[ANN] Catcoin - Scrypt meow!

[Maverickthenoob]

Well one of the big multipools agreed to leave CATcoin alone maybe we can ask the others? Like middlecoin and hashcows to leave us be for 3 days or something? Maybe they can find it in their hearts! 

Hashcows agreed to leave CAT alone for a few days, I don't think Middlecoin has even been contacted yet though.

Besides I doubt Middlecoin would leave us alone.

This is the fundamental problem that the fork didn't fix: until there is enough constant hash to make the effect minimal, multipools are going to cause huge difficulty swings by jumping in when the difficulty is low, mining out all the blocks, pushing the diff up, and jumping out. We only have to mine 36 blocks now to get through the high diff, but we don't get many of the easy ones for all that work.

Until the net hashrate is too big for a multipool to significantly effect (which means 20+ GH/s), the difficulty NEEDS to depend on more than just the last 36 blocks. The 36 block retarget is great, but the new difficulty should be calculated from the last 2016 block times so that fast 4x diff swings dont happen.

The only issue is that if the diff got pushed up and then the network hash fell a bunch, it would take a lot longer for the diff to drop back down again. Though it would take a lot more to push the diff up. We honestly just need more hashrate.

[1714714056]

1714714056

[1714714056]

1714714056

[1714714056]

1714714056

[1714714056]

1714714056

[campeck]

Yeah I had all my rigs pointed at Middlecoin recently and DOGE as a backup pool. Well Middlecoin kept interrupting connection so much that I mined 100k Doge and held the top hash spot on doge.luckyminers! So something has definitely been happening with the middlecoin algo. Maybe the fast coin switching was knocking me off and keeping me off since the cgminer --failover-only command seems to want 5 minutes of stability in on pool 0 before returning to it.

Point those rigs at CAT! 

They are, they are! 

[mavis369]

Maybe this phone http://www.dailymail.co.uk/sciencetech/article-2534593/The-invincible-phone-New-handset-waterproof-heatproof-drop-proof-cancels-noise-phone-calls.html can use catcoin in some way, not as daft as it sounds???

[envy2010]

Well one of the big multipools agreed to leave CATcoin alone maybe we can ask the others? Like middlecoin and hashcows to leave us be for 3 days or something? Maybe they can find it in their hearts! 

Hashcows agreed to leave CAT alone for a few days, I don't think Middlecoin has even been contacted yet though.

Besides I doubt Middlecoin would leave us alone.

This is the fundamental problem that the fork didn't fix: until there is enough constant hash to make the effect minimal, multipools are going to cause huge difficulty swings by jumping in when the difficulty is low, mining out all the blocks, pushing the diff up, and jumping out. We only have to mine 36 blocks now to get through the high diff, but we don't get many of the easy ones for all that work.

Until the net hashrate is too big for a multipool to significantly effect (which means 20+ GH/s), the difficulty NEEDS to depend on more than just the last 36 blocks. The 36 block retarget is great, but the new difficulty should be calculated from the last 2016 block times so that fast 4x diff swings dont happen.

The only issue is that if the diff got pushed up and then the network hash fell a bunch, it would take a lot longer for the diff to drop back down again. Though it would take a lot more to push the diff up.

The difference being that the pools' profitabilities would change every 36 blocks but NOT by enough to drive ALL the hash off the coin. Having the retarget use more than one period for the calculation introduces an integration factor that tends to reduce variability in difficulty. Its a simple system dynamics theory that's easily mathematically provable.

We honestly just need more hashrate.

Absolutely. The best way right now is to get more people interested.

[Maverickthenoob]

Well one of the big multipools agreed to leave CATcoin alone maybe we can ask the others? Like middlecoin and hashcows to leave us be for 3 days or something? Maybe they can find it in their hearts! 

Hashcows agreed to leave CAT alone for a few days, I don't think Middlecoin has even been contacted yet though.

Besides I doubt Middlecoin would leave us alone.

This is the fundamental problem that the fork didn't fix: until there is enough constant hash to make the effect minimal, multipools are going to cause huge difficulty swings by jumping in when the difficulty is low, mining out all the blocks, pushing the diff up, and jumping out. We only have to mine 36 blocks now to get through the high diff, but we don't get many of the easy ones for all that work.

Until the net hashrate is too big for a multipool to significantly effect (which means 20+ GH/s), the difficulty NEEDS to depend on more than just the last 36 blocks. The 36 block retarget is great, but the new difficulty should be calculated from the last 2016 block times so that fast 4x diff swings dont happen.

The only issue is that if the diff got pushed up and then the network hash fell a bunch, it would take a lot longer for the diff to drop back down again. Though it would take a lot more to push the diff up.

The difference being that the pools' profitabilities would change every 36 blocks but NOT by enough to drive ALL the hash off the coin. Having the retarget use more than one period for the calculation introduces an integration factor that tends to reduce variability in difficulty. Its a simple system dynamics theory that's easily mathematically provable.

We honestly just need more hashrate.

Absolutely. The best way right now is to get more people interested.

Envy, I actually agree with the previous 2016 block average, I'm just wondering if that in certain extreme situations might not make the diff "liquid" enough and could hurt the profitability of the coin, especially if we gained a whole bunch of hash for a long period of time. I can think of a couple of extreme scenarios in either direction where that could cause problems. I also am not sure that it's a significant enough change to make it worth another fork, at least right now. The major issue got fixed last time, this is something we can watch more long term IMO.

[etblvu1]

Well one of the big multipools agreed to leave CATcoin alone maybe we can ask the others? Like middlecoin and hashcows to leave us be for 3 days or something? Maybe they can find it in their hearts!  

Hashcows agreed to leave CAT alone for a few days, I don't think Middlecoin has even been contacted yet though.

Besides I doubt Middlecoin would leave us alone.

This is the fundamental problem that the fork didn't fix: until there is enough constant hash to make the effect minimal, multipools are going to cause huge difficulty swings by jumping in when the difficulty is low, mining out all the blocks, pushing the diff up, and jumping out. We only have to mine 36 blocks now to get through the high diff, but we don't get many of the easy ones for all that work.

Until the net hashrate is too big for a multipool to significantly effect (which means 20+ GH/s), the difficulty NEEDS to depend on more than just the last 36 blocks. The 36 block retarget is great, but the new difficulty should be calculated from the last 2016 block times so that fast 4x diff swings dont happen.

Under your proposed algorithm, what would happen, if the difficulty appears to be too high for the value of the coin over an extended period of time (i.e., the coin sits near the bottom of profitability lists), so it keeps reducing difficulty inch by inch by means of the 2016 block average, then one day, the difficulty got low enough, that it climbs up to the middle of the profitability list - then for unrelated reasons - maybe Catcoin gets some press - the price of Catcoins gets driven up to double or triple the price from increased demand - and now Catcoins are at the top of the profitability charts - massive hash power gets directed onto the coin? Does this mean we may have to go through 1,000 blocks or so of very rapid block generation (e.g., once every 10 seconds), until the 2016 block average starts to move the difficulty up? Would such a scenario not represent a rapid inlation and influx of 50,000 newly minted CATs being dumped into the exchanges? Also, once the difficulty rises to the point where it is no longer at the top of the profitability charts, and the 50,000 CATs being dumped in the exchange causes the prices to drop, and the massive hash power gets directed elsewhere, would we not be back to having very slow block times (e.g., every 30 minutes or an hour), until the averages can get caught up, perhaps in another 1,000 blocks?

If you see it playing out differently, please describe how you see it playing out.

[DJTrey]

CatcoinFlip.com has launched - coin toss betting game for catcoins.

Enjoy :3

[rentarocker]

Promote CAT. 

- The price dive after IPO was expected with 1,000,000 coins created in such a short period.  Now we're creating what...7200/day?  Pocket change.

- The fork was a perfect opportunity to spread FUD about CAT.  Many of readers of this thread fell for it.  Watch Cryptsy and its clear that anytime price dips big players swoop in and vacuum up CAT.  This is a good sign.  These actors aren't buying CAT for the hell of it.  Bitcointalk forums only fueled this fire.  Stop throwing your CAT on it.

- This coin is on target to become scarce and expensive.  I just bought more.  Hashrate problems will go away when the value of CAT increases.

- CAT team no longer needs to be distracted with the nuts-and-bolts of starting a crypto.  It's done.  Arguing semantics is a waste of time. 

A small amount of promotion and all of these problems will be gone by the end of the week.  An influx of new money to BTC from mainstream investors is what drove DOGE to a multi-million dollar market cap.  What are you doing to attract them?  I can guarantee you that most 'new' money to crypto will know and care nothing for hashrates.  They will care that CAT has a good name; that it can be used on online payment/gambling sites; that it has a strong community ect.

[Maverickthenoob]

Well one of the big multipools agreed to leave CATcoin alone maybe we can ask the others? Like middlecoin and hashcows to leave us be for 3 days or something? Maybe they can find it in their hearts!  

Hashcows agreed to leave CAT alone for a few days, I don't think Middlecoin has even been contacted yet though.

Besides I doubt Middlecoin would leave us alone.

This is the fundamental problem that the fork didn't fix: until there is enough constant hash to make the effect minimal, multipools are going to cause huge difficulty swings by jumping in when the difficulty is low, mining out all the blocks, pushing the diff up, and jumping out. We only have to mine 36 blocks now to get through the high diff, but we don't get many of the easy ones for all that work.

Until the net hashrate is too big for a multipool to significantly effect (which means 20+ GH/s), the difficulty NEEDS to depend on more than just the last 36 blocks. The 36 block retarget is great, but the new difficulty should be calculated from the last 2016 block times so that fast 4x diff swings dont happen.

Under your proposed algorithm, what would happen, if the difficulty appears to be too high for the value of the coin over an extended period of time (i.e., the coin sits near the bottom of profitability lists), so it keeps reducing difficulty inch by inch by means of the 2016 block average, then one day, the difficulty got low enough, that it climbs up to the middle of the profitability list - then for unrelated reasons - maybe Catcoin gets some press - the price of Catcoins gets driven up to double or triple the price from increased demand - and now Catcoins are at the top of the profitability charts - massive hash power gets directed onto the coin? Does this mean we may have to go through 1,000 blocks or so of very rapid block generation (e.g., once every 10 seconds), until the 2016 block average starts to move the difficulty up? Would such a scenario not represent a rapid inlation and influx of 50,000 newly minted CATs being dumped into the exchanges? Also, once the difficulty rises to the point where it is no longer at the top of the profitability charts, and the 50,000 CATs being dumped in the exchange causes the prices to drop, and the massive hash power gets directed elsewhere, would we not be back to having very slow block times (e.g., every 30 minutes or an hour), until the averages can get caught up, perhaps in another 1,000 blocks?

If you see it playing out differently, please describe how you see it playing out.

This actually is quite an articulate way of describing exactly the issues and "extreme situations" I was concerned about. I see it playing out exactly like that, yeah the diff would lower or raise every 36 blocks, but in tiny fragments rather than jumps, even if they are necessary jumps. This could in turn grossly affect the profitability of the coin, and could lead to huge surges and quiet periods of hash. Etblvu1 hit this one right on the head I think.

Unrelated: I've made contact with the Catcoinz Facebook page owner and he has agreed to hand the page over to me, at least for now, and I will in turn either do the content myself, hand it over to another member, or remove the page depending on what the community wishes. Unfortunately I don't see Catcoinz (with a z) being a page that is taken particularly seriously or professionally, so we may just have to turn it into an obvious unofficial page while the other one is made more official. I'm open to suggestions from the community on this.

[envy2010]

Envy, I actually agree with the previous 2016 block average, I'm just wondering if that in certain extreme situations might not make the diff "liquid" enough and could hurt the profitability of the coin, especially if we gained a whole bunch of hash for a long period of time. I can think of a couple of extreme scenarios in either direction where that could cause problems. I also am not sure that it's a significant enough change to make it worth another fork, at least right now. The major issue got fixed last time, this is something we can watch more long term IMO.

No, it's not worth forking again now. I'm just throwing it out there to keep in mind for the next wallet update, which is going to happen eventually.

As for difficulty liquidity, I don't think it's that big a problem. If the hashrate slowly builds up (with the diff following it), then suddenly dissapears for some reason, the resulting long block time will quickly drive the 2016 block average time way up, and consequently the difficulty way down (since it would still be retargeted every 36 blocks).

Say you spend 4 weeks at 10 net GH/s, getting 10-minute block times. Then 8 GH/s goes away and you get 50-minute block times (b/c you have 1/5 the hashpower) for 36 blocks (30 hours). With the current mechanics the diff would retarget to 1/5 it's previous rate, allowing the 8 GH/s to come right back in and get 2-minute block times for 36 blocks (1.2 hours). Instead, using a straight average over the previous 2016 you would have 10.7 minute blocks and the difficulty would drop by 7%. Now if the 8 GH/s comes back, you get 9.3 minute blocks (instead of 2-minute). If it doesn't, you get 45 minute blocks for another 28 hours until the next retarget, where it would drop about about another 7%. It would keep dropping by about 5-10% every day until hashpower picked up again, and under the circumstances I assumed above, it would hit half the difficulty in about a week. The advantage over a regular 1 or 2 week retarget is that the difficulty isn't stuck forever if hash drops, but can also respond quickly.

Say the network is steady at 1 GH/s and rises to 10 GH/s between retargets. The blocks come every minute for at most 1/2 an hour, then difficulty retargets to 4x, the 9 GH/s jumpers leave, and the 1 GH/s core is stuck with 40-minute block times (this is what we see now). Instead, the 2016 block average time would be 9.84 minutes and the first retarget would be only 1.2% higher. If the jumpers stick around another 36 blocks (36 minutes), the 2016 average goes to 9.68 minutes, and the diff jumps another 1.5%. After 360 blocks (6 hours, the planned retarget TIME), the difficulty would have gone up 10 times to 17% higher than it started - which might cut into profitability. After 640 blocks (~12 hours), it would have gone up about 18 times to a 40% higher difficulty. This lets profitability (which is heavily based on the coin's value) truly determine what the difficulty should be.

In short, using a longer average would be like setting a time limit on the maximum amount the difficulty can change. If it very recently changed it would not be able to jump or drop by 4x. The cumulative effect of many smaller changes would let the difficulty sync with the coin's value.

[overseerio]

CatcoinFlip.com has launched - coin toss betting game for catcoins.

Enjoy :3

Awesome

[dotnetmin]

Pump and Dump coins are dying. LTC is place at #3 of most profitable coins at multipool.us
So miners can stop mine scrap coins and change to LTC and BTC

Free markets regulate itself   

[TheTribesman]

Well one of the big multipools agreed to leave CATcoin alone maybe we can ask the others? Like middlecoin and hashcows to leave us be for 3 days or something? Maybe they can find it in their hearts! 

Hashcows agreed to leave CAT alone for a few days, I don't think Middlecoin has even been contacted yet though.

Besides I doubt Middlecoin would leave us alone.

This is the fundamental problem that the fork didn't fix: until there is enough constant hash to make the effect minimal, multipools are going to cause huge difficulty swings by jumping in when the difficulty is low, mining out all the blocks, pushing the diff up, and jumping out. We only have to mine 36 blocks now to get through the high diff, but we don't get many of the easy ones for all that work.

Until the net hashrate is too big for a multipool to significantly effect (which means 20+ GH/s), the difficulty NEEDS to depend on more than just the last 36 blocks. The 36 block retarget is great, but the new difficulty should be calculated from the last 2016 block times so that fast 4x diff swings dont happen.

Under your proposed algorithm, what would happen, if the difficulty appears to be too high for the value of the coin over an extended period of time (i.e., the coin sits near the bottom of profitability lists), so it keeps reducing difficulty inch by inch by means of the 2016 block average, then one day, the difficulty got low enough, that it climbs up to the middle of the profitability list - then for unrelated reasons - maybe Catcoin gets some press - the price of Catcoins gets driven up to double or triple the price from increased demand - and now Catcoins are at the top of the profitability charts - massive hash power gets directed onto the coin? Does this mean we may have to go through 1,000 blocks or so of very rapid block generation (e.g., once every 10 seconds), until the 2016 block average starts to move the difficulty up? Would such a scenario not represent a rapid inlation and influx of 50,000 newly minted CATs being dumped into the exchanges? Also, once the difficulty rises to the point where it is no longer at the top of the profitability charts, and the 50,000 CATs being dumped in the exchange causes the prices to drop, and the massive hash power gets directed elsewhere, would we not be back to having very slow block times (e.g., every 30 minutes or an hour), until the averages can get caught up, perhaps in another 1,000 blocks?

If you see it playing out differently, please describe how you see it playing out.

This actually is quite an articulate way of describing exactly the issues and "extreme situations" I was concerned about. I see it playing out exactly like that, yeah the diff would lower or raise every 36 blocks, but in tiny fragments rather than jumps, even if they are necessary jumps. This could in turn grossly affect the profitability of the coin, and could lead to huge surges and quiet periods of hash. Etblvu1 hit this one right on the head I think.

Unrelated: I've made contact with the Catcoinz Facebook page owner and he has agreed to hand the page over to me, at least for now, and I will in turn either do the content myself, hand it over to another member, or remove the page depending on what the community wishes. Unfortunately I don't see Catcoinz (with a z) being a page that is taken particularly seriously or professionally, so we may just have to turn it into an obvious unofficial page while the other one is made more official. I'm open to suggestions from the community on this.

I was even thinking that it might be best for new altcoins to change difficulty block by block. That way the multipools will have a real job of trying to pile in on easy block periods. I can imagine it's not workable mathematically but it's a thought. Keep everyone guessing. Miners committed to that particular altcoin won't care about difficulty fluctuations.

[envy2010]

Well one of the big multipools agreed to leave CATcoin alone maybe we can ask the others? Like middlecoin and hashcows to leave us be for 3 days or something? Maybe they can find it in their hearts!  

Hashcows agreed to leave CAT alone for a few days, I don't think Middlecoin has even been contacted yet though.

Besides I doubt Middlecoin would leave us alone.

This is the fundamental problem that the fork didn't fix: until there is enough constant hash to make the effect minimal, multipools are going to cause huge difficulty swings by jumping in when the difficulty is low, mining out all the blocks, pushing the diff up, and jumping out. We only have to mine 36 blocks now to get through the high diff, but we don't get many of the easy ones for all that work.

Until the net hashrate is too big for a multipool to significantly effect (which means 20+ GH/s), the difficulty NEEDS to depend on more than just the last 36 blocks. The 36 block retarget is great, but the new difficulty should be calculated from the last 2016 block times so that fast 4x diff swings dont happen.

Under your proposed algorithm, what would happen, if the difficulty appears to be too high for the value of the coin over an extended period of time (i.e., the coin sits near the bottom of profitability lists), so it keeps reducing difficulty inch by inch by means of the 2016 block average, then one day, the difficulty got low enough, that it climbs up to the middle of the profitability list - then for unrelated reasons - maybe Catcoin gets some press - the price of Catcoins gets driven up to double or triple the price from increased demand - and now Catcoins are at the top of the profitability charts - massive hash power gets directed onto the coin? Does this mean we may have to go through 1,000 blocks or so of very rapid block generation (e.g., once every 10 seconds), until the 2016 block average starts to move the difficulty up? Would such a scenario not represent a rapid inlation and influx of 50,000 newly minted CATs being dumped into the exchanges? Also, once the difficulty rises to the point where it is no longer at the top of the profitability charts, and the 50,000 CATs being dumped in the exchange causes the prices to drop, and the massive hash power gets directed elsewhere, would we not be back to having very slow block times (e.g., every 30 minutes or an hour), until the averages can get caught up, perhaps in another 1,000 blocks?

If you see it playing out differently, please describe how you see it playing out.

I described in more detail above, but to answer your specific question, the number of coins generated would depend on how much extra hash was turned on the network. If the hash goes up 10x, the block time will drop to 1 minute. After 640 blocks the difficulty would go up 40%, and after 1000 blocks (50,000 coins) the difficulty would double. After 2016 blocks (104,800 coins) the difficulty would stop rising at 10x and you would be back to 10-minutes between blocks.

The key is that the 40% rise in difficulty (after only about 32,000 coins even with a 10x hash increase) will drive off miners who are only in for the profits, unless the value of the coin rises as well. Since more coins being generated quickly would lower the price (as you noted), those miners will leave very quickly as the difficulty rises. These trends will work together to keep profitability and value BOTH stable.

[envy2010]

I was even thinking that it might be best for new altcoins to change difficulty block by block. That way the multipools will have a real job of trying to pile in on easy block periods. I can imagine it's not workable mathematically but it's a thought. Keep everyone guessing. Miners committed to that particular altcoin won't care about difficulty fluctuations.

Some coins actually do this already, but unless it's tied to a long average over the last x number of block times, it still leads to cyclical pumping and jumping of the difficulty.

[etblvu1]

I described in more detail above, but to answer your specific question, the number of coins generated would depend on how much extra hash was turned on the network. If the hash goes up 10x, the block time will drop to 1 minute. After 640 blocks the difficulty would go up 40%, and after 1000 blocks (50,000 coins) the difficulty would double. After 2016 blocks (104,800 coins) the difficulty would stop rising at 10x and you would be back to 10-minutes between blocks.

The key is that the 40% rise in difficulty (after only about 32,000 coins even with a 10x hash increase) will drive off miners who are only in for the profits, unless the value of the coin rises as well. Since more coins being generated quickly would lower the price (as you noted), those miners will leave very quickly as the difficulty rises. These trends will work together to keep profitability and value BOTH stable.

I think intentionally generating tens of thousands of extra Catcoins ahead of schedule and causing it to be dumped in the exchanges, as a means to control desirability of mining the coin, is inflationary and suppresses the value of the coin in a way that is not necessary. If we know that instability comes from miners who pile onto the coin during easy difficulty times, then leave when difficulty goes up - then the best solution to the problem should specifically address why people would choose to engage in this conduct. I submit people engage in this conduct because they are doing what is necessary to "maximize profits." If we change the reward structure so this conduct does not produce more profit (and I submit limiting rewards when this conduct is detected is more elegant than suppressing the value of the coin itself by inflation), I would submit the behavior would stop. Or if you believe something is contributing to erratic difficulty levels other than miners jumping in and out to try to maximize profit, then a more general solution may be necessary. But I have not seen anyone suggest yet that there is any other cause for this instability.

Etblvu1

[Maverickthenoob]

For the time being we've negotiated some freedom from hashcows high hash rate, and I also believe that ATM we are too small for middlecoin to effectively deploy 10GH against. We shall see how the next low diff round goes and if it much closer to the 10 minutes, if it is the diff should be a lot more stable going forward. Remember we've only had like 3 diff changes since the fork, so I expect things are still leveling out. That process should be less problematic without the cows jumping on.

Edit: Coinium getting close to 51% again, please redistribute to the smaller pools to even the load or use p2pool.

[rav3n_pl]

Attention USA CAT miners!
I have bought a VPS on US ground and put P2pool node on it!
It should be way better than mining on oversea nodes
http://p2pool-us.gotgeeks.com:9333

It is 1% fee node.

Happy mining

[vondi1122]

Free markets regulate itself   

Best quote ever.

[envy2010]

I described in more detail above, but to answer your specific question, the number of coins generated would depend on how much extra hash was turned on the network. If the hash goes up 10x, the block time will drop to 1 minute. After 640 blocks the difficulty would go up 40%, and after 1000 blocks (50,000 coins) the difficulty would double. After 2016 blocks (104,800 coins) the difficulty would stop rising at 10x and you would be back to 10-minutes between blocks.

The key is that the 40% rise in difficulty (after only about 32,000 coins even with a 10x hash increase) will drive off miners who are only in for the profits, unless the value of the coin rises as well. Since more coins being generated quickly would lower the price (as you noted), those miners will leave very quickly as the difficulty rises. These trends will work together to keep profitability and value BOTH stable.

I think intentionally generating tens of thousands of extra Catcoins ahead of schedule and causing it to be dumped in the exchanges, as a means to control desirability of mining the coin, is inflationary and suppresses the value of the coin in a way that is not necessary. If we know that instability comes from miners who pile onto the coin during easy difficulty times, then leave when difficulty goes up - then the best solution to the problem should specifically address why people would choose to engage in this conduct. I submit people engage in this conduct because they are doing what is necessary to "maximize profits." If we change the reward structure so this conduct does not produce more profit (and I submit limiting rewards when this conduct is detected is more elegant than suppressing the value of the coin itself by inflation), I would submit the behavior would stop. Or if you believe something is contributing to erratic difficulty levels other than miners jumping in and out to try to maximize profit, then a more general solution may be necessary. But I have not seen anyone suggest yet that there is any other cause for this instability.

Etblvu1

Such generation and inflation is only possible with influx of enormous amounts of hash, so this is not an end-game scenario, just how things play out before stabilizing. The coin will reach a STABLE point where adding 10x hashpower on a whim is not profitable, and this generation would not be feasible.

Also, for every period where there is fast generation, there would be a slow period - just not as noticeable as with a regular difficulty adjustment. Think of it as friction on the difficulty, that tends to return everything to a stable midpoint.