18 days have passed since my last look at the top GDI decks. A lot has happend since then: The Grenadier was introduced, and a larger balance patch had a large impact on the meta*. Time for a renewed look at what the high-ranked players bring to the battlefield!
*: Wikipedia on "meta": "the metagame consists of the deck types that are currently popular and expected to show up in large numbers"
How to read the table:
- Frequency is how often a unit / commander appears in a deck
- "Low Lvl / High WR" is the average of (a) 12 players with low levels (10-11) and (b) the 17 decks with 80%+ winrate. The idea is that both is indicative of "good decks", so the mix of both can be an interesting subset of players to look at:
- Low level users probably have decent skill (and need good deck selection) to still be in the Top 50
- High Winrates indicate that these players know what they are doing and have a strong deck
- Δ vs Mar 12 is the change in frequency to the previous analysis, given in percentage points (example for the Militant: this time, 40% of decks contain it, this is 20 percentage points higher than last time, when it was just 20%)
- Δ vs average shows how the winrate and unit levels of decks with these cards deviated from the average; for example, decks with Bikes had a 2% lower winrate and and 0.1 lower levels than the average for the Top 50 decks
Observations on Infantry:
- With the Sniper buff, Machine Gun buff and Grenadier introduction, GDI infantry play has become highly diverse, which I'd generally consider a good thing.
- Winrates seem balanced between the units. The higher Grenadier winrate is very likely due to it being played by a couple of strong players in the Grenadier event, which props up it's winrate (strong players rerunning their 9-1/10-0 runs)
- Rifle usage grew, albeit not as much as the Nod Militant – explanations might include:
- Nod playing more Militants means they play less Flame Troopers, so the respective need
- Sniper making the expensive Shock/less attractive
- MG popularity might make people feel they need a Rifle opener and then rather Talon as secondary anti-Infantry
- GDI also wanting to save Tiberium in early game for their own tech (e.g. Sandstorm)
- Jumpjets (slightly nerfed vs harvesters) and Shock Troopers are less used; both had been ocre units and now you see way more Rifle/Missile
- I was a bit surprised that MG usage was not higher, as many people seem to complain on Reddit and Discord about them
Observations on Vehicles:
- The Pitbull stays highly popular as GDI's highly micro-able allrounder vs light vehicles and air
- Predator Tank usage is slightly down after the nerf; the cost nerf might have hit it harder than it's counterpart due to it's lower speed (early timing attacks now are much more easy to defend)
- The APC has grown a bit in popularity; it was indirectly buffed with nerfs on Tanks and MLRS
Observations on GDI Air:
- With the exception of drones, all units see a good amount of usage; winrates are also in line
- Hammerhead usage went down a good bit, possibly because there were fewer Nod air decks around due to the popularity of the Avatar
Observations on GDI Tech:
- All Tech units outside of the Disruptor see play with this selection of decks
- The Sandstorm was buffed and increased in popularity accordingly; as a relatively cheap multi-purpose tech unit, it's both playing the role as the sole tech unit in an otherwise Aggro deck, or as part of some Tech combo
- Juggernaut decks see play in some variants (e.g. with Zone Troopers, or even with Wolverine and Kodiak); from my experience, they are hard to get out (harder than Avatar, partly because you need to protect the Juggernaut with other units), but once the Jugger player gets their desired setup, it can be hard to break
Observation on Commanders:
- Unsurprisingly, Jackson is the most used commander; it's very versatile and can help Aggro decks, Air decks and Tech decks
- Liang usage has grown significantly (albeit the data does not show this). In the past, it was mostly seen with the (rare) dual harvester tech decks, but now it's also used for MG decks and even for an Orca deck (player: TonyCasto)!
Observations on Deck composition:
- Notation: X-X-X-X as in # infantry, # vehicle, # air, # tech; -0 at end is omitted
- Half of decks (12 each) are 2-2-2 or 2-3-1
- Next in line is 2-3-0-1 with 4 decks
- No other "pattern" is used by more than 3 decks, so the diversity is very high
- The patterns with 3 decks each are
- 3-0-0-3
- 3-0-2-1
- 3-1-2
- 2-0-3-1
Is it an issue if a certain unit appears in many decks? Not necssarily. High usage rates not always indicative of a unit being overpowered. It can be good for a game to have a number of core units that are often played. Broodwar, the most established and best-balanced RTS out there, certainly has core units that are played every single game and others that only see fringe use. As long as the gameplay around the core units is diverse (both in playstyles and deck composition), I do not see an issue with e.g. 96% of Nod decks containing Laser Troopers.
Unit Levels & Winrates
In the chart below, you can see the distribution of deck win levels (rounded to .5) and their winrates. As you can see, the most common deck in the Top 50 is an L12 deck. And winrates don't deviate that much, which is of course because the medal count is already an indicator of strength (so the lower level players in Top 50 partly compensate their lack of unit levels with their skill).
Algorithmically generated deck
I compared the unit usage rates between all, and the "low level or high winrate" players. If you then take the units that are more popular with the possibly stronger group of players, you get to the following deck:
- Missile, Sniper
- Dog, Pitbull, Tank
- Orca Bomber
- Strongarm or Jackson
The "anti deck", which is less played by the supposedly stronger players would be:
- MG, Shock
- MLRS, APC
- Mohawk
- Juggernaut
- Liang
Yes, this indeed looks awkward... Wonder what winrate Bikerush might achieve with it on ladder? ;)
How I selected the decks of players:
- Ideally, it's the recently played deck with a large sample size and high or representative winrate for the player
- If there are multiple decks with large sample sizes played this season by the player, I'll chose the one with the higher winrate
- If winrates are similar, I chose the more recent one
What issues can you have with this analysis?
- I don't analyse all decks of the top 50, but just one per player (see above)
- Champions & Blitz might mean certain specific decks are played a lot (e.g. for repeated reruns), which can have high winrates. In some cases, I thus might select such a deck over the most appropriate ladder deck of the player
- High winrates for a unit is not necessarily conclusive of a unit being "overpowereed", it can also be indicative of strong players liking to play with the unit
- But directionally, I feel this data is pretty good to get a picture on which units and which deck compositions are seen as viable / strong
Related links:
No comments:
Post a Comment