Topic: [ANN][MOTO] Motocoin - page 71. (Read 178259 times)

I think it's not true, this is the type of game in which every minor difference in speed or position can lead to huge and unpredictable trajectory changes, so bot will have to use real physics and not "fake" one.

You still don't understand what I'm talking about. If you just attached some transactions to your block without securing them then no one cares about them. Realying nodes may change them while relaying and next miner can just ignore them and add there any transactions he wants.

Centralized banking system is also difficult to attack by anyone "not them". But the whole point of cryptocurrencies is that no one should have any large control.

Yes, I didn't thought about it, this makes it almost useless as protection from bots.

Maybe some one-way function can be used. For example client can be required to search for some nice hash each tick or second of playing so it will slow down playing, but checking solution will be very fast.

You would still attach the PoW to a new block with new transactions, your map seed would just reference older data.  Yes, when you solved this map it would do nothing to confirm the integrity of those transactions at that time (instead adding confirmation only to the blocks up to where your seed block was) but in the next N blocks when someone else solved against your block (or any block after it) it would serve to secure those transactions.  This is why I say it would require N times as many confirms to be sure of a transfer.

(EDIT to clarify: Really I think there are a lot of problems and potential problems with this approach anyway, particularly related to the extra confirmation delay and "S.E. attacks," so it should probably be considered as a nearly-last resort approach.  I think something based on proof-of-activity(/stake) blocks is likely a much more viable solution.)


Right, just adding computational effort to the path is not a good solution.


OUCH if this is just PoW checking time something is very wrong, then.  You should get significantly better performance than this.


Yup.  Not to mention the "clincher" here which is that the bots could simply run their solutions without this "added legwork" and then re-run their path against the "full physics" once a solution is found to verify that it will be accepted by the network.  This would just hamper normal users a LOT and the bot heuristics very very little.

If you are mining new block and secure some transactions with it then we say that these transactions belong to that block. If you say that they were part of previous block but wasn't secured by it then this is just a game of words, because you are that guy who secures them and you deside which transactions will belong to previous block, not the one who mined previous block, so it is still the same as if you added these transactions to your block. Each block includes hash of previous block, therefore currently all previous blocks are used for seed.


I'm not quite sure what you mean here. If they mine only coinbase then this is just a method to distribute coins, not to secure the network, right?


You don't need to change time, you can just change your address for coinbase transaction. With some other tricks you can have even more insanely larger set of possible maps.


Yeah, it is exceptionally secure. And you can perform 99% attack any time you wish, sounds very secure.


Remember that physics is calculated not only while playing but also when other nodes check blocks for proper proof-of-play. Right now I can check about 60 blocks per second in one thread on my Core i7-3770. With 4 threads I will probably be able to check 240 blocks per second but many people have much slower processors. With your proposal physics simulation will become much slower and to really harm the bots we will need to make it at least 100 or 1000x slower. This will make block checking very slow, on some computers it will take several seconds to check one block, some probably would check new blocks slower then they are generated, inital synchronization with network will take days or even months, botnets will perform spam attacks sending a lot of invalid blocks and nodes will use 100% of their CPU time to check these blocks.

My only complaints about the interface are
A) Night mode: just plain doesn't work right.
B) Updating:  MOTO Charts freeze constantly, and it is often not clear if the buy/sell list you are looking at is currently up to date or not.

Thank You for feedback. What else in interface could be better in Your opinion?

DOMOB you are awesome.


Yah!  I'd love to see something that is explicitly for such a purpose...  evolutionary computation based PoW maybe? Proof-of-annealing?  (Would be nice to have a proper "quantum hard" chain, too, as currently I don't think there are any? (Yah yah, I know, quantum just breaks security of addrs and the merkle chaining anyway...))


This is no different whether the "hashing function" is sha, scrypt, blake, x11, moto racing, raytracing, sat solving, or anything.  The first sha asics could have 51%ed bitcoin.  If I go make an X11 fpga array tomorrow I could 51% attack a lot of networks for awhile.  If someone made a raytracing based coin and someone else devised some new crazy fast raytracer math they'd be able to 51%.  If there were a 3SatCoin and I found some crazy new lower bound on general NP solution... well you get the drift.

You just have to hope that the people who make the technological breakthroughs are inclined toward using them to the benefit of securing the network (instead of to the detriment of the network's users) and that competition quickly causes others to procure and adopt the same new technologies, so the network remains secure.  As I said before, I personally intend benevolence.  Any of us botters could have already actually killed motocoin any time now, by just 51%ing c-cex just enough to where the exchange decides to remove the listing because of the losses.  As each new botter comes online (btw, hello botter number 4-or-5(-we're-not-sure-yet), welcome to the game!) such an attack becomes more difficult/unlikely as you have ever increasing hashing strength from the other botters to compete against.


I think it is fine to use such techniques as long as they are actually computationally hard (an open question for moto... I'm so far unconvinced that there isn't a direct linear regression solution allowing deterministic constant time block solution at any difficulty) and are allowed to continue to actually *secure* the network!  (Meaning the bots are allowed to do their thing and strengthen the consensus history with their ever increasing ramp of energy spend commitments.)

The hard part ofc is doing it in a way that it doesn't just degrade into nothing more than a really convoluted sort of alternative hashing function, so you don't end up with "just another altcoin" that might as well be just another x11 ltc fork.  That, as I see it, is the open problem.... how to achieve machine-scale driven network security around consensus and still leave room for human competition for coinbases.  I suspect it is possible, but I'm afraid that I don't have all of the answers just yet.

Yes, as I said there are quite a few fatal problems with my original plan.  I have two rough directions that I'm thinking now... either allowing the last N blocks to be used for seed (so when you win a new block you might not necessarily actually be securing the current block, but adding security to any of the last N blocks.  This would, however, mean that people would need to wait N times as many confirmations to be confident in their coins... I still think this is fraught with problems, however, and I probably won't continue exploring it much further.

Another path that I am looking down is using something along the lines of the proof-of-activity proposal to allow user to collect up proof of stake data, and mine against that data to create blocks.  Stake data would have to carry some "freshness" metric to avoid replay withholding, and manual human miners might only be able to mine coinbase TX in their blocks (unless they accept a map reset) but otherwise I'm thinking that it might work.


Ok, I was assuming people were "playing fair" on their timestamp, as well.  (My bots don't do any timestamp manipulation, they just take the "***Work" block that the client spits out.)  Even with the timestamp the block count isn't quite infinite (there's an limit on timestamp drift allowed by the network ofc) but this does make the search space so large that, pragmatically, it may as well be infinite.


No, not at all!  What I'd like to get to is a state where bots can mine without affecting human miners, and human miners can still solve blocks to win coins, and the two still compete not directly "block for block" as currently but instead compete only on TargetTime thresholds.  Under this scenario the network is maximally secured (by the large computational effort of the bots) and still human mineable.

At launch we had a very insecure network that was great for human miners of all skill levels.
Right now we have an exceptionally secure network that very very few humans are able to mine on.
The goal is to keep the network as computationally strong as it is now, but with a possibility for (ordinary, sane) human miners to continue to compete for coinbases. (EDIT to clarify: I mean secure the network without just punting and adding hashing based PoW like HUC does.)

Deposits were quick and easy, overall it was good. The interface however needs a lot of improving, for example: the value graph needs to be a bit more streamlined in my opinion

I'v got only one answer. Maybe anyone else could give several words for our platform?

And what about adding some computationally intensive randomness to physics (like computing many many hashes each tick and adding it to velocity). This will make bot's speed comparable to normal human playing speed. And without careful planning bots will definitely lose.
Along with 4x sized map it will knock out bots for at least a couple of months.

At least be more prudent about the whole situation. There are members of the community who have invested in this project fully believing in it, and you, as the person who has spent most time on the project, should value your own efforts more highly. The sell orders on c-cex are extremely thin, 2 BTC would get us back to where we were before. As long as another exchange does not add MOTO and in the meantime the necessary fixes are implemented, everything should be back to normal. It's still a very young project - if everyone were to give up in the first few days whenever a new coin is created, the crypto-world would have long since died

People must know about current situation. Do you want lie to everybody that it is currently human-mineable?

What's the reason for adding "now dominated by bots" in the ANN title? It's a childish move, scares away investors and shows the Dev's lack of commitment to the project if he gives up easily at the first noticeable problem

You contradict to yourself, you say that it is good the way it is but propose to make fundamental changes. If map is not depended on transactions in block (to limit generation of new maps) then we need to mine blocks in some other way.

You can just change transaction in block to generate new map and always use nonce=0.

Yeah, I proposed that a long time ago, I still like the idea...

Just make reward 100-nonce*10 (in other words decrease reward each time player presses F6)

My be add several coins? Each player must collect of this coins. And difficalty will be depend on not only time, but number of coins.

Its good the way it is. I suspect most of these botting programs are going through hundreds if not thousands of maps looking for the optimal one and then brute forcing it, they might have even modified the game so that it goes really really fast, like faster than you can adjust the game normally.

By charging 1moto for a set of maps, or charging per play, like how an arcade game used to charge 25cents per play, you could make it very hard for bots. For 1moto they get 16 presses of f6 and/or a turn. This means the bots have to pay for all the maps they blow through in seconds. If the idea was refined I think it could make bots unprofitable.