The default RNG algorithm in the Unity engine is substantially different from the one in Adobe Air, so gem cascades is something that is guaranteed to change (unless of course the devs hack the RNG).
Right now I would say the console AI gets an equal priority on cascades relative to the user, or close to it, but it is severely hampered on PC.
[/quote]
No tweaking of sliders, No Famine. 47.3%. I think I’ll tap down Skull Priority a smidgeon, as see how it goes.
Both of you are high level, do either of you play PvP much? When you don’t you end up getting invaded by weaker players.
What’s PvP?
I Get invaded by all levels mostly 1k+. The last few were of my wins on the log were from GoT. So obviously decent players.
EDIT: this week I haven’t played as much. I never grind the leaderboard anymore, except the first week when @Rickygervais plugged in his controller bot on and stole #1 from me.
I wouldn’t say I play a ton of PVP. Before Christmas, I was getting up on the leaderboard occasionally, but these days, I probably do about 100-150 battles a week. I’ve been focusing on traiting troops, so doing a ton of explore over the last couple months. I seem to get a pretty good variety of invaders, but it might be a touch more lower-level since I went into a more casual guild 6 weeks ago.
look like you will never forget that lol
Ha! Never.
I don’t have that much time to play and try for the leaderboards anymore. So that was my chance.
@Stan
I’m curious why only the last few weeks it’s been above 50%??
Since PvP update came out my defense team NEARLY always has a positive score by the end of the week. Lol, gem sliders and ai improvements only increased my defense winning %.
I’m not alone in this and most of my friends have positive scores. Its like a friendly little comp between us.
Not currently at home but I’m about 300-350 defensive wins more then loss overall.
I can’t comment for xbone players but that’s MY experience on ps4
Prior to guild wars, I’d been using a 1-troop defence or intentionally weak/funny defence teams to try to generate easier revenges, or using event teams to try help people get their event points (on weeks when that mattered).
@death im currently 59w 85l
For anyone who is yet to read how the Console AI works, this was a forum post from Feb explaining how the new AI functions differently from the current PC/Old Console AI code.
Just to explain how the sliders work, during each turn, the AI will determine what priorities each gem is worth depending on the troops on either team. Mana colors that you require is prioritized more over one that you do not need. Mana colors that the opponent has is prioritized over Mana colors that no one has. And Skulls in general is prioritized slightly higher than any Mana color. That is a simple overview of how the Mana Color Priorities function.
The Sliders directly modify these numbers via multiplication, where the middle is default, and is multiplied by 1 (effectively changes nothing from the AI’s priority mentioned above). Pushing the slider to the right will effectively increase the priority of the AI picking that mana color, and since it is multiplication, it has considerable effects on the scoring. This is similar with decreasing the slider to the left, where the priority is now a fraction of what it recommends. This is a simple overview of how the Defense Gem Priority sliders work.
From the forum post provided above, each spell is also given a priority. This priority is generally higher than skull matches, but that is all dependent on what steps that spell has (e.g. summoning a troop with a full team is pointless, so that lowers the priority of the spell, where if there is room to summon, it has a higher priority). As mentioned by @Strat, when colors are involved in spells, the priorities above will determine the priority of that spell (e.g. exploding a gem. If there is a spot on the board where exploding the gem is beneficial, meaning that those prioritized gem values above are taken into account, then that will change the priority of the spell to cast in either a positive or negative way).
Taking all the above into account, spells and moves can now be quantifiable on the same scale (using priorities). Before, this was not the case, like where the AI could have an awesome spell ready to cast, but ignore it because there is a skull match. This means that changing the slider will change the AI. Changing the slider to MAX will now MULTIPLY into gem matching, meaning it could also be preferred over casting spells, or may even still try to match that Gem Color when it is not needed.
Basically, it is recommended to NOT put the sliders to MAX or MIN, but to give more subtle movements to them. But, there may be cases where you want to completely ignore or favor certain gem colors, hence why we provide the option, but I would still recommend to keep it close to the middle, as that is the recommended priority for each gem (using the statement from the first paragraph of this section).
TL;DR - Gem Priorities affect what gem matches are made by using multiplication. Using Extremes (low or high) WILL effect what the AI does, including when it should cast a spell or take skulls.
So can you explain to me what happened in this situation?
It was before bone dragon nerf
I had
Frost sprite***
Courage***
Courage ***
Bone dragon***
So in this case i wanted to priorise purple and brown, since i wanted to get my bone dragon ready quick.
So if we start from middle, i puted purple and brown a little bit to the right
I dididin’t touch skulls and red sliders (in middle)
And puted blue, green and yellow completely to the left
And when i tested my team it was going for blue first
Edit: just to be clear, AI also had the option for brown purple and skulls but decided to go for blue
@GoldPhoenix0 Thanks for the detailed write up. In a future update could you please provide the multiplier (or fraction) that is going to be applied to the AI for each color selection on screen. Because there is no feedback provided to the user in the slider it feels very imprecise.
Also would be nice if we could have a training room where we could put any defense team and test it without changing pvp defense?
Cause i don’t like to be attacked wile im trying to adjust my defense
Appreciate the very detailed write up. Thanks @GoldPhoenix0!
Well, as these weightings are values behind the scenes, I’m not sure how useful giving numbers would be, as without knowing what the base priorities are, having a number saying how effective a change is would be counter-intuitive. If we put a number for max and min (e.g. times x and 1/x) on this menu, that could cause more confusion, as what does these numbers actually do? In place of this, we change the scale of the Gem on the slider to show that it is more or less prioritized, as this is closer to the result that players will see, rather than times x or 1/x.
That does sound a bit weird that it would still target blue first over purple or brown. There is also the random factor which is also part of the calculation, where it could have boosted the chances of a blue match, and lowered the chances of the other gem matches.
Just interested to know, was it just a single case of this happening, or was it constantly choosing blue over other the higher priority colors?
With this particular team, it was constantly choosing blue over other.
Fair enough, maybe not the raw multipliers exactly. But think of it from the users perspective. Some feedback or indication of priority would be very helpful. As it stands its too vague.
This might be a silly example: What if your oven didnt have a temperature gauge or display just a slider bar? Or your car without a speedometer? Yes you can ballpark everything, but then how do you discuss and share that information with others? How can you reproduce your settings with a new Defend team after a change/swap?
Please consider something on screen. Thanks again for getting back to me.
must be some sort of multiplier of 0 going into a maxed value? XD