| 1 |
@PRO_Dregs do tests and get data. Either live in the matchmaker or in dedicated testing. I can only use what data I see and what data people bring to my attention. The stats are fallible, in that they take a week or so to update sometimes, and they don't include tournament games yet.
|
1 |
@PRO_Dregs do tests and get data. Either live in the matchmaker or in dedicated testing. I can only use what data I see and what data people bring to my attention. The stats are fallible, in that they take a week or so to update sometimes, and they don't include tournament games yet.
|
| 2 |
\n
|
2 |
\n
|
| 3 |
I
don't
see
evidence
of
your
claimed
mastery
of
these
matchups.
Here
are
what
the
post-Bulkhead
(
12/11/20
to
31/1/21)
stats
say
games
between
2800+
players:
|
3 |
I
don't
see
evidence
of
your
claimed
mastery
of
these
matchups.
Here
are
what
the
post-Bulkhead
(
12/11/20
to
31/1/21)
stats
say
about
games
between
2800+
players:
|
| 4 |
* Amph vs. Tank has never occurred.
|
4 |
* Amph vs. Tank has never occurred.
|
| 5 |
* Spider vs. Hover occurred three times, each on Mechadansonia. Hover won two games.
|
5 |
* Spider vs. Hover occurred three times, each on Mechadansonia. Hover won two games.
|
| 6 |
[spoiler]
|
6 |
[spoiler]
|
| 7 |
https://i.imgur.com/0ntZ5Zd.png[/spoiler]
|
7 |
https://i.imgur.com/0ntZ5Zd.png[/spoiler]
|
| 8 |
\n
|
8 |
\n
|
| 9 |
The 'Pro 1v1' host games are now in, although casual isn't. I'm not sure whether including casual makes sense, but it could be something to ask for. At least in the case of Spiders, a replay search says that Mechadansonia has never been played on the casual host.
|
9 |
The 'Pro 1v1' host games are now in, although casual isn't. I'm not sure whether including casual makes sense, but it could be something to ask for. At least in the case of Spiders, a replay search says that Mechadansonia has never been played on the casual host.
|
| 10 |
\n
|
10 |
\n
|
| 11 |
Without data and testing this [i]]looks[/i] like a bunch of theory-crafting. Buoy has been a theoretical threat to Tanks for a long time, but it has taken many other factors for this to even come up as an issue. I know you have game knowledge behind what you're saying, but it takes a lot more work to devise a good solution than to point to a problem. I want replays and testing so that we have the data to do something reasonable, rather than breaking other matchups while running in circles.
|
11 |
Without data and testing this [i]]looks[/i] like a bunch of theory-crafting. Buoy has been a theoretical threat to Tanks for a long time, but it has taken many other factors for this to even come up as an issue. I know you have game knowledge behind what you're saying, but it takes a lot more work to devise a good solution than to point to a problem. I want replays and testing so that we have the data to do something reasonable, rather than breaking other matchups while running in circles.
|
| 12 |
\n
|
12 |
\n
|
| 13 |
Balance is iterative, and the simple-looking iterations are often the ones that move us towards blandness. For example, a Recluse with less range or a Tank skirmisher would mess with the identity of the factories. Maps are also iterative. Every map is an experiment, and not every experiment succeeds. Learning that a particular factory is really good on a particular map helps with future maps.
|
13 |
Balance is iterative, and the simple-looking iterations are often the ones that move us towards blandness. For example, a Recluse with less range or a Tank skirmisher would mess with the identity of the factories. Maps are also iterative. Every map is an experiment, and not every experiment succeeds. Learning that a particular factory is really good on a particular map helps with future maps.
|
| 14 |
\n
|
14 |
\n
|
| 15 |
I don't find the lessons of Sparkles Reef particularly surprising.
|
15 |
I don't find the lessons of Sparkles Reef particularly surprising.
|
| 16 |
* Amph and Hover are not intended to be equal plops to Ships. Maps that force this can be made, but success is not a guarantee. Sparkles Reef is mostly sea and impassible terrain, with the two exposed sandbar-islands in the corners being the only notable Amph/Hover zones. Ship is likely to be good on maps that are mostly sea.
|
16 |
* Amph and Hover are not intended to be equal plops to Ships. Maps that force this can be made, but success is not a guarantee. Sparkles Reef is mostly sea and impassible terrain, with the two exposed sandbar-islands in the corners being the only notable Amph/Hover zones. Ship is likely to be good on maps that are mostly sea.
|
| 17 |
* Siren is particularly good on Sparkles Reef because it is a small map with many chokepoints. Land units/compositions that fill the same role would be similarly good on a dry version of the map.
|
17 |
* Siren is particularly good on Sparkles Reef because it is a small map with many chokepoints. Land units/compositions that fill the same role would be similarly good on a dry version of the map.
|
| 18 |
\n
|
18 |
\n
|
| 19 |
I file @Bakuhatsu's thoughts on Venom, Felon and Siren away in the "something to keep an eye on" cabinet, alongside Buoy vs. Tank. The comparison between Blitz and Venom [i]is[/i] very interesting, yet the stats don't support a noticeable winrate for spiders (45% under the same filters as above) and Venom hasn't dominated in games I've seen or played. It's good, sure, but it isn't like old incarnations of monospam Blitz when it was a little more powerful. Maybe it shows how important speed is, or weight. In any case, I keep an eye on these things and wait to see whether potential problems will be demonstrated to be actual problems.
|
19 |
I file @Bakuhatsu's thoughts on Venom, Felon and Siren away in the "something to keep an eye on" cabinet, alongside Buoy vs. Tank. The comparison between Blitz and Venom [i]is[/i] very interesting, yet the stats don't support a noticeable winrate for spiders (45% under the same filters as above) and Venom hasn't dominated in games I've seen or played. It's good, sure, but it isn't like old incarnations of monospam Blitz when it was a little more powerful. Maybe it shows how important speed is, or weight. In any case, I keep an eye on these things and wait to see whether potential problems will be demonstrated to be actual problems.
|
| 20 |
\n
|
20 |
\n
|
| 21 |
It's like Flail. Flail has terrible stats on paper, but always looks really effective. The lesson I learn here is "it's interesting that stats can combine in that way" rather than "Flail must be bad regardless". Reworking it preemptively would be a waste of a unique unit.
|
21 |
It's like Flail. Flail has terrible stats on paper, but always looks really effective. The lesson I learn here is "it's interesting that stats can combine in that way" rather than "Flail must be bad regardless". Reworking it preemptively would be a waste of a unique unit.
|
| 22 |
\n
|
22 |
\n
|
| 23 |
Spider vs. Shield looks like a more credible problem. There is more data behind it and Shield has always had a lack of ways to deal with Recluse. Venom makes Bandit pretty sad. The next step is to grab a partner, pick a normal seeming map, and try out a bunch of strategies in a fairly controlled environment. How good is Racketeer? Clumped vs. Spread Bandits? Greed into LLT? Switch sides regularly. If you find an approach that works, great, but if you don't you have still produced a set of replays that explore the interactions. Start a thread and share them around. Maybe other people can suggest some things you missed, or do their own testing. Look at the common problems with the matches and use the tweakunits modoption to experiment with patches. Try different maps to see what sort of terrain is good for Shields.
|
23 |
Spider vs. Shield looks like a more credible problem. There is more data behind it and Shield has always had a lack of ways to deal with Recluse. Venom makes Bandit pretty sad. The next step is to grab a partner, pick a normal seeming map, and try out a bunch of strategies in a fairly controlled environment. How good is Racketeer? Clumped vs. Spread Bandits? Greed into LLT? Switch sides regularly. If you find an approach that works, great, but if you don't you have still produced a set of replays that explore the interactions. Start a thread and share them around. Maybe other people can suggest some things you missed, or do their own testing. Look at the common problems with the matches and use the tweakunits modoption to experiment with patches. Try different maps to see what sort of terrain is good for Shields.
|
| 24 |
\n
|
24 |
\n
|
| 25 |
This work is part of making Spider viable. It was fine before Venom became viable (and moreso, before Redback existed), but apparently the aim was to make Spider ploppable, so there is work to be done. The answer to this problem is unlikely to be a Recluse range nerf, but even if it was I'd want to know it rather than guess. Spider still seems to be losing most of its matchups, although it is fairly even with Jump, which was an intended starting point, so the testing is required for us to find a solution that is specifically aimed at the Shield matchup. Given that Shield is also struggling, a better change could be a targeted Shield buff.
|
25 |
This work is part of making Spider viable. It was fine before Venom became viable (and moreso, before Redback existed), but apparently the aim was to make Spider ploppable, so there is work to be done. The answer to this problem is unlikely to be a Recluse range nerf, but even if it was I'd want to know it rather than guess. Spider still seems to be losing most of its matchups, although it is fairly even with Jump, which was an intended starting point, so the testing is required for us to find a solution that is specifically aimed at the Shield matchup. Given that Shield is also struggling, a better change could be a targeted Shield buff.
|
Zero-K v1.9.1.0 - Mod Page and Slower Kodachi