AI War - Strong Vs Weak
From ArcenWiki
What Exactly Do "Strong Vs" And "Weak Vs" Indicate?
A: Contrary to what a lot of new players seem to assume, these are not indicative of "bonuses" or "penalties" against other ship types. Rather, this is empirical performance data based on the results of simulated matchups. Here's how it works:
1. The simulation takes equivalent numbers of each type of ship, and then makes them fight.
2. By "equivalent numbers" this means that the relative ship population caps ARE taken into account -- so the simulation would be 10 bombers against 10 cruisers, 10 cruisers/bombers against maybe 3 or 4 vampires, or 10 cruisers/bombers against maybe 30 laser gatlings, based all on their relative pop caps.
3. The results of all the various matchups are stored in the game data and reported to you through the Strong Vs and Weak Vs sections of your hovertips. The percentage is how much health the victor still had left when all of the losing ships were destroyed.
- If your ship is 95% Strong against an enemy ship type, you know that given equivalent numbers of the two ship types, your ships will win and only take 5% damage.
- If your ship is 50% Weak against an en enemy ship type, you know that you will probably need at least 2X the normal equivalent number of ships to even force a draw. If you have 3X the normal equivalent number of ships, then you will almost certainly win, but you'd be likely to take severe losses.
In this way, the Strong Vs and Weak Vs can help you really predict what the outcome of battles are, and what the best counters to enemy ships will be. There's so much else to focus on that we felt like having this information right in the interface was very needed.
Same Type Vs Same Type: The effectiveness of any given ship type against other ships of its same type is not generally reflected in Strong Vs and Weak Vs. This is because, given two same-size groups of identical ships, you are likely to have one ship barely survive from whichever group shoots first (unless there was superior tactical maneuvering, but the Strong/Weak simulations intentionally do not account for this since it is too dependent on player skill).
Some ship types are very strong against themselves, and others are unable to even hit ships of their same type (Shield Bearers, for instance, are unable to penetrate the shields of other Shield Bearers). So sometimes a ship can be an effective counter against itself, but not always. And there is usually a better option found in another ship type. Snipers are one big exception to this, since their extreme range makes it necessary that they be an effective counter to themselves.
Non-Military Ship Types: The Strong Vs and Weak Vs only includes military ship types, turrets, and other things of that sort which can fire. Some ships are stronger than others against other ships such as harvesters, command stations, constructors, force fields, or other heavy defenses. Mostly it just takes experimentation to see what works best with these, but some ships that have a specific weakness will note it.
Force Fields, Fortresses, And Other Heavy Defenses: Bombers and Space Tanks both have very significant bonuses against these types of ships. This is generally noted in the hovertext for the target itself.
Starships: Fighters have a moderate bonus against starships, and tend to be an excellent counter for them. For some of the more extreme AI-only starships (such as the Core Starship), other measures may also need to be employed (Mobile Force Fields, other ships pitching in, etc).
Raw Bonuses
Why Use Strong/Weak Vs Instead Of Just Showing The Raw Bonuses?
A: A number of players over the life of the game have wondered why, when there are raw bonuses that are used under the hood, the game shows Strong Vs and Weak Vs data in the hover tooltips instead. This is sort of related to the question of why there are bonuses in the first place, and how those fit into the overall design of the game.
The short answer is: these bonuses are extremely, extremely misleading when viewed on their own. Or, put another way, you have to do an absolute ton of mental math when looking at the various raw stats of a ship to really understand how it will perform against other ships. And, generally speaking, there are enough factors that you'll overlook something and come to a slightly-to-majorly wrong understanding. The Strong Vs and Weak Vs data runs a literal simulation pitting some ships of one sort against equivalent ships of the other sort, and records the result. Even this isn't a "master number" that tells you everything you need to know, though.
What Exactly Is The Raw Bonus Data Indicating?
The bonuses are, simply, linear multipliers. So if standard fighters have a bonus of 12 against raiders, then they will do 12x as much damage to raiders as they normally do to any other ships. There is a ton of information of why we designed it this way, and what exactly this is simulating also here on the wiki.
Where the Strong/Weak Data Is Weak Or Misleading
Thus the strong/weak view of things takes out all the ambiguity except for:
- How the ships will perform relative to one another if they are at REALLY far ranges from one another, or really close range (in the simulation used for strong/weak, they start at around 20,000 distance from one another).
- Needless to say, if a bunch of ships pop out of a wormhole right on top of a bunch of other ships, this can lead to wildly different results based on shorter-ranged ships (or area of effect ships) not having to waste time closing distance while being shot by the longer-ranged ships.
- In other words, shorter-ranged and area-of-effect ships are very much undervalued in those sorts of close-combat situations in the strong/weak data. Melee ships and electric shuttles are much more powerful when used in these sorts of situations.
- How the ships perform if the player uses any tactics other than just running the two groups at one another.
- This is of course something that we can't account for, but there are a number of ships that seem undervalued by the simulation, but which advanced players can use more effectively by "kiting," making surprise raids, or so forth.
- How the ships perform if they are used in drastically out-of-balance relative numbers.
- How one fighter does versus one MLRS isn't always completely predictable from the strong/weak data, nor is how 150 fighters do against one MLRS.
- But, the strong/weak data gives a pretty good idea in this case, anyway -- it's just not a completely linear scale when you have nonequivalent concentrations of ships -- with ships that get cluster bonuses or penalties, that is even more of an effect.
- This data doesn't address the economics of battles. Maybe one Ship A is 10% poorer against Ship B than is Ship C, but Ship A costs 50% less than Ship C to build. The strong/weak data is a military simulation only, not paying attention to the economic side of things.
- This data doesn't address how multiple ship types will do in combination with one another. If you use two or three ship types together, often they will be much stronger together than they would be individually: the sum is greater than the whole of its parts.
Why The Strong/Weak Data Is Still The Best Brief Guide
In the end, what this whole thing boils down to is this: AI War uses a very complex simulation with a ton of factors, and there's no way to boil ship-to-ship performance down to a single number that will be relevant in all types of ship to ship encounters. The strong/weak data was arrived at with much testing and with player feedback, and provides the best possible brief summary of how ships will interact in most situations, but they still under or over value ships in certain circumstances as noted above.
As the US mortgage industry unfortunately demonstrated, you can use single rollup numbers like this as a guide, but not as a sole source of making decisions. There is no substitute for experimentation and observation -- trying out funky tactics and then seeing if they work or not. Or watching what happens when the AI sends 1000 infiltrators against your planet, and thus figuring out clever ways to counter them. That's why all of the other stats (ship speed, number of simultaneous shots, etc) are also shown in the game. Paired with the strong/weak data, generally that's all you need to know to make informed assumptions about how ships will perform in various uneven matchups.
Why The Raw Bonuses Are So Misleading
So, bringing this discussion back around to the original question, why not just use the raw bonuses to make your decisions, instead of the strong/weak data? Well, the main problem is this. Based on a long list of factors, ships inherently fall into a strong/weak hierarchy without any of the extra bonuses. What those bonuses do is modify the existing hierarchy, simulating armor, special weaknesses, etc. In other words, the bonuses don't create the hierarchy, they only augment it.
To take an example, if you were to look at standard fighters and were to see that they had (for instance) at 12x bonus against zenith viral shredders, what does that really tell you? Well, it tells you that the fighters do 12x their normal damage against viral shredders, and that's obviously a good thing for the fighters. But you can't assume that fighters will even win against viral shredders based on that fact alone, and that's why this becomes so misleading so quickly:
- Perhaps that 12x bonus is there to make fighters absolutely destroy viral shredders
- Maybe it's there to make them more evenly matched.
- Maybe it's there to make them not die so incredibly quickly against their otherwise-overpowered foe.
Just looking at the bonus number, you don't have any context to be able to make larger assumptions about ship vs ship performance. Players innately assume that the first situation, above, is the correct case -- but that's often incorrect, because as a designer I am frequently using all three reasons for creating ship bonuses. In the interest of not slowing the gameplay down, I rarely use ship-to-ship penalties, as that can make battles between two ships that are ill-suited to fight one another take forever.
Generally if Ship A needs to be tweaked to be a bit weaker against Ship B, and Ship B is another military ship, Ship B is given a bonus against Ship A, rather than giving a penalty to Ship A. The main time we use penalties for the attacking ship is if the defending ship doesn't have any firepower, or if the defending ship would just get chewed through immediately by the attacker despite any firepower adjustments due to range or other considerations.
Still not convinced? If you're wanting to figure out which of two ship types will win in a given engagement, here are all the various factors you will have to consider, which the strong/weak vs data inherently accounts for (and this isn't even a full list):
- Attack power of both ships
- Attack Range of both ships
- Attack Speed of both ships
- Number of secondary shots beyond the first, if any
- Any shot type immunities of both ships
- Health of both ships
- Movement Speed of both ships
- Any combat-relevant special abilities (such as tractor beams, cloaking, health regen, cluster bonuses/penalties, etc -- there are dozens of these) of both ships
That's quite a lot of factors to try to consider for some "quick mental math" to find out which ships are stronger than another ship! In alpha versions of the game, we found that players routinely made bad choices because they overestimated their ability to extrapolate from the bonuses when we were showing those in-game. Of course, players quickly realized that they were miscalculating, and this was in turn quite frustrating. The strong/weak vs data was created as a way to solve that intense frustration they were experiencing, and to provide a much better set of guide numbers than they could quickly construct on their own.
Exporting Raw Bonuses
How Do I Get Access to The Raw Bonuses And Other Ship Stats?
A: There is a page on the community wiki that has a lot of that data. However, it's not always completely up to date, as it's something that is maintained by the community, not the developer directly. But! You can always get the data for the very latest versions of the game at any time (in Excel XML format). Here's how:
- 1. Open up a copy of AI War and start a new game.
- 2. Press F3 so that the debug text is shown.
- 3. Press Ctrl+Shift+F8 to initiate writing of the "secondary excel exports."
- 4. The game program will immediately close and will have written a variety of interesting files into a new Data\ directory inside the game data directory.
- Your game data directory the file path in the textbox at the bottom of the settings screen.
This data is in Excel XML format, and is the raw data that was used to create that page in the first place. It also includes some analysis-style Excel XML files that show how many ships each ship is strong vs and weak vs, and things of that nature.
