added to every result we get
BASE_SCORE * (FULL_COLOR_BONUS + FULL_SURVIVAL_BONUS + PARTIAL_SPEED_BONUS + FULL_MANA_BONUS + FULL_DAMAGE_BONUS) = 2424.
Now, we know that the speed bonus is linear, going from full to zero in 30 moves. People also like to pick “nice looking” number instead of weird fractions. B within the speed bonus is likely something reasonably simple that will make 1014 and 2424 look “nice” when removed from the equation. One pretty good fit would be 6, that makes every action less than 30 worth 6% of 300 = 18 points. Assuming measurement 4 contained 12 actions (for a 30 - 12 = 18 speed multiplier) and measurement 3 contained 7 actions (for a 30 - 7 = 23 speed multiplier), the numbers without the partial speed bonus would look like this:
BASE_SCORE * (FULL_COLOR_BONUS + FULL_SURVIVAL_BONUS) = 1014 - 23 * 18 = 600
BASE_SCORE * (FULL_COLOR_BONUS + FULL_SURVIVAL_BONUS + FULL_MANA_BONUS + FULL_DAMAGE_BONUS) = 2424 - 18 * 18 = 2100
That’s great, a perfect fit, we can remove BASE_SCORE to arrive at
FULL_COLOR_BONUS + FULL_SURVIVAL_BONUS = 2
FULL_COLOR_BONUS + FULL_SURVIVAL_BONUS + FULL_MANA_BONUS + FULL_DAMAGE_BONUS = 7
and also
FULL_MANA_BONUS + FULL_DAMAGE_BONUS = 5
Hey, nice looking numbers, not a single fraction at all. COLOR_BONUS and SURVIVAL_BONUS are probably weighted the same, at 1 each. MANA_BONUS and DAMAGE_BONUS are quite interesting, they use the same formula, so they might also be weighted the same, at 2.5 each. In any case, they outweigh color and survival to such and extent that those can safely be ignored. Speed contributes next to nothing, it’s just a small bonus for a fast match.
TLDR: Lots of guessing. If it fits, forget about colors and surviving troops. Pick the one perfect loop/nuke team every fight each day, you’ll get about twice the score than with any other approach if you play it well.
yes, i used ragnagord among others
Yeah, Gorg and TDS were the core of the team.
I have to imagine that full color bonus is +200%, or 600 points. If it isn’t, then the schedule screen is lying to us about the bonuses from our guild statues.
I have still a battle to fight left.
I’m tempted to forget it’s brown day on PS4 and go with the kraken team to see what happens.
Sirrian claimed somewhere that the final full color bonus was halved, which would fit perfectly. It’s not really lying, the scaling gets applied afterwards.
What he said was that the surviving troops bonus was halved, not the color bonus. So we went from 40% to (roughly) 20% for having all four troops survive.
Isn’t that what people wanted? To be able to try more strategies? Which in turn means the drain/devour defense meta isn’t as punishing? It seems nice you can still score okay and not have to rage when the opposition gets a lucky devour…
Oh I’m not complaining, not at all I’m just making fun of the seemingly randomness of the whole thing and us frantically trying to figure things out
And if you ask me, it really can be random. I don’t care that much
Good point, it’s in the top post. It says very little about the color bonus though, it doesn’t even mention how COLOR_BONUS gets calculated. There should be an E% in there, which likely halves the result. In order to verify that, some more 0 turn scores using mismatched colors would be helpful.
It’s in the formula:
FINAL_SCORE = BASE_SCORE + BASE_SCORE * (COLOR_BONUS + SURVIVAL_BONUS + SPEED_BONUS + MANA_BONUS + DAMAGE_BONUS)
That part about it is amazing.
The thing is, most people score around 1400, while some players rocket out to 2700 points. By “something”.
It’s like getting to a game of poker. Where you slap the table with a straight Royal Flush, grinning… And the dealer tells you that someone else won with a full house.
… But won’t tell you why.
At the end of the day - you’ll just throw the game to the dustbin. Very few people like to play a game where you are not given the rules, but rewards are distributed anyway. It feels unfair, if anything.
If the bonuses are capped in a way… you could basically only get a certain amount of points (not like double the points for winning on turn 2) - that would solve things. But currently - we are just throwing cards blindly, trying to figure out what do the developers want from us.
It’s super common for games to not reveal the underlying scoring system/math. Almost always, players have to reverse engineer formulas. I think the only mistake they made was ever revealing it.
All I hear is a lot of laziness. Experiment for a few days and figure out what works best for you!!!
No one is speaking about Math here. This is not about “chances you’ll win”.
… This is about determining the winner.
In a poker - rules are set so you know which cards beat others in value.
… We don’t have that.
All games start with a set of rules. All of them. Exceptions are games that tell you “can you figure out how this works?”. And I don’t think people payed their money for a game like that.