Engineers are detected by their use of RSF/MW, and Scis are detected by their Photonic Fleets.
If you had easy matches and did not use RSF/MW, one cannot tell the difference between an Eng and a Tac. Also I only recently resolved an issue with the log file output of german STO clients that prevented correct Eng detection.
You know, Hilbert, if you wanted an extra source of information to determine usage of abilities you could probably set up something using demos. Although it's true that they're mainly focused for creating movies (and as such lack any sort of ability usage data), you could probably infer a great deal from the vfx that is generated (e.g. identify when SNB is used and who it's used on; see who's deploying their Scattering Fields when, etc.). It'd probably require lots of extra parsing and correlating vfx IDs to the abilities that they're used by, though.
I remember trying this myself a month or so back; I could see when flashy things like Scattering Field were deployed, and there'd be an entry in the demo that instructed the player to create and center the vfx on my ship at that time.
http://hilbertguide.com/leaderboard/?encounter=2013-10-27%2003:42:53&
At the time I posted it, this match, a copy of the one before it on my score list, is listed as a Zero. It's driving everyone on it's score down, figured you might like to know. I'm assuming there were too many discrepancies between it and the other copy and it caused an issue. Figured you might want to know to avoid this in the future.....and because I don't like my score being so low! Thanks XD
Apparently both players started reporting within one second of each other and thus the first data set was still incomplete when the second player wanted to start a new one and my merging algorithm could not yet detect the duplicate. And then the finalising command was applied to the wrong encounter since I did not expect the mergin algorithm to fail.
I have removed the duplicate encounter and will fix the underlying problem once I have some time.
It seems one of my matches on Skittles was double posted (in a way, it was one match but it came out with 2 completely different stat lines).
Also my matches are recording for both my characters I pvp with Ne'Hi and Skittles but it seems Skittles name is not appearing in the 24 hour list although NeHi's are even though I have higher scores with Skittles than on Ne'Hi and the required number of games to appear on the last 24 hour list.
I still think there is a bug with it. Last time i used it, a few matches we had was 5v5, everyone active, balance was well above .25 (more like .40 or .50) and still it wasnt submitted or even recognized by the tool.
I even removed the log and rebuild it, something must be off.
The double post thing happens if you are out of combat for more than 60s, and far enough away (like after a respawn) from any combat that your log doesn't record any activity for 60s. Basically it creates 2 separate encounters. This can also happen if you hit the ceiling and both teams agree to fly back down. If there's no shooting for 60s Hilbert's parser ends the encounter. When shooting starts again, a new encounter begins.
As already mentioned, the tool decides that an encounter has ended if there are 60s without combat between players. This is how I attempt to detect the end of a match because the log does not indicate map changes. Occasionally this leads to partial matches being uploaded with scores like 12-5 that look to the tool as if they could be complete matches where some kills have been missed due to log distance limitations.
I think there is still a problem in the server side code that can lead to duplicate matches of two players start uploading the same match at almost exactly the same time , so if you happen to see this, I would be great if you could send me the link to the duplicate encounters so that I can fix it in the database. Same goes if you see anything else really weird, or if you get detected as a sci or eng despite not being that captain type (tac is the default if the tool cannot detect signs of a player being an eng or a sci).
Also already mentioned is the 0.25 cutoff for matches that are included in the daily/weekly/monthly lists. This is to avoid that really lopsided matches skew the average scores.
Just let me say I am impressed by the mere fact that some community member worked this out. I am pretty much a non-player or at least a non-PvP player these days, but I find this very interesting.
Star Trek Online Advancement: You start with lowbie gear, you end with Lobi gear.
As already mentioned, the tool decides that an encounter has ended if there are 60s without combat between players. This is how I attempt to detect the end of a match because the log does not indicate map changes. Occasionally this leads to partial matches being uploaded with scores like 12-5 that look to the tool as if they could be complete matches where some kills have been missed due to log distance limitations.
The results of the multiple PvP encounters occurring during a match can be merged into one record, by checking the boxes and merging the selections applicable to the specific pvp encounter. I may be wrong, but I also thought you could change the length of time before the ACT dps counter stops an encounter, as well as manually ending it.
Also, what about going as far as you can with utility and style? Just keep the impact on the players overall score minimal since the combat log doesn't track all abilities. It could make support players a little happier, even if their score is low.
I think the task is impressive but I'll never buy into anything that involves a group of players claiming some items, tools or abilities should not be used. Sounds way too much like qq and tac driven gameplay
* Please remove the negative style for Nimbus Pirates. The subnuc has been removed
True. I still think that Nimbus pirates are a completely unnecessary element of spam, but I suppose that I can now significantly reduce the style penalty for using it. (Also, style has a rather small impact on the overall score unless you really push it.)
* What about giving heals that increase damage resistance, like A2SIF and HE, a slightly higher ranking than ET and ST?
I don't think that attempting to tweak such details will do much good at this point. The overall formula that relates damage, kills, healing and utility needs to get better before something like that can reasonably be done.
Also, what about going as far as you can with utility and style? Just keep the impact on the players overall score minimal since the combat log doesn't track all abilities. It could make support players a little happier, even if their score is low.
The problem with giving utility a greater impact is that it starts favouring people who simply spam the relevant abilities and thus in the end leads to high scores even for some useless players. It is nigh impossible to tell from the combat log what the exact impact of the utility abilities are. Of course one could do some fancy attempts at correlating kills and utility events, but this would require significant effort - and a lot of coding of the type that is a chore, not fun.
Well ehhhh hilbert ehhh this is ehh quite a good tool u made em... i reinstalled my windows and lost key/program
U need 2 distribute it to more ppl cuz ive played quite some matches and didnt got reported once. Its still low populated. Isnt it out of beta by now for other ppl to use ?
The problem with giving utility a greater impact is that it starts favouring people who simply spam the relevant abilities and thus in the end leads to high scores even for some useless players. It is nigh impossible to tell from the combat log what the exact impact of the utility abilities are. Of course one could do some fancy attempts at correlating kills and utility events, but this would require significant effort - and a lot of coding of the type that is a chore, not fun.
Hilbert.
You mention that damage which leads to an actual kill is rated higher. Does this also count for these "style" abilities? Like a well timed CPB or sensor scan helps to make the kill? If no, is it possible to implement such?
You mention that damage which leads to an actual kill is rated higher. Does this also count for these "style" abilities? Like a well timed CPB or sensor scan helps to make the kill? If no, is it possible to implement such?
The examples you have chosen are already part of the problem with utility abilities. Neither CPB nor sensor scan is visible in the combat log, the same goes for other shield drains (glider, tachbeam, ...) and of course SNB which has a huge impact, yet remains entirely invisible in the log. (I suppose that with massive correlation techniques one could guess when an SNB happens, but it would still remain essentially impossible to figure out which player fired it.)
Rating the importance of a GW or Tykens for a kill just from the log data is nearly impossible. Tractor beam is maybe the only thing that comes close to being an effective utility skill that is visible in the log, but since PH and APO are not visible in the log, I don't even know whether the tractor beam was actually holding the target.
With the current state of the log file, I think it is nigh impossible to generate meaningful utility-kill relations.
The examples you have chosen are already part of the problem with utility abilities. Neither CPB nor sensor scan is visible in the combat log, the same goes for other shield drains (glider, tachbeam, ...) and of course SNB which has a huge impact, yet remains entirely invisible in the log. (I suppose that with massive correlation techniques one could guess when an SNB happens, but it would still remain essentially impossible to figure out which player fired it.)
Rating the importance of a GW or Tykens for a kill just from the log data is nearly impossible. Tractor beam is maybe the only thing that comes close to being an effective utility skill that is visible in the log, but since PH and APO are not visible in the log, I don't even know whether the tractor beam was actually holding the target.
With the current state of the log file, I think it is nigh impossible to generate meaningful utility-kill relations.
I have to give you praise for this tool to be available. Even though its far from being perfect due to lack of information provided in the combatlog, the work you have put into this is is amazing.
Very very good work, and if you need any help to test this please let me know.
"Better were the days when mastery o' space came not from bargains struck with eldritch creatures... but from the sweat of a man's brow and the strength of his back alone. Ye all know thi's to be true!"
Yo Hilbert, today i got a notification to patch to a new version, patched it, and ever since it is throwing an error that the application cannot be opened.
ERROR DETAILS
Following errors were detected during this operation.
* [3/8/2014 1:54:25 PM] System.ArgumentException
- Value does not fall within the expected range.
- Source: System.Deployment
- Stack trace:
at System.Deployment.Application.NativeMethods.CorLaunchApplication(UInt32 hostType, String applicationFullName, Int32 manifestPathsCount, String[] manifestPaths, Int32 activationDataCount, String[] activationData, PROCESS_INFORMATION processInformation)
at System.Deployment.Application.ComponentStore.ActivateApplication(DefinitionAppId appId, String activationParameter, Boolean useActivationParameter)
at System.Deployment.Application.SubscriptionStore.ActivateApplication(DefinitionAppId appId, String activationParameter, Boolean useActivationParameter)
at System.Deployment.Application.ApplicationActivator.Activate(DefinitionAppId appId, AssemblyManifest appManifest, String activationParameter, Boolean useActivationParameter)
at System.Deployment.Application.ApplicationActivator.PerformDeploymentActivation(Uri activationUri, Boolean isShortcut, String textualSubId, String deploymentProviderUrlFromExtension, BrowserSettings browserSettings, String& errorPageUrl)
at System.Deployment.Application.ApplicationActivator.ActivateDeploymentWorker(Object state)
The situation has been resolved. It seems most likely that it was a corrupted download during the automated updating process.
I have started adding some damage charts to the datamining page. It seems that Beam Arrays are the big thing these days and Disruptors and Phasers are the most popular energy types.
I intend to keep the charts with daily values for a while and then eventually switch to weekly averages once I have enough data.
Comments
If you had easy matches and did not use RSF/MW, one cannot tell the difference between an Eng and a Tac. Also I only recently resolved an issue with the log file output of german STO clients that prevented correct Eng detection.
I remember trying this myself a month or so back; I could see when flashy things like Scattering Field were deployed, and there'd be an entry in the demo that instructed the player to create and center the vfx on my ship at that time.
Just a thought.
At the time I posted it, this match, a copy of the one before it on my score list, is listed as a Zero. It's driving everyone on it's score down, figured you might like to know. I'm assuming there were too many discrepancies between it and the other copy and it caused an issue. Figured you might want to know to avoid this in the future.....and because I don't like my score being so low! Thanks XD
Apparently both players started reporting within one second of each other and thus the first data set was still incomplete when the second player wanted to start a new one and my merging algorithm could not yet detect the duplicate. And then the finalising command was applied to the wrong encounter since I did not expect the mergin algorithm to fail.
I have removed the duplicate encounter and will fix the underlying problem once I have some time.
Also my matches are recording for both my characters I pvp with Ne'Hi and Skittles but it seems Skittles name is not appearing in the 24 hour list although NeHi's are even though I have higher scores with Skittles than on Ne'Hi and the required number of games to appear on the last 24 hour list.
-Edit- Just checked your scores and yes, they are below .25
I even removed the log and rebuild it, something must be off.
I think there is still a problem in the server side code that can lead to duplicate matches of two players start uploading the same match at almost exactly the same time , so if you happen to see this, I would be great if you could send me the link to the duplicate encounters so that I can fix it in the database. Same goes if you see anything else really weird, or if you get detected as a sci or eng despite not being that captain type (tac is the default if the tool cannot detect signs of a player being an eng or a sci).
Also already mentioned is the 0.25 cutoff for matches that are included in the daily/weekly/monthly lists. This is to avoid that really lopsided matches skew the average scores.
* Please remove the negative style for Nimbus Pirates. The subnuc has been removed (http://sto.perfectworld.com/news/?p=1089141)
* Please add style/utility for the Assimilated Tractor Beam similar to Tractor Beam. It is recorded in the log.
* Why do I sometimes get a negative style when the only item I own which causes a negative is the pirates - and I don't use them anymore?
* What about giving heals that increase damage resistance, like A2SIF and HE, a slightly higher ranking than ET and ST?
The results of the multiple PvP encounters occurring during a match can be merged into one record, by checking the boxes and merging the selections applicable to the specific pvp encounter. I may be wrong, but I also thought you could change the length of time before the ACT dps counter stops an encounter, as well as manually ending it.
Good point. I will do that.
This should not happen. If you have such a match, it would be great if you could send me the combatlog so that I can investigate the details.
I don't think that attempting to tweak such details will do much good at this point. The overall formula that relates damage, kills, healing and utility needs to get better before something like that can reasonably be done.
The problem with giving utility a greater impact is that it starts favouring people who simply spam the relevant abilities and thus in the end leads to high scores even for some useless players. It is nigh impossible to tell from the combat log what the exact impact of the utility abilities are. Of course one could do some fancy attempts at correlating kills and utility events, but this would require significant effort - and a lot of coding of the type that is a chore, not fun.
U need 2 distribute it to more ppl cuz ive played quite some matches and didnt got reported once. Its still low populated. Isnt it out of beta by now for other ppl to use ?
Maybe just flag them as verified / unverified, in case I need to remove some results due to abuse?
Refracting Tetryon Cascade
Fleet Technician
Fleet Physicist
Fermion Field
Graviton Shield
If I haven't said it already, thank you for the leaderboard and the parser!
Hilbert.
You mention that damage which leads to an actual kill is rated higher. Does this also count for these "style" abilities? Like a well timed CPB or sensor scan helps to make the kill? If no, is it possible to implement such?
@rudiefix Feds: Rudiefix / Thron / Opa
@rudiefix KDFs: Lill / Xifeidur / Dehr / Ugly
@rudiefix Roms (KDF alligned): Chicita
Rating the importance of a GW or Tykens for a kill just from the log data is nearly impossible. Tractor beam is maybe the only thing that comes close to being an effective utility skill that is visible in the log, but since PH and APO are not visible in the log, I don't even know whether the tractor beam was actually holding the target.
With the current state of the log file, I think it is nigh impossible to generate meaningful utility-kill relations.
I have to give you praise for this tool to be available. Even though its far from being perfect due to lack of information provided in the combatlog, the work you have put into this is is amazing.
Very very good work, and if you need any help to test this please let me know.
@rudiefix Feds: Rudiefix / Thron / Opa
@rudiefix KDFs: Lill / Xifeidur / Dehr / Ugly
@rudiefix Roms (KDF alligned): Chicita
The charts are dynamically updated and will show more time with every passing day.
I'll probably add something similar for the major weapon abilites and maybe also energy types in the future.
ERROR DETAILS
Following errors were detected during this operation.
* [3/8/2014 1:54:25 PM] System.ArgumentException
- Value does not fall within the expected range.
- Source: System.Deployment
- Stack trace:
at System.Deployment.Application.NativeMethods.CorLaunchApplication(UInt32 hostType, String applicationFullName, Int32 manifestPathsCount, String[] manifestPaths, Int32 activationDataCount, String[] activationData, PROCESS_INFORMATION processInformation)
at System.Deployment.Application.ComponentStore.ActivateApplication(DefinitionAppId appId, String activationParameter, Boolean useActivationParameter)
at System.Deployment.Application.SubscriptionStore.ActivateApplication(DefinitionAppId appId, String activationParameter, Boolean useActivationParameter)
at System.Deployment.Application.ApplicationActivator.Activate(DefinitionAppId appId, AssemblyManifest appManifest, String activationParameter, Boolean useActivationParameter)
at System.Deployment.Application.ApplicationActivator.PerformDeploymentActivation(Uri activationUri, Boolean isShortcut, String textualSubId, String deploymentProviderUrlFromExtension, BrowserSettings browserSettings, String& errorPageUrl)
at System.Deployment.Application.ApplicationActivator.ActivateDeploymentWorker(Object state)
COMPONENT STORE TRANSACTION DETAILS
* Transaction at [3/8/2014 1:54:25 PM]
+ System.Deployment.Internal.Isolation.StoreOperationStageComponent
- Status: Installed
- HRESULT: 0x0
- Manifest: 5VLBKGDJ.Z78.application
+ System.Deployment.Internal.Isolation.StoreOperationSetDeploymentMetadata
- Status: Set
- HRESULT: 0x0
+ System.Deployment.Internal.Isolation.StoreOperationSetDeploymentMetadata
- Status: Set
- HRESULT: 0x0
+ System.Deployment.Internal.Isolation.StoreTransactionOperationType (27)
- HRESULT: 0x0
Fleet Admiral Rylana - Fed Tac - U.S.S Wild Card - Tactical Miracle Worker Cruiser
Lifetime Subscriber since 2012 == 17,200 Accolades = RIP PvP and Vice Squad
Chief of Starfleet Intelligence Service == Praise Cheesus
I have started adding some damage charts to the datamining page. It seems that Beam Arrays are the big thing these days and Disruptors and Phasers are the most popular energy types.
I intend to keep the charts with daily values for a while and then eventually switch to weekly averages once I have enough data.
Data in absolute values
Data as percentages