The Hilbert Guide
A cookie cutter approach to PVP
The Hilbert Guide
A cookie cutter approach to PVP
Random rambles through PVP in Star Trek Online by Hilbert@mancom
So Cryptic has decided to install pug queues without matchmaking. This is particularly disappointing because it is surprisingly easy to balance matches without pre-formed groups. Such matchmaking is already happening for private pug matches via data from the PVP Leaderboard.
The current algorithm for that balancing tool calculates an imbalance score for each possible 5v5 that one can form from 10 players (or less) and then returns three setups with the lowest imbalance scores.
So how does one get this imbalance score? First the algorithm discards all setups that have a lopsided science captain distribution. This means that if there is an even number of sci captains among the ten players, both teams will get the same number of sci players and if there is an odd number, one team will have one more than the other, but there will never be a setup where one side has two more than the other.
Now after discarding the setups with unbalanced SNB distribution, the imbalance score for each possible pairing “Team 1 vs Team 2” is calculated as
score = [
[(Damage1 - Damage2) / (Damage1 + Damage2) + (Kills1 - Kills2) / (Kills1 + Kills2)]^2
+ [(CrossHealing1 - CrossHealing2) / (CrossHealing1 + CrossHealing2)]^2
+ [(SelfHealing1 - SelfHealing2) / (SelfHealing1 + SelfHealing2)]^2
+ [(VapeScore1 - VapeScore2) / (VapeScore1 + VapeScore2)]^2
]
If there are no engineers in the game, this is the final imbalance score, if there is at least one engineer, the score is modifed by an additional factor where EngineersX denotes the number of engineers in team X.
score * [1 + 1/5 * abs[(Engineers1 - Engineers2) / (Engineers1 + Engineers2)] ]
This is like a “soft” version of the hard science captain distribution. An imbalance in the number of engineers is allowed, but setups with such an imbalance get an additional imbalance penalty to discourage a lopsided distribution of the engineers.
Now what do the other variables in the imbalance score mean?
DamageX is the sum of the average net dps values (i.e. after taking resists into account) of the players in team X. For each payer this is calculated as the average over all matches played during the last week or during the last 25 matches, whichever provides more data.
KillsX is the sum of the average weighted kills per match of the players in team X. But what is a weighted kill? Whenever a team scores a point in a match, this point is distributed among the team members as weighted kills; depending on how much of the total damage against the defeated target during the 30s prior to the kill was done by each player, they get that fraction as a weighted kill.
SelfHealingX and CrossHealingX are the sums of the hps values of the players in team X, where hps is recorded separately for self- and crosshealing.
VapeScoreX finally is an attempt to avoid placing too many vapers on the same team. Again it is the sum of the VapeScores of the players in team X. It is calculated for each player as
VapeScore = sqrt[ Kills * dpsHull / (dpsRaw + 1) + Kills / ((hps + 1) / 200)] * {1/3 if the player is not a Tac; 2/3 if it is a Tac} + .001
If this values is bigger than 3.5, it gets reduced to 3.5. What this does is give a high value to players who have a very good hull/raw damage ratio and very low healing (self- and crosshealing combined). Only tacticals can truly be vapers, so they get twice the value that the other classes get. The + .001 is only there to avoid division by zero errors in the imbalance score calculation.
Finally the imbalance score is multiplied by 40000000 to get the numbers into a range that humans can easily grasp. This does not change the ordering of the scores and is purely for display purposes.
So what the algorithm does is that it minimises the sums of the squares of the differences between the teams in four categories (while forcibly balancing the number of sci captains and penalising engineering captain imbalances):
Just enter some names into the balancing tool and see for yourself!
A Photonic Shockwave is one of the most useful tools in organised PVP as it can be used to break extends and interrupt certain abilities. Usually PSW is found on support ships, but one can also equip it on some escorts.
| Fore Weapons |   |
| Deflector |  |
| Impulse |  |
| Warp |  |
| Shields |  |
| Aft Weapons |    |
| Devices |   |
| Engineering Consoles |   |
| Science Consoles |    |
| Tactical Consoles |  |
Ship Stations
See Abbreviations for the full names of the bridge officer abilities.
The cruiser in the guide is a multi-purpose build that can do everything. Hence it is not really optimised for any particular task. This makes it a good generic template for beginners, but if one wants a more specialised cruiser, a different build can be useful.
The following build should be used by an engineer who is the main healer of a team. This means that this cruiser will rely on help from the team to get ST for clearing SS/SNB.
| Fore Weapons |  |
| Deflector | |
| Impulse | |
| Shields |  |
| Aft Weapons | |
| Engineering Consoles |  |
| Science Consoles |  |
| Tactical Consoles |  |
Depending on the rarity of the maintenance engineers you have available you can use three blues or two purples. The goal is to reduce ET3 to its global cooldown of 15s.
Ship Stations
Skill EPtW1 only to level 6 (60s cooldown). Put EPtS1, EPtS2, FAW1 and FAW2 on autofire (tray 7). Use TT only if it is really necessary. Usually your team should provide the TT. Your job is to distribute as many heals as possible, therefore your ET has priority over TT.
Make sure that you have set the option Auto attack (in Options -> Controls) to "Toggle, non-combat cancels", so that your your weapons keep firing even when you are switching between allies to heal them.
Not every shot you fire in STO hits its target. How many shots miss is determined by a formula that takes the bonus accuracy (Acc) of the attacker and the bonus defense (Def) of the attacked player into account.
Bonus accuracy can be obtained by speccing into Starship Targeting Systems and by using weapons with [Acc] modifiers. The [Acc] modifiers only apply to the individual weapons and thus do not show up in the character's info tab.
Bonus defense can be obtained by speccing into Starship Maneuvers, by increasing the ship's speed and from some abilities like APO and Evasive.
To determine the hitrate of a weapon the difference (Diff = Acc - Def) between bonus accuracy and bonus defense is used. This means that same values in Acc and Def cancel each other out. When Diff is 0, then the hitrate is 100%. If Diff > 0, then the hitrate remain at 100% and supposedly there is some overflow into CritD and CritH. Against moving targets Diff will be negative most of the time though. When Diff < 0, the chance to hit is determined by formulas that differ between beam and cannon weapons. Make sure you use decimal representations of the percentages when you use these formulas, i.e. 12,3% = 0,123.
For cannons, the chance to hit is
Hit = 1 / (1 - Diff)
For beams, the chance to hit is
Hit = 1 / (1 - Diff) * 0,75 + 0,25
To illustrate the difference, the following graph shows the theoretical hit rates for beams and turrets (which are cannon weapons) and real world data points that were gathered in a controlled 1v1 environment.
According to a STOked episode where CaptainGeko gave an incomplete description of the accuracy mechanics (only for cannons, he did not touch the beam accuracy issue and the STOked guys did not ask him), there is a minimum hit chance of 25% that is independent of the Diff value. Since 25% hitrate is only attained at Diff = -300% for turrets, there is no way to test whether this is actually true, since one cannot get a defense rating that is high enough to produce such a Diff.
This 25% minimum chance explains why the beam curve looks the way it does. The hitrate formula for cannons gives values that asymptotically approach 0 when Diff goes to negative infinity which means that it is necessary to implement a hard cut-off to get the 25% minimum chance. The formula for beams on the other hand goes asymptotically to 25% and thus does not require a special cut-off. In a way the beam formula is the smarter way to implement the 25% minimum chance, but obviously the curve declines more slowly and thus for identical Diff values, beams will hit more reliably than cannons.
Special cases are FAW and CSV. FAW used to have a 100% hitrate and after a "fix" it now takes a player's innate accuracy rating into account, but ignores the weapons' [Acc] modifiers. And despite being a beam ability, it uses the cannon formula. This means that currently FAW has the same hitrate as a turret without any [Acc] modifiers.
CSV has a higher hitrate than regular cannon fire (but not 100% like FAW used to have). As far as I know, the reason for this is not known with certainty. I suspect that it has to do with how CSV's multi-target damage is calculated since it appears that against a single target CSV behaves as expected from cannons. One would need to conduct tests in a controlled environment with multiple ships of known defense rating as targets to gather data to formulate a theory.
Giving advice regarding science ships is difficult. In contrast to escorts and cruisers there is no default layout for science ships that you can find (with minor modifications) on almost every ship. Another difference to cruisers and escorts is the fact that there are practically no non-science captains flying science ships in a premade setting.
| Fore Weapons | |
| Deflector | |
| Impulse | |
| Shields | |
| Aft Weapons | |
| Engineering Consoles | |
| Science Consoles |  |
| Tactical Consoles | |
Since this is just the informal blog, the following proposal for a science ship is less detailed than the builds in the guide and aimed at more advanced players. This means that instead of "easy" abilities like RSP or AtSIF, you will see an Extend 1.
Ship Stations
Here is an idea for a spec: > Spec at STO Academy <