Good job compiling these “coincidences” @Neil_AU. That amount of common spikes are incredibly unlikely. Would be interesting to hear a good explanation… Given the evidence and lack of response, we support of this chill.
5 Likes
FTSOLondon supports this proposal based on past and present evidence for both providers.
2 Likes
Unless FLR Labs gives us a good answer promptly, we support the proposal also.
2 Likes
Same as Andrew, without any correct answer from FLR LABS I support this proposal.
According to a post on their Discord the only one willing to remain anonymous seems to be the dev which looks even more suspicious by the way.
3 Likes
we support the proposal, and the fact that the developer is the only member of the team who wishes to remain anonymous leaves no doubt : they pay this developer to have the prices with an API, and they have no control from what they submit, I think that they have no idea of the sources used to calculate the prices.
in the current context, guilty until proven guilty
4 Likes
I have reviewed the evidence and I am in favour of this.
2 Likes
1 Like
Thanks for the reply! We have a hard time understanding your comparison due to the amount of common deviations and the amplitude of them. There are a difference in magnitudes, check the number of zeroes, between the examples you make and the examples from @Neil_AU. Even if the examples you make were examples of collusion, it’s not the parties we are discussing in this thread.
2 Likes
Hello and thanks for taking the time to reply.
I will say that we did not decide yet like others on where we lean before you posting your reply.
I want to say that we partly agree with you that some price spikes that are below 1% are insignificant enough that do not show really anything. But like you say out of tens of thousand of epochs that you hit the median within 0.1% you have an outlier of 1% difference a few hundred times and it happens that someone else out of tens of thousands of epochs has an outlier at the same exact epoch as you and then it happens multiple times.
Any repeated spikes above 1% take a crazy percentage of chance to be repeated. Just because there were 14 examples that neil showed it doesnt mean its only those, there are more spikes with differences larger than 1% if you want i can gather and show some of them.
Your claim that its from an exchange feed seems strange as everyone are using the same feeds sockets or rest so it would affect the rest of us and it usually takes longer to fix.
A lag between your server or a desync of the book would explain it if it was more epochs that had the problem.
For discussion purposes i just want to correct some false information you gave so it doesnt muddle the conversation.
LoveFTSO only got his provider started in December 2022 in fact the first transfer to that address is on December 17th so it doesnt have anything to do with the September happenings or ApeFTSO.
Looking forward to more discussion.
5 Likes
This site doesn’t allow to upload pictures so I hosted it here:
These are crazy spikes for ADA. ADA ranging from 9 to 40$ is not a 1% spike.
Ironically only the “good fellows” have them, let me name them:
LoveTSO
FLRLabs
O1
OLD_TOADZ (that we know it was hiro running it)
Coffee_TSO
There are 2 more that I can’t find info about it.
4 Likes
Let’s simplify our kernel algorithm equation.
For example, we can calculate the median price using the formula: price = calc_median_price(binance_eth_price, bitrue_eth_price). However, sometimes exchange services are not available due to maintenance, which can cause issues with our function.
For instance, calc_median_price works well if provided values are in range [1500~2000] but if bitrue_eth_price becomes 0, our result may become a strange value that could cause a spikes in our submissions which can become problematic if LoveFTSO were using the same exchange as us.
I find your response to be severely lacking and erroneous. Is this just to make a show for your supporters that you are a unfortunate target by jealous bullies? Anyone with reasonable logical skills sees you have not at all defended your position but in a sad way are attacking others and not answering allegations.
As previously mentioned by bifrost_oracle, the order of magnitude of your counter examples is completely different than the ones provided by Neil_AU. The charts even look completely different. You didnt even put in a good effort to find better suiting charts, which Im sure could be found due to the large amount of data processed by the FTSO. The difference is you would find maybe one or two such instances (still with different orders of magnitude) whereas in the FLRLabs-LoveFTSO case we see multiple with huge differences. Its hard to say you are pioneers in this space if this simple concept truly eludes you.
As for the collusion tool - what does your response even mean? What exactly is 5 = 5000 x 0.999? Each value of the collusion tool is below 1, so you can make any such mathematical equation. 69420 = 69420 x 0.999. You are trying to explain the collusion tool and you don’t understand how it works. It does exactly what you are saying it doesn’t, it shows similarities between submissions and that is 99.9. Of course this doesn’t directly mean you are colluding, but it makes it easier to identify others to put under the microscope. Which was done and the most important results, the charts with orders of magnitude difference, were proposed.
Of course Flare will have the final decision. We are just putting forward proposals to them and providing evidence. But if you think saying that you are a target and are legit without providing actual proof will be enough to save you I think you will be surprised. You keep saying you want discussion since this is the easiest way to show you are being bullied by the envious group. Most users won’t read this long thread and your lacking response, but it is easy for you to sway your following into thinking we just want to take your rewards.
But the evidence speaks for itself, whereas your response does not speak in your defence. I look forward to you bringing more to this discussion than some quickly found similarities between your competitors (who dont need your rewards anyways) and then some weird math which doesnt even make sense.
5 Likes
I think what you post is wrong for a simple reason.
If your price was missing and you get a 0 it will basically remove that coefficient from the submission calculation.
This will result in a lower submitted price, it cannot explain the spikes higher than the median.
For anyone following not accustomed. You basically have different weights of prices from different feeds.
Example of what you claim.
Binance has a 10% weight, Kraken has a 20% weight and rest exchanges another 70%. The total of all feeds have a 100% result. If lets say binance has a 0 price the weight will be 0 so you will basically post a price 10% lower than intended.
7 Likes
Why would it be a problem that LoveFTSO uses the same exchange as you if their algorithm is different? Are you implying that every provider using the same exchange as you that caused the spike should get the same spike on their submission?
3 Likes
Sure, it you handle that outlier identically… and sample at the same time… and use the same exchanges. I feel that this is just a good argument for the purpose of this thread and the chill we are discussing.
3 Likes
From the Collusion tool Doc …
“The edges are then sorted according to their weights. Then a percentage threshold is defined, usually around 0.98 (meaning 98%) and this percentage of the smallest closeness value edges are omitted (e.g. 98% of lower value edges are omitted, 2% of edges with biggest closeness values are kept). Then such a graph is drawn using a spring embedding algorithm which actually produces clustering.”
3 Likes
Thank for your the reply, but thinking about your answer we do not agree and we think you are using same prices provided by an external source or the same algorithm.
The reason is basic math: you cannot use same sources at same times and submit your signal in the same way without colluding.
Best wishes.
2 Likes
Is your tso dev the only member or your team that’s annon?
2 Likes