Buff Assault Class

mcompany
Autonomous Entity
Autonomous Entity
Posts: 865

Re: Buff Assault Class

Post#41 » 19 Oct 2017, 18:15

TheKidPunisher wrote:@mcompany: so, you modify damage (at SR/MR/LR) from 1350/750/150 to 1440/800/160 and it is a nerf? I need more explication or i dont get it...

I made a miscalculation somewhere (I didn't go to verify with my spreadsheet, I used a new calculation for that). Apparently that is actually 106.66...% of the damage current assault does. My b.

pier4r wrote:Chris has some good point.

One cannot adjust assault because there will be 4vs4 matches . Let them come first.

Otherwise it is like going naked around the city saying ' eventually I will take a shower , so I need to be naked' . It is silly.

When there will be in general less bots on a map, then yes, maybe it will be needed to buff the assault a bit.

Sure, that makes sense, but the problem is that assault has literally been confirmed to be ≤ shotgun in 99.99% of categories (assault's firing rate is technically better, but that is kinda meaningless compared to everything else). Not to mention that assault has been classified as the worst bot since 5.2. Maybe buffing assault because it *might* get better under a new system *might* be worthwhile, at the very least, we new to make it so that assault is not a strictly worse shotgun/shotgun is not a strictly better assault. Because currently, there is never a situation that in the long run that having an assault is better than having a shotgun (unless it is forced to attack at long range by itself in a 1v1, something that is not the most likely thing to happen, and something that arguably isn't even strictly bad as a shotgun)

pier4r
Skynet
Skynet
Posts: 3190

Re: Buff Assault Class

Post#42 » 20 Oct 2017, 08:33

I did not test but surely someone did. In a 1vs1 (that rarely happens but still) who wins, assault or shotgun?

Note that a 1vs1 or even better a 2v2 (2 ass vs 1 ass + 1sg) one has a simplified model of real battles but at least all the factors count. Speed, damage, tactics , etc.
http://www.reddit.com/r/Gladiabots/wiki/players/pier4r_nvidia_shield_k1 -> Gladiabots CHAT, stats, insights and more ;

MrChris
Automaton
Automaton
Posts: 157

Re: Buff Assault Class

Post#43 » 20 Oct 2017, 09:54

Assault 1v1 long range happens all the time
------------------------------------------------
My in-game name is MrChris

Creator of the unofficial Gladiabots stats page: https://gladiabots-stats.info.tm/mrchris
And the Gladiabots retreatment simulator: https://cmrichards.github.io/glad_simulation

pier4r
Skynet
Skynet
Posts: 3190

Re: Buff Assault Class

Post#44 » 20 Oct 2017, 10:40

yes, I am not interested in assault 1 vs 1 long range. (aside from the fact that in normal battles there are more factors)

I mean, with the current bots (and give one's tactics): assault vs shotgun, who wins?

and assault + sg vs assault x 2 , who wins?
http://www.reddit.com/r/Gladiabots/wiki/players/pier4r_nvidia_shield_k1 -> Gladiabots CHAT, stats, insights and more ;

sollniss
Automaton
Automaton
Posts: 164

Re: Buff Assault Class

Post#45 » 20 Oct 2017, 10:45

Long: Assault > SG
Mid: Assault > SG (50/50 if the SG moves into short, but depending on the range, you just die sometimes)
Short: SG > Assault

User avatar
Revenge
Neural Network
Neural Network
Posts: 361
Contact:

Re: Buff Assault Class

Post#46 » 20 Oct 2017, 16:53

sollniss wrote:Long: Assault > SG
Mid: Assault > SG (50/50 if the SG moves into short, but depending on the range, you just die sometimes)
Short: SG > Assault

Tell me again how a class with 60% more shield(5k to 3k) loses midrange to assault with same damage. ;)

If your shotgun is having trouble against assault. Visit the Strategies forum for improvement: viewforum.php?f=3

If you're still having trouble, feel free to message me on discord, or you can get some help on Telegram.

User avatar
GFX47
Dev Team
Posts: 2603

Re: Buff Assault Class

Post#47 » 20 Oct 2017, 17:10

I think I'll only implement the "one bot of each class max" rule when once there's more classes.

sollniss
Automaton
Automaton
Posts: 164

Re: Buff Assault Class

Post#48 » 20 Oct 2017, 17:46

Revenge wrote:[
Tell me again how a class with 60% more shield(5k to 3k) loses midrange to assault with same damage. ;)

If your shotgun is having trouble against assault. Visit the Strategies forum for improvement: viewforum.php?f=3

If you're still having trouble, feel free to message me on discord, or you can get some help on Telegram.


I'm rank 10 grandmaster, I have a 1v1 engine, you can send me a challenge if you want.

mcompany
Autonomous Entity
Autonomous Entity
Posts: 865

Re: Buff Assault Class

Post#49 » 20 Oct 2017, 18:14

pier4r wrote:I did not test but surely someone did. In a 1vs1 (that rarely happens but still) who wins, assault or shotgun?

Note that a 1vs1 or even better a 2v2 (2 ass vs 1 ass + 1sg) one has a simplified model of real battles but at least all the factors count. Speed, damage, tactics , etc.

Short range is an obvious winner of shotgun, and long range shotgun can't land enough bullets and would lose too much time getting to medium range

Medium range... it is impossible to give a straight answer. If we assume both bots start attacking and haven't fired a single bullet all game then you get this chart
1-600 shotgun wins
601-900 assault wins
901-1200 shotgun wins
1201-2700 assault wins
2701-3000 shotgun wins
3001-4500 assault wins
4501-4800 draw
4801-5400 assault wins
5401-6000 shotgun wins
6001-6300 assault wins
6301-6600 shotgun wins
6601-8700 assault wins
8701-9000 draw
After this, it cycles

Even then, this is incredibly dependent on which bot starts aiming first (basically impossible to detect currently and absolutely ridiculous to attempt to detect with instant tags), and has already been proven to be minorly affected by the current "chance" of the bot. For example, it has been tested that if shotgun have shot 3+n*4 bullets and starts the battle at the edge of medium range that while assault will shoot its finishing bullet first (8000 armor), shotgun will just barely have enough time to shoot as it is dying and can turn the battle into a draw

Also note that if shields never recharged, the battle would look very similar in long range that it does in medium range

pier4r
Skynet
Skynet
Posts: 3190

Re: Buff Assault Class

Post#50 » 20 Oct 2017, 18:43

sollniss wrote:Long: Assault > SG
Mid: Assault > SG (50/50 if the SG moves into short, but depending on the range, you just die sometimes)
Short: SG > Assault


And again, I am not interested in "turrets". We don't use bots in stationary way, so comparing results in fixed positions is naive. I am interested in dynamic 1vs1 or max 2vs2.

So if a shotgun is a better assault, and 1vs1 and 2vs2 (then with 3 or more bots, the marginal change may be too small) the team with shotgun should win.

Does anyone has optimized Ais that he cannot find a way to let the assault survive a shotgun?
http://www.reddit.com/r/Gladiabots/wiki/players/pier4r_nvidia_shield_k1 -> Gladiabots CHAT, stats, insights and more ;

pier4r
Skynet
Skynet
Posts: 3190

Re: Buff Assault Class

Post#51 » 20 Oct 2017, 18:45

mcompany wrote:Short range is an obvious winner of shotgun, and long range shotgun can't land enough bullets and would lose too much time getting to medium range

Medium range... it is impossible to give a straight answer. If we assume both bots start attacking and haven't fired a single bullet all game then you get this chart

but then shotgun is not a better assault, otherwise would always win with similar optimized tactics.
http://www.reddit.com/r/Gladiabots/wiki/players/pier4r_nvidia_shield_k1 -> Gladiabots CHAT, stats, insights and more ;

TheKidPunisher
Autonomous Entity
Autonomous Entity
Posts: 563

Re: Buff Assault Class

Post#52 » 20 Oct 2017, 19:01

pier4r wrote:And again, I am not interested in "turrets". We don't use bots in stationary way, so comparing results in fixed positions is naive

+1

pier4r wrote:I am interested in dynamic 1vs1 or max 2vs2.

The game isn't a 1v1 at the base but it can be possible that the game finish with only 1 bots in each team...

pier4r wrote:Does anyone has optimized Ais that he cannot find a way to let the assault survive a shotgun?


In a 1v1 pure (on the 1v1 map in test mode for example), i do not see how an assault can survive a shotgun...

mcompany
Autonomous Entity
Autonomous Entity
Posts: 865

Re: Buff Assault Class

Post#53 » 20 Oct 2017, 19:04

pier4r wrote:And again, I am not interested in "turrets". We don't use bots in stationary way, so comparing results in fixed positions is naive. I am interested in dynamic 1vs1 or max 2vs2.

In a 1v1, this isn't like shotgun vs sniper. It is pretty much always a worse idea to move in any way on this battle. If assault flees from short range, shotgun I believe can get free health damage or a free push, so fleeing is not an option, and for long range, considering how even medium range is, shotgun moving to medium range is usually disadvantagous (and I forget if moving to short range was better). So turrets usually gets the job done of testing how this plays out for 1v1.

pier4r wrote:So if a shotgun is a better assault, and 1vs1 and 2vs2 (then with 3 or more bots, the marginal change may be too small) the team with shotgun should win.

Does anyone has optimized Ais that he cannot find a way to let the assault survive a shotgun?

Not quite sure what is being asked here

pier4r
Skynet
Skynet
Posts: 3190

Re: Buff Assault Class

Post#54 » 20 Oct 2017, 20:30

TheKidPunisher wrote:The game isn't a 1v1 at the base but it can be possible that the game finish with only 1 bots in each team...


Yes I know, but it is too complicated to figured out which bots contributes what in a 3vs3, 4vs4, 5vs5. So a 1vs1 or 2vs2 that are not fixed gives a slight idea how things goes.
http://www.reddit.com/r/Gladiabots/wiki/players/pier4r_nvidia_shield_k1 -> Gladiabots CHAT, stats, insights and more ;

mcompany
Autonomous Entity
Autonomous Entity
Posts: 865

Re: Buff Assault Class

Post#55 » 21 Oct 2017, 18:45

The assault vs shotgun medium range timings
1000 ms: shotgun 600 assault 300
1100 ms: assault 600
1200 ms: assault 900
2000 ms: shotgun 1200
2200 ms: assault 1200
2300 ms: assault 1500
2400 ms: assault 1800
3000 ms: shotgun 1800
3400 ms: assault 2100
3500 ms: assault 2400
3600 ms: assault 2700
4000 ms: shotgun 3000
4600 ms: assault 3000
4700 ms: assault 3300
4800 ms: assault 3600
5000 ms: shotgun 3600
5800 ms: assault 3900
5900 ms: assault 4200
6000 ms: shotgun 4200 assault 4500
7000 ms: shotgun 4800 assault 4800
7100 ms: assault 5100
7200 ms: assault 5400
8000 ms: shotgun 6000
8200 ms: assault 5700
8300 ms: assault 6000
8400 ms: assault 6300
9000 ms: shotgun 6600
9400 ms: assault 6600
9500 ms: assault 6900
9600 ms: assault 7200
10000 ms: shotgun 7200
10600 ms: assault 7500
10700 ms: assault 7800
10800 ms: assault 8100
11000 ms: shotgun 7800
11800 ms: assault 8400
11900 ms: assault 8700
12000 ms: shotgun 9000 assault 9000

TheKidPunisher
Autonomous Entity
Autonomous Entity
Posts: 563

Re: Buff Assault Class

Post#56 » 21 Oct 2017, 19:08

Ok but it is only on «turrets» mode where bots just fights as retards; with a good AI shotgun does not lose 1v1 vs assault...

mcompany
Autonomous Entity
Autonomous Entity
Posts: 865

Re: Buff Assault Class

Post#57 » 21 Oct 2017, 19:19

TheKidPunisher wrote:Ok but it is only on «turrets» mode where bots just fights as retards; with a good AI shotgun does not lose 1v1 vs assault...

Well, I haven't tested moving to short range against an assault when the battle starts in medium range, so I wouldn't know if shotgun can definitively win in a 1v1

Anyways, for 2v2s and beyond, this information is generally helpful, and shows that shotgun is not always guaranteed to lose in medium range for all health values as sollniss would have you believe. It also is yet another reason that it is accurate to say that shotgun is basically the same bot as assault, just better

TheKidPunisher
Autonomous Entity
Autonomous Entity
Posts: 563

Re: Buff Assault Class

Post#58 » 21 Oct 2017, 19:23

Assault is better at long, as he takes only 2,4s to deals damage when shotgun take 4s

mcompany
Autonomous Entity
Autonomous Entity
Posts: 865

Re: Buff Assault Class

Post#59 » 21 Oct 2017, 20:13

First off, landing a single bullet once every 5 shots on a 3 shot cycle would never give you a straight answer. But if we ignore that, then yes, shotgun is worse in a 1v1 at long range. But if we apply the "don't use turrets" insistence you and pier have, then you should test if shotgun would win by getting to short range. I just did: shotgun won. So this is kinda meaningless. And if we looked at the timings of when bullets hit for a long range battle (while assuming shields don't recharge) we would see very similar results to what happens in medium range. Shotgun will be ahead for a few times, assault would be ahead for others, neither would be definitively better. So it is not quite accurate to say that assault is better at long range, but rather than shotgun is worse only when it doesn't have an ally to attack with

pier4r
Skynet
Skynet
Posts: 3190

Re: Buff Assault Class

Post#60 » 21 Oct 2017, 20:36

mcompany wrote: that shotgun is basically the same bot as assault, just better

Wouldn't be better mean: always better?

From your checks does not seem so.
http://www.reddit.com/r/Gladiabots/wiki/players/pier4r_nvidia_shield_k1 -> Gladiabots CHAT, stats, insights and more ;

Return to “Feature Requests”

Who is online

Users browsing this forum: No registered users and 4 guests