Tuesday, March 13, 2018

Vintage SurfWatch: Griefer Testing Findings Originally Posted December 13, 2009

SurfWatch sponsored a series of tests to observe the effect of some activities on competitive Second Life (SL) surfing. The testing was conducted on Wednesday, December 9, 2009, at 6:30 AM SLT. The results of this testing could also apply to any competitive SL sports, i.e., simboarding, sailboat racing, racecar driving, etc. and will be made available to any other SL parties that have expressed interest.


BACKGROUND

An SL Surfing Association (SLSA) expert was scheduled to conduct the testing, however, due to last minute Real Life (RL), the expert was unable to attend and members of the panel were able to recreate the test elements and conduct the tests. A diverse panel was invited to observe, ask questions and participate. The testing was unable to be videotaped during this session, however, the documented results of this testing will be used to educate surfers and officials on effects of these tested activities during a competition; provide a written record of observable results; and develop a baseline for future and/or further griefer testing.

This testing was undertaken by SurfWatch with the assistance of the SLSA as they are the governing body of SLSA competitions and training. The intent is that it be used as a training tool for SLSA officials and surfers in general to have a better understanding of what can happen when a surfer during a competition claims to have been interfered with; to help surfers identify when they have been interfered with as opposed to sim lag; and to help identify in general ways various SL competitions - not just surfing - can be effected for identification and, if possible, prevention.

This was a SurfWatch event and panelists were invited to participate by SurfWatch - the SLSA had no input on who would be sitting on the panel and, in fact, did not know who was invited on the panel until the actual testing. A wide range of the surfing community were invited to sit in on the panel and everyone is invited to review and comment on the results of the testing. Specific comments by panelists are identified below by initials.

SurfWatch did have testing critera in place which was not distributed in advance. We are sharing this information as well as asking for input on our findings and would be very interested in seeing any additional tests, especially by others that have technical expertise and knowledge. Prior to this testing, we developed a test outline. We tested on a homestead sim because that is where the majority of competitions take place, and we tried to duplicate conditions that officials and surfers would face when trying to identify these issues. Our findings and the observations and recommendations of the panel follow. We invite public comment as well, and would be willing to host future test sessions if there is interest.

TESTING

Time: 6:30 AM SLT
Date: Wed Dec 9, 2009
Location: SLSA sim Archipelago

Moderator: SurfWatch Tauri Tigerpaw (TT)
Surfers: Abel Halderman (AH), Desirae Beaumont (DB)
Panel: Lissa Pinion (LP), Marianna Monentes (MM), Kantbe Thursday (KT), Socks Clawtooth (SC), Buffy Munro (BM), List of Invited panelists who did not attend available on request

Tests:
1. Demonstrate holding a no-mod spherical object in observable edit mode while tapping or clicking on boards and/or waves (masking):

SW: We were able to hold a visible edit stream on a no-mod spherical object within the sim while clicking on waves and surfboards sitting in line up or surfing along a wave and affecting the wave and the rider almost undetectably.
MM: I witnessed the edit stream locked while the wave was being clicked. I saw the wave jerk and saw the surfer go down.
LP: While locked on a rider, it would be difficult for a judge to see the stream due to several factors including color of stream and wave colors. It was observed as well that while an individual was in edit on one object, they were also able to affect the rider by clicking on them or the wave w/ no detectible edit stream.
KT: Agreed

2. Demonstrate the effect on sim performance while holding any object on the sim in edit mode, and measure the impact on sim performance:

SW: Using Advanced, Consoles, Fast Timers - specifically, the network indicator (which we were advised is a measure of how much time the sim needs per frame to update adjacent sims and avatars that are being messaged from the sim with data information - if the network time goes up, so does time dilation and lag), we were able to observe the measurable impact and effect (by the strictest definition of "interference") on sim performance of holding various items in edit mode.
MM: I know I had some questions as to whether or not when in edit mode sim performance was affected. I saw that, in fact, when in edit mode while locked does change the sim performance.
LP: Agree with both of above posts and add that a Second Technical Security position would be needed during competitions due to having all the tools open would hamper a judge from properly observing the rider.
KT: Agreed

3. Demonstrate "script-loading" and its effect on other surfers in the sim.

SW: Taking two surfers (that could be competing in the same heat, for example), we had one load up with all the scripts they could find while monitoring sim active scripts using View, Statistics Bar, Active Scripts. The unloaded surfer took a wave and indicated he experienced what felt like "alot of spikes". We then had that script-loaded surfer remove all scripts and confirmed the Active Scripts dropped by 600. The original rider took another wave and could clearly tell the difference in quality and performance of ride, indicating it "felt like only one spike".
MM: I witnessed the scripts change when loading and unloading scripts when looking at Statistics in the Active Scripts.
LP: Observed same thing as well as on Sim Stats tool, the FPS spiked when scripts were added or removed quickly.
DB: In addition, high prim or scripted objects need to be added or removed before you start, as this action actually causes lag spikes, otherwise the sim has to update its object pointer references and update every avatar in the sim with that information.
KT: Agreed. I observed that the sim's Net Time figures jumped up in a similar fashion to when an av TPs into the sim.
SC: This test didn't take into account avatars sitting on an adjacent spectator sim, nor any scripts or effects going on in that sim that can also change from moment to moment and change how a rider perceives the wave. I would recommend rendering numbers be no higher than 2000 ARC. This still allows for 95% of all avatars - furry, neko or otherwise - to come without major alteration to their avatar.

4. Demonstrate right and left clicking on boards and waves during runs - measurable impact

SW: The panel tested various combinations of clicking or tapping on boards and waves to observe the impact on both, as well as testing the distance of an avatar still having the ability to edit click waves and boards. Our tests showed at a distance of two sims from the competition site using Advanced Menu, Disable Cam Constraints, an avatar can edit click waves and boards undetected. Within one sim, if the observer has Advanced, Character, Show Look At activated, they can see an identifiable cam lock on the rider and the wave. The wave demonstrated stop and start when in edit click. The board stopped, resulting in the rider sitting. Other observable actions resulted from various speeds and combinations of edit clicks on both.
MM: I saw that the waves and rider can be affected when locked on from 2 sims away.
KT: Agreed
SC: My concern with using "Show Look At" for surf events is that the stands can hold in excess of 35 avatars and sometimes up to 50. Isolating one out of them may be impossible. Also, some people will accidentally right-click a wave without intent of interference.

5. Demonstrate a decoy particle stream while editing items with invisible particle stream.

SW: While the panel was not able to demonstrate this last item, an SL scripting expert unrelated to surfing advised us of this possibility and stated it was fairly easy to create a decoy particle stream. A setting in Debug Settings allows the actual in-use particle stream to be switched off for everyone so it is unobserved.
LP: I have personally seen a builder use no visible stream before this test was done today.

RECOMMENDATIONS

Recommendations from the panel:

1. Communicate and reiterate again to SLSA members that being in edit lock at any time during an active competition within the competition sim is in itself grounds for Disciplinary Action due to the ability to "mask".
2. Create a competition official position of Technical Security (TS) assigned to remain within the competition sim during the competition and to monitor sim stats, including active scripts. If there is a surfer complaint or a major spike observed, the TS contacts the Head Judge to make a determination. The TS also consults with the HJ for sim performance issues in a re-do determination.
3. Remind all competitors to disable and remove all scripts, including AO, HUD, TE, etc. well before the event, and the TS log numbers before and after each heat. TS may possibly remain in the rez area and record script levels as part of the pre line up activities.
4. Request all competitors and competition officials to lower rendering numbers to a predetermined reasonable level and the TS verify before each heat.

While these recommendations in no way protect against or prevent all aspects of competition interference, accidental or deliberate, they serve as a starting point to address some major issues and generate further discussion and investigation.

4 comments:

November Sales said...

Excellent Work by all involved!

Bobbi Laval said...

Thanks so much Surfwatch for doing this. I'm curious, did people locking their cammera on surfers have any negative impacts on performance?

Tauri Tigerpaw said...

Thanks Bobs! We did not include camming in our initial test, however, we are scheduling another testing event and will include camming as well as broadcasting impact and other additional tests. The information regarding this is scheduled to post tomorrow morning.

Kris said...
This comment has been removed by a blog administrator.