I looked into it more, and am pretty convinced that he cheated the 2:31.554 run.
For background, I will describe the Live Replay TAS code. A variant of this code is what people use to record runs, it basically plays back inputs from the ghost and drives it in real time. The TAS version of this code allows you to activate/deactivate it at certain times, allowing you take over the ghost and correct mistakes.
By doing this multiple times you can use this code to cheat and basically have save states. Example:
-Set a good lap 1, but fail the run. Save the ghost by playing on a license that doesn't have your best time.
-Use the live replay code on that run. Take over the driving after lap 1, get a good lap 2, and save the new ghost.
-Use the live replay code again until you get a good lap 3 on that run and save the completed run. You could do these attempts on your main license so the ghost saves to your real top 5.
This method is undetectable as long as we never see the ghosts that were used to build up to the completed run. The most publicly available version of the Live Replay TAS Code was for GCN controller only.
Now, I will show why it appears that ruso cheated with Live Replay TAS:
I checked the inputs for both the 2:31.554 and 2:31.710 runs AFTER they diverge (2:08.5 in game time in the comparison posted:
https://www.youtube.com/watch?v=zpbo7mNgyaQ). The inputs AFTER this point, for both ghosts, suggests that a GCN controller was used. All inputs BEFORE that point appear to be from WiiChuck.
To see this for yourself, download the Input Viewer (
https://www.mariokartboards.com/threads/ghost-input-display-introduction-and-...) and watch the ghosts. Up until 2:08.5 (128.5 seconds), the ghosts clearly use WiiChuck since you can see when he wheelies it triggers other inputs such as down trick. Then, after 2:08.5 the wheelie inputs seem more like GCN controller inputs, since it only shows D-pad UP inputs.
Therefore, I think he used the Live Replay TAS code to complete both runs, taking over during the lap 3 shroom. This is why both ghosts say they were done with GCN, even though he is a WiiChuck main.
This is also why Star's check shows "illegitimate" inputs for these ghosts. Since the first 2.5 laps were originally driven on WiiChuck, and WiiChuck has controller stick inputs that are impossible on GCN (video explaining this:
https://www.youtube.com/watch?v=KUjS7qWWu9c).
A WiiChuck player having GCN live replay copies of their ghosts is not alone evidence of cheating, since it's possible the GCN ghost was made when they or someone else recorded the ghost (though most live replay recorders used a version that doesn't save the ghost). However, if this was the case the ENTIRE run should still be WiiChuck inputs.
But in ruso's case, he has two ghosts with the exact same inputs up until lap 3, and both of them have GCN inputs after this point. This strongly suggests he used the live replay TAS code to correct a failed run, then tried to pass the cheated run off as a WR.
A potential timeline of events:
2013-06-??: ruso drove a failed run where he got 50.334 (new BKS) + 50.626 then failed on or after the last shroom. This fail was likely played legitimately with WiiChuck and the ghost was saved, however we do not have this ghost.
2013-06-13: ruso decides to cheat a WR, by using the live replay TAS code on this failed run. This means he had to do attempts of getting a good ending on the failed run using the GCN controller. In one of those attempts he gets the 2:31.710 and the ghost is saved.
2013-06-13: On a different license, he does more attempts at getting a good ending and gets 2:31.554. He claims that as WR and uploads a live replay video. Note that this run wasn't counted on the WR page since he used custom textures.
2013-06-??: The 554 run is uploaded to WFC, and is automatically added to the ghost database.
2013-08-01: ruso sets a 2:31.402 (this run might be legit). In order to add the ghost to the database, he uploads his save file to ninrankings. This also adds his 2:31.710 ghost, which must have still been saved on a different license. The 710 ghost is the main evidence that he cheated, and has the same date as the 554 (2013-06-13).
2:31.402 ghost ID = 29385
http://www.ninrankings.org/ghostdatabase/ghostviewer.php?id=293852:31.710 Ghost ID = 29414
http://www.ninrankings.org/ghostdatabase/ghostviewer.php?id=29414Other ghosts around these IDs are also from ruso (e.g. 29413 and 29415). This suggests they were all added at the same time from his save file (2013-08-01).
To me all the evidence points to this run being cheated. For it to be legit the events would have to be something like this:
-ruso sets 2:31.554 legitimately with WiiChuck. For some reason, during the end of lap 3 his wheelie inputs don't trigger additional inputs like down trick.
-ruso uses the live replay code to record the 2:31.554 run, so a GCN version of the ghost is uploaded to the database.
-For some reason, ruso uses the live replay code on the 2:31.554 and saves a ghost which has a different ending (2:31.710). Maybe as a test or something. The 710 would have to be AFTER the 554 otherwise the 554 is cheated.
-The wheelie inputs at the end of the 710 run (which must be done with GCN since he used live replay), for some reason, look similar to the wheelie inputs at the end of the 554 run (which must be done with WiiChuck to be legit).
Basically, he needs a good explanation as to why he has 2 ghosts with different endings and why both endings seem to have GCN inputs (even though he is a WiiChuck main). I don't see any explanation that makes sense other than he cheated.
The times set after this could potentially be legit as he did stream quite a bit of rBC. We need to decide which times to remove from the WR site now, I am wondering what the community thinks should be kept/removed.