Talking Turrets

[Waladil]

I came to this forum just to discuss the OP-ness of turrets. My build that absolutely ROLLED hard (and would handle even harder difficulties with ease) called for one ninja exo, specced to turrets (with a minor in stealth actions), one science exo, specced to radar range and hacking (with a minor in viruses), and a siege exo specced to having lots and lots of shadow torpedos.

Now, I was playing with Genji 'cause I like my exos epic. But this was my second serious playthrough and I ended up curbstomping the game. As I got up to 40-50 radar range and 20+ trap skill (eventually reaching 130 or something), every mission I'd just drop a few turrets in the entry zone, swap to my science exo to get sensor readings, use the siege to blow up doors (Yeah. I used shadow torpedoes to blow up DOORS from the spawn point), and whistle in most of the level to my turrets.

Who would then proceed to massacre everything. Lion's Den missions, assassination runs, bot command centers, you name it. My turrets would one-shot most things, with tough things taking at most ~5 shots. Some bosses were kinda tough. After the main brunt was cleared out, I'd go in with a sniper exo or the rockets on the sieger to clear out those oddbots who don't path right, and if I saw anything scary I'd just whistle and scarper back to my mommy turrets. Or drop some more, since I was carting around plenty and never needed all of them.

Let me put it this way: I ran the numbers. My ninja had over 12,000,000 in potential damage by endgame. If I'd played differently (by playing out the endgame slowly, trying to boost trap level even higher) and perhaps started with a second ninja exo, I might have been able to kill a murderbot. Just outright kill him. They've got 100,000,000 health and are supposed to be "unkillable."

So as much as I hate to say it, turrets are kinda mad op, yo.

Oh, and my final mission... I almost lost. Almost. Teleporters were a new concept and I hadn't fully figured out tactics vis-a-vis them and that cost me a couple good bots, plus something about that mission (either the teleporters or the human reactors) was really screwing with AI pathing. Like they'd dance around at a corner, unable to decide if they wanted to come at me or the reactor. So that made the whole "sit and wait" plan less good. But eventully I managed to clear it on day one, with two barely-specced bots.

[Logorouge]

By the way, is it intended that you get cash when you kill your own turrets?

[FrostyThePyro]

Could this be fixed with adding a diminishing returns mechanic to the stats?

yes, by adding a root to the equation it can take signifignatly more trap points to gain higher level bonuses while still requiring only a few to get it started, this is a very sound aproach as trap skill effects many things on the turret resulting in an exponential growth of power instead of a geometric one, so adding a root reduces growth exponentialy, evening the overall power into somethin closer to geometric.   

Or you could set it to an absolute max limit with an equation like power=minpower+maxpower(level/(level+weight))    lets say you wanting the starting power to be 10, but you never wanted to go above 1010 (so mine power 10 max power 1000), and a weight of 100 (which is what level you reach 50% max) so at level 0 you would have 10 power, at level 50 you would have 343 power 50/(50+100) is 1/3, at level 100 you would have 510 power, at level 300 you would have 760, etc.  you would allways be aproaching but never reaching the max of 1010.

In the steam forum thread it has been sugested that having each exo have a unique style of turret may also work, weather its in the form of entirely new turrets (such as seige having long range, splash and realy poor ammo,  or ninja's being a decoy with realy high health and bots will prioritize attackin, science having no damage to speak of but a emp effect), or in the form of letting one or two of the the stats scale in an extreme fashion while the rest get nerfed down (for example a sniper exo gets to keep the extreme range scaleing we have now, and an assault gets to keep the extreme sheild scaling with other stats applying diminishing returns to be less curb stompy)


One thing that should be looked at if/when turrets are nerfed is ease of placement, being able to place turrets at a range, or being able to place multiple turrets at a time could help ease the issue of the slippery slope of turrets being too weak to bother with or being so strong as to trivialize the challange of the game.

[Cherubael]

I came to this forum just to discuss the OP-ness of turrets. My build that absolutely ROLLED hard (and would handle even harder difficulties with ease) called for one ninja exo, specced to turrets (with a minor in stealth actions), one science exo, specced to radar range and hacking (with a minor in viruses), and a siege exo specced to having lots and lots of shadow torpedos.

Now, I was playing with Genji 'cause I like my exos epic. But this was my second serious playthrough and I ended up curbstomping the game. As I got up to 40-50 radar range and 20+ trap skill (eventually reaching 130 or something), every mission I'd just drop a few turrets in the entry zone, swap to my science exo to get sensor readings, use the siege to blow up doors (Yeah. I used shadow torpedoes to blow up DOORS from the spawn point), and whistle in most of the level to my turrets.

Who would then proceed to massacre everything. Lion's Den missions, assassination runs, bot command centers, you name it. My turrets would one-shot most things, with tough things taking at most ~5 shots. Some bosses were kinda tough. After the main brunt was cleared out, I'd go in with a sniper exo or the rockets on the sieger to clear out those oddbots who don't path right, and if I saw anything scary I'd just whistle and scarper back to my mommy turrets. Or drop some more, since I was carting around plenty and never needed all of them.

Let me put it this way: I ran the numbers. My ninja had over 12,000,000 in potential damage by endgame. If I'd played differently (by playing out the endgame slowly, trying to boost trap level even higher) and perhaps started with a second ninja exo, I might have been able to kill a murderbot. Just outright kill him. They've got 100,000,000 health and are supposed to be "unkillable."

So as much as I hate to say it, turrets are kinda mad op, yo.

Oh, and my final mission... I almost lost. Almost. Teleporters were a new concept and I hadn't fully figured out tactics vis-a-vis them and that cost me a couple good bots, plus something about that mission (either the teleporters or the human reactors) was really screwing with AI pathing. Like they'd dance around at a corner, unable to decide if they wanted to come at me or the reactor. So that made the whole "sit and wait" plan less good. But eventully I managed to clear it on day one, with two barely-specced bots.

I ended up in much the same situation. First game (Normal/ironman/meg with standard setup), using the science exo for a combo of turrets and hacking, ended up being absolutely trivial with turrets. Thought 'eh, lets turn up the difficulty so they don't trivialize everyting' and tried Expert/ironman/genji I went with the ninja exo (for the 50% propulsion boost) speccing for around 15+ turrets per mission and maximum trap skill (though I ended up with more like 50 sentries at the end, because of the scaling).

End result? Sentry guns still kill everything and I finished the ironman/expert playthrough having lost one mission (Wasn't paying attention, ninja exo got oneshot before I had gotten more than two or three sentries down). Biggest difference seemed to be that it was quite tough in the first ~5 missions before I got the trap skill up, and the exo weapons simply did not scale to bot health in a meaningful way, so if the sentry exo died at the start of a mission it was basically game over.

So, whichever way is chosen, turrets absolutely need less scaling with trap skill. If you spec even a bit for it, you end up with sentries that have more durability than your assault exo, firing what amounts to sniper railgun shots for damage/range, with chaingun levels of ammo. And you can easily deploy 40 or 50 of them during a mission.

Personally I would be in favor of limiting the number of sentries per mission (maybe a fixed 2 sentries per module and +1 per "additional sentries" stat elsewhere, instead of scaling with module level?), so you end up being able to deploy around 5-15 per mission if you go for it. Then you could improve the baseline sentry gun abilities so they are useful at the beginning/lower difficulties sans trap skill and reduce their scaling to be on around the same level as weapon improvements.

[chemical_art]

Could this be fixed with adding a diminishing returns mechanic to the stats?

yes, by adding a root to the equation it can take signifignatly more trap points to gain higher level bonuses while still requiring only a few to get it started, this is a very sound aproach as trap skill effects many things on the turret resulting in an exponential growth of power instead of a geometric one, so adding a root reduces growth exponentialy, evening the overall power into somethin closer to geometric.   

Or you could set it to an absolute max limit with an equation like power=minpower+maxpower(level/(level+weight))    lets say you wanting the starting power to be 10, but you never wanted to go above 1010 (so mine power 10 max power 1000), and a weight of 100 (which is what level you reach 50% max) so at level 0 you would have 10 power, at level 50 you would have 343 power 50/(50+100) is 1/3, at level 100 you would have 510 power, at level 300 you would have 760, etc.  you would allways be aproaching but never reaching the max of 1010.

In the steam forum thread it has been sugested that having each exo have a unique style of turret may also work, weather its in the form of entirely new turrets (such as seige having long range, splash and realy poor ammo,  or ninja's being a decoy with realy high health and bots will prioritize attackin, science having no damage to speak of but a emp effect), or in the form of letting one or two of the the stats scale in an extreme fashion while the rest get nerfed down (for example a sniper exo gets to keep the extreme range scaleing we have now, and an assault gets to keep the extreme sheild scaling with other stats applying diminishing returns to be less curb stompy)


One thing that should be looked at if/when turrets are nerfed is ease of placement, being able to place turrets at a range, or being able to place multiple turrets at a time could help ease the issue of the slippery slope of turrets being too weak to bother with or being so strong as to trivialize the challange of the game.

I don't like this, because enemies increase geometrically. Doing this would cause turrets to be OP in early game but fall off hard late game. Better to adjust how trap skill effects the stats directly.

[ScrObot]

In the steam forum thread it has been sugested that having each exo have a unique style of turret may also work ...

I like the idea of different sentry types, I'd much rather see them as typed differently on the parts themselves rather than be exo type-specifc (although it's a whole additional slew of components to have to compare). Sure, it promotes having a varied exo line up, but I would rather not be shoehorned into only using an exo type for it's intended purpose -- perhaps I'm buffing up a Ninja exo with hacking and computers instead of bringing along a Science exo, etc.

Alternately, keep only one "type" of turret but add a different deployable type ("decoy"?) or two to fit the other proposed roles.

[FrostyThePyro]

I don't like this, because enemies increase geometrically. Doing this would cause turrets to be OP in early game but fall off hard late game. Better to adjust how trap skill effects the stats directly.

Are you talking about the limits part, or the roots.  Both of which can be manipulated to give just about any curvature you like.   Though it should be noted that part of the issue is that turrets don't grow geometrically as it is, but exponentially, enemies and the players do as well in their ways.  If you double something's health and damage, you haven't doubled its power, but quadrupled it.  Ammo and range also factor in to that, and as they all key off of one stat for turrets it gets to be a bit of a god stat, especially a stat that can appear on just about any peace of equipment.  And with turrets you have the effect of dropping multiple turrets which is another exponential effect.  2 turrets are generally more than twice as good as 1 turret, and four turrets are more than twice as good as 2.


It should also be noted that with both systems its very easy to have early parts be near indistinguishable from a linear equation, and neither require any sort of starting multiplier (aka no effect on low trap level turrets compared to now).  On the limits, with a very high max and very high weight it will be nigh impossible to notice the difference until you get to very high trap stat (for a quick example weight 1000 and max of 1001 at trap level 50 you get an answer of 48, at level 100 you get an answer of 91, and at trap level 400 you get an answer of 286).  Similarly if using a root you will want to be using a low root, such as 1.1 or 1.05 or less.  In a similar example as above a 1.05 root will answer 41.5, 80.3, 300.7 at trap levels 50, 100, 400 respectively. 

That all said the official word is now

Ok, whew, Sentries were REALLY overpowered. Them being good is cool, but certain cheesemasters were trivializing high-difficulty endgames with 300+ trap skill, etc... wily players. Anyway:  Sentries no longer get an ammo boost from trap skill (unless they were already placed in an old save).
 Sentry range boost from trap skill is capped at 5 (max total range of 11) (unless they were already placed in an old save).
 The +TrapSkill effect now costs 4 item levels per point instead of 3 (from a player perspective this basically means trap skill values will be 33% lower on the items you find in the future, has no impact on items from old saves).
 The +%Propulsion-related effect's range from 12%-to-60% => 9%-to-45%. (again, won't affect items from old saves).
 The +SentryCount effect magnitude from 2+level => 1+(level/2) (same deal with old save items)
 Overall, with the +attack and +shields boost from trap skill still intact (albeit with lower trap skill, and fewer sentries to work on), these are quite possibly still too strong. We have a sneaking suspicion that players will find out and let us know.
 Thanks to Sooty, Misery, Frumple, Reasonance, Waladil, The_Ring-Bearer, Tridus, Mick, and others for inspiring these changes,

So it looks like they are primarily attacking the trap skill equipment.  though I am a little concerned with how the propulsion related effects part will impact stealth moves.  I may want to do an all ninja play through before this goes live.

[chemical_art]

Not sure what to think of the range changes. the very chessmasteries who abuse turrets most knoww w how to setup chokes that don't need that range. doesn't reduce their cheese but hurt casual use more.

[chemical_art]

Will flesh out the idea later, but perhaps the trap skill itself is too broad. It seems if it was split a little bit the traps could have a few strengths based on build rather them being good everything.

[Teal_Blue]

There needs to be a way to nerf turrets without destroying their scalability completely. I don't think you can hit any one stat though.

If you upgrade turrets to a high level and manage to get them into a good location, then you should be rewarded for that. However, it seems that almost everywhere is a "good location", because turrets are so hard to overrun. High level = high ranger = less hits on the turrets (or none) as well as more shields. It feels a bit silly when a turret has a longer range than my sniper, hits harder than my siege bot, and has more shields than my assault. It's like a super exo!

Here are my thoughts.

- Put a timer on their attack, say they fire every four rounds or something. This means that you can no longer whistle the entire world and feel secure they can't get overrun because only 1-2 bots can get in range at a time.

- They should have very weak shields. You should be punished if you place them in a location where bots can actually reach them.

- Limited ammo, each placed turret is like the equivalent of a who new weapon worth of ammo on a exo, which sorta trivializes the whole ammo scarcity issues if you use them.

- I think it's OK if they have strong range, so probably not much tweaking there.

With the above changes, you may no longer be able to setup turrets and then go hide in a corner somewhere. If you lure enemies into a turret killzone, you may have to actually contribute to the fight yourself with your exos. With weak shields, they could be much more vulnerable to stealth bots and long range firing bots that can hit them, so you can't just brainlessly hit the next turn button.

Obviously the goal here shouldn't be to make turrets useless, but it should be to better define their role as a way to support your exos instead of replacing them.

Just a thought, but if you added a weapon 'cooldown' period, so the turrets don't just fire-fire-fire all the time, it might mitigate some of the problem with them being over-powered. I am not sure how much. I suppose it depends on how long a cool down, and whether or not any other stats, such as distance, ammo, shields, gun strength are lowered or not.

-Teal

[FrostyThePyro]

There needs to be a way to nerf turrets without destroying their scalability completely. I don't think you can hit any one stat though.

If you upgrade turrets to a high level and manage to get them into a good location, then you should be rewarded for that. However, it seems that almost everywhere is a "good location", because turrets are so hard to overrun. High level = high ranger = less hits on the turrets (or none) as well as more shields. It feels a bit silly when a turret has a longer range than my sniper, hits harder than my siege bot, and has more shields than my assault. It's like a super exo!

Here are my thoughts.

- Put a timer on their attack, say they fire every four rounds or something. This means that you can no longer whistle the entire world and feel secure they can't get overrun because only 1-2 bots can get in range at a time.

- They should have very weak shields. You should be punished if you place them in a location where bots can actually reach them.

- Limited ammo, each placed turret is like the equivalent of a who new weapon worth of ammo on a exo, which sorta trivializes the whole ammo scarcity issues if you use them.

- I think it's OK if they have strong range, so probably not much tweaking there.

With the above changes, you may no longer be able to setup turrets and then go hide in a corner somewhere. If you lure enemies into a turret killzone, you may have to actually contribute to the fight yourself with your exos. With weak shields, they could be much more vulnerable to stealth bots and long range firing bots that can hit them, so you can't just brainlessly hit the next turn button.

Obviously the goal here shouldn't be to make turrets useless, but it should be to better define their role as a way to support your exos instead of replacing them.

Just a thought, but if you added a weapon 'cooldown' period, so the turrets don't just fire-fire-fire all the time, it might mitigate some of the problem with them being over-powered. I am not sure how much. I suppose it depends on how long a cool down, and whether or not any other stats, such as distance, ammo, shields, gun strength are lowered or not.

-Teal

On a similar thought to cool downs you could make turrets a collective entity, by which I mean all turrets collectively share one action (so no mater how many turrets you have out only one will shoot per turn).  placing multiple turrets would simply give the turret collective a broader selection of firing angles(in addition to the obvious ammo and shield benefits).



And if all else fails, turn turrets into a weapon, put it on the paper doll with 3 slots that can take shield and weapon equipment, but shield ones would effect the turret instead of the exo.  So instead of keyin off trap skill it would key off whatever equipment you put on it just like any given gun would, except this gun has a shield stat as well as damage/range/ammo. 

[Misery]

Not sure what to think of the range changes. the very chessmasteries who abuse turrets most knoww w how to setup chokes that don't need that range. doesn't reduce their cheese but hurt casual use more.

On casual/easy/normal though, aside from not really needing the turrets in the first place, they're not going to be hurt too much by range changes, particularly as many enemies simply wont have much range.  Even on Expert though, a range of 11 is..... quite good, quite good indeed.  And turrets arent abusable just by chokepoints;  right now, they're *very* abusable by merely being at range, which gets extreme really fast.  Having to set up chokepoints to use them is actually rare.

The upcoming changes look good overall, but based on those I'm expecting these will still need more nerfing after that. 

[Winge]

The upcoming changes look good overall, but based on those I'm expecting these will still need more nerfing after that.

I'm inclined to agree, although it's hard to say for sure.  The most obvious change would be to limit where trap skill appears again.  Right now, I'm seeing it in places it doesn't seem to belong, which is probably why people are able to stack it so high (OK, and the buffed +% Propulsion stat that I thought was too weak at one point).

IMO, only the following parts should be able to have Trap Skill:
1.  Propulsion-specific parts (Hover Thruster, Stealth Generator, Overload Circuit)
2.  Propulsion Boosters (NONE of the other boosters)

And maybe the following:
3.  Sentry Turret/Minelayer
4.  Reactor

Right now, it looks like I can get +Trap Skill on anything except Weapons and Shields parts, which makes it a little bit too easy to get.

[Misery]

The upcoming changes look good overall, but based on those I'm expecting these will still need more nerfing after that.

I'm inclined to agree, although it's hard to say for sure.  The most obvious change would be to limit where trap skill appears again.  Right now, I'm seeing it in places it doesn't seem to belong, which is probably why people are able to stack it so high (OK, and the buffed +% Propulsion stat that I thought was too weak at one point).

IMO, only the following parts should be able to have Trap Skill:
1.  Propulsion-specific parts (Hover Thruster, Stealth Generator, Overload Circuit)
2.  Propulsion Boosters (NONE of the other boosters)

And maybe the following:
3.  Sentry Turret/Minelayer
4.  Reactor

Right now, it looks like I can get +Trap Skill on anything except Weapons and Shields parts, which makes it a little bit too easy to get.

Yeah, I was thinking the same thing.

And when you combine that with %propulsion boosts, it inflates way too much.  Being able to stack the percentage boosts makes sense and works for pretty much every other stat type, like attack damage, this being because nothing other than weapon parts gives any specific weapon a big blob of damage boosting.

It definitely should be propulsion-only, that stat.  It also makes the propulsion stuff in general just make more sense.

[BobTheJanitor]

This gets to the loot vapor-lock problem of another thread. It's hard to optimize when it seems like every slot can have just about every stat. I would be all for stats in general being more focused, so that you know when you look at one system you're only going to see some specific stats, which thus makes it easier to compare loot. I'd be all for trap skills just going on certain slots, in the greater context of making all skills only found in certain slots.