OpenLedger trade engine bug or ..?

in bitshares •  9 years ago 

Sometime ago I read a post where a user of OpenLedger complained about the trade engine not matching the trades properly. In specific, higher "Ask" trades were executed before his lower ones, if I remember well.
I did check out his claims and they are true. Though, I would like to point out that it might be a bug, or that someone is trying to inflate the price of an asset by selectively trading between or among different accounts.

Authors get paid when people like you upvote their post.
If you enjoyed what you read here, create your account today and start earning FREE STEEM!
Sort Order:  

That is a known issue that doesn't matter to 99.999% of the trades only affects so called "dust trades" .. so those are low volume anyways ...
https://github.com/cryptonomex/graphene/issues/460

Excuse me. If you want to be taken seriously, as you intend to and advertise, you know that the above answer is not professional at all. Don't expect OpenLedger to gain support before you fix this bug, plus some others that you most probably have spotted already.

I am not affiliated with Openledger and and as such don't intend anything, nor do I advertise anything but merely point you to the facts.

The fact remains if that answer was one coming from an OpenLedger affiliated individual, it would still be an amateurish one.
In any case thank you for the clarification.

See https://bitsharestalk.org/index.php/topic,22851.msg296530.html#msg296530 for an explanation.

Basically, the skewed prices result from rounding errors on orders with very low volume. This is to be expected and not a bug.

The resulting trade price is "opticallly" skewed in the sense that the price of a specific trade is off by a large factor. However, price comparisons should always be done on volume-weighted numbers, in which case these rounding errors will disappear.

I checked the link. The case I refer has nothing to do with this. I mentioned that I had the Lowest Ask trade set and instead the higher one was executed. I distinguished a couple events that reason with the facts.
If you still insist that has something to do with rounding, then you should take rounding off the equation or find a way to execute trades properly if you decide to keep it.
Do not expect widespread adoption if such behavior is continued, explained, expected or not.

It would be helpful if you were more specific about your experiment. Any attempt to answer an unspecific question is only a shot in the dark.

In the example I cited, the actual trade price as calculated from paid/received amounts was off from the price of the offers on the market. The point here is that the paid/received amounts are calculated by applying the price of the market offer to the amount available for paying, and rounding the result. The rounding is necessary due to lack of precision in the target asset.

If that is not the problem that you verified in your experiment, again, please be more specific about what you perceive as the problem.

My experience is the following. Names and prices changed, essence remains untouched.
I set a trade of Asset A at an "Ask" price of 1. Gradually it becomes the "Lowest Ask". At that moment the "Market History" section reports trades executed at "Ask" price of 1.5, which is not the "Lowest". So, it is easy to conclude that there is either a bug there, or someone was just inflating the price.

Now since you mentioned a rounding issue as known, never mind if irrelevant to my experience, I just underlined the fact that traders do not like trade engines which produce unexpected results.
That's all and thank you for your time.

Thanks. Without knowing the details (volume, assets involved) of the trade that happened at 1.5 it is impossible to tell if the rounding issue is the cause.

What I take from this is that we're talking about a UI issue here. Your point that this is unexpected and leads to confusion is absolutely valid. I do not see this as a bug in the market engine, though, but merely as a display problem.

Edit: added comment to github issue https://github.com/cryptonomex/graphene/issues/460#issuecomment-236355266