When do you attempt to sell and rebuy lower...

[Joe200]

The setup. You think that we are at the peak and the price is about to drop. You want to sell now in order to rebuy later. How much of a drop should you expect in order for it to be worth your while to sell and then attempt to rebuy lower? Considerations:
- risk of selling too early and losing your coins
- taxes

tl;dr.
- I think a typical person would need to expect a drop of 65-70% in order for it to be worth their while to sell and attempt to rebuy later at a lower price.
- Interestingly, in the last 2 crashes, price dropped by this amount.
- For the people who trade more frequently - I just do not understand why you do it / how you can justify it. Please explain. Do you make coins?

The equations. You buy C0 bitcoins at price Pb at the beginning of a cycle. Once the price peaks, you sell them at Ps, near the top, with the hopes of rebuying more bitcoins at Pr after the crash.

Since the price increased,
Ps = f * Pb, f >> 1.

The fiat gains per bitcoin are:
g = Ps - Pb = (f - 1) * Pb = (1 - 1/f) * Ps

Total fiat gains: G = g * C0 = (1 - 1/f) * Ps * C0

At the end of the year, you'll have to pay taxes on G, at a rate of t. If, after you sell, you reserve the money to pay these taxes, then, post-tax, you gained in fiat G' = (1 - t) * G = (1 - t) * (1 - 1/f) * Ps * C0. [This assumes that you did not have any coins before buying at Pb.]

The total amount of fiat after you sell is the post-tax gains plus what you had invested originally:
L = G' + Pb * C0
L = [(1 - t) * (1 - 1/f) + 1/f] * Ps * C0
L = [1 - t + t / f] * Ps * C0

When you eventually buy back at Pr, you will have C1 coins. The ratio a = C1 / C0 is the coin gain
a = [1 - t + t / f] * Ps / Pr

The price drop from Ps to Pr is d':
d' = 1 - Pr / Ps = 1 - [1 - t + t / f] / a

The actual price drop from the peak to the trough will be greater than d' because you can't pick the exact top and bottom. Peak to trough drop: d = d' + c. Let's say c = 0.15. So

Code:

d = 1 - [1 - t + t / f] / a + 0.15.

We are interested in is the price drop d as a function of the tax rate t, the runup f, the coin gain a.

Variable values.
t depends on where you live and whether you decide to pay taxes. If you don't pay taxes or live in a country that doesn't tax bitcoin trades, t = 0. Here is the US federal tax bracket: http://www.moneychimp.com/features/tax_brackets.htm Plus, you'll pay at the state level, maybe 5%. Some reasonable values for t are 0 and (0.25, 0.28, 0.33) + 0.05 = 0.30, 0.33, 0.38.

Based on previous crashes, some reasonable values for f are 8, 17, and 36*.

a = 1 is interesting because it tells us how much price has to drop just to get all of your coins back. However, selling your coins is very risky, since you might lose them if price keeps rising. For most people, they require a >> 1 to sell. I really think a = 1.25 is the minimum any sane person would require. a = 2 is a very good gain, and a = 1.5 is probably what most people will settle for. -- What do you think are good values of a?

Results.

Consider the case when a = 1 -- you just want to get your coins back. This is not realistic. Just a baseline.

Code:

      t  f a     d
1     0  8 1  0.15
2     0 17 1  0.15
3     0 36 1  0.15
4   0.3  8 1 0.413
5   0.3 17 1 0.432
6   0.3 36 1 0.442
7  0.33  8 1 0.439
8  0.33 17 1 0.461
9  0.33 36 1 0.471
10 0.38  8 1 0.483
11 0.38 17 1 0.508
12 0.38 36 1 0.519

When t = 0, d' = 0. Meaning, you could sell and rebuy the next instance and keep all your coins. d = 0.15 because that's what I've assumed above -- by the time you notice the price going down, by the time you sell, by the time you rebuy again, with all the slippage and panic, for d' = 0, d has to be > 0.

If you are paying taxes, you would require a drop in the mid-40%, all the way up to 50% if you are in a higher tax bracket. That's just to keep your coins.

a = 1.25 -- you are pretty risk tolerant.

Code:

      t  f    a     d
1     0  8 1.25  0.35
2     0 17 1.25  0.35
3     0 36 1.25  0.35
4   0.3  8 1.25  0.56
5   0.3 17 1.25 0.576
6   0.3 36 1.25 0.583
7  0.33  8 1.25 0.581
8  0.33 17 1.25 0.598
9  0.33 36 1.25 0.607
10 0.38  8 1.25 0.616
11 0.38 17 1.25 0.636
12 0.38 36 1.25 0.646

Even if you are paying 0 tax, you still want the price to crash at least 35%. Once you figure the taxes in, you require a price drop of 55-65%. For a chance to increase your stash by 25%.

a = 1.5 -- where I think most people are in terms of risk aversion.

Code:

      t  f   a     d
1     0  8 1.5 0.483
2     0 17 1.5 0.483
3     0 36 1.5 0.483
4   0.3  8 1.5 0.658
5   0.3 17 1.5 0.672
6   0.3 36 1.5 0.678
7  0.33  8 1.5 0.676
8  0.33 17 1.5  0.69
9  0.33 36 1.5 0.697
10 0.38  8 1.5 0.705
11 0.38 17 1.5 0.722
12 0.38 36 1.5  0.73

Even non-tax payers need the price to drop 50%. With taxes, people require a drop of 65-70%.

a = 2 -- what I think is very risk averse.

Code:

      t  f a     d
1     0  8 2  0.65
2     0 17 2  0.65
3     0 36 2  0.65
4   0.3  8 2 0.781
5   0.3 17 2 0.791
6   0.3 36 2 0.796
7  0.33  8 2 0.794
8  0.33 17 2 0.805
9  0.33 36 2  0.81
10 0.38  8 2 0.816
11 0.38 17 2 0.829
12 0.38 36 2 0.835

65% required drop even without taxes. With taxes, you would have to expect a drop of ~80% to sell.

For comparison, here are the actual crashes that we've had:

Code:

    pk.date pk.price    tr.date tr.price drop    ft.date ft.price n.dest n.be
1 2010-07-19     0.09 2010-07-24     0.05 44.4 2010-07-27     0.06     NA   83
2 2010-11-07     0.36 2010-12-10     0.19 47.2 2010-12-11     0.22     11   68
3 2011-02-14     1.06 2011-04-05     0.67 36.8 2011-04-07     0.75     14   62
4 2011-06-09     29.6 2011-11-18     2.14 92.8 2011-11-24     2.42     42  617
5 2013-04-09      215 2013-04-16     65.3 69.6 2013-07-08       77     17  209
6 2013-11-30    1,130 2014-04-11      392 65.4 2014-05-07      441     17   NA

Interestingly, the past two crashes were 70%, which is exactly what I think a reasonably risk-averse person who is paying taxes would require in order to attempt to sell at the top and rebuy at the bottom.

Here is what I do not understand -- people trading coins on a more frequent basis. Why would you do this? Are you really so risk tolerant that your a is close to 1? Do you not pay taxes? Do you actually gain coins by frequent trading? Even on an after-tax basis? How?

---
* Appendix: reasonable values of f. See my discussion of previous crashes for more details.

Code:

     pk.date pk.price    ft.date ft.price r.pk.ft
1 2010-07-19     0.09 2010-07-27     0.06      NA
2 2010-11-07     0.36 2010-12-11     0.22       6
3 2011-02-14     1.06 2011-04-07     0.75    4.82
4 2011-06-09     29.6 2011-11-24     2.42    39.4
5 2013-04-09      215 2013-07-08       77    88.7
6 2013-11-30    1,130 2014-05-07      441    14.7

Based on this, here is a prediction for f (which, in the above table, is r.pk.ft):

Code:

  p_25 p_50 p_75
1 8.13 17.2 36.2

[piramida]

Here is what I do not understand -- people trading coins on a more frequent basis. Why would you do this? Are you really so risk tolerant that your a is close to 1? Do you not pay taxes? Do you actually gain coins by frequent trading? Even on an after-tax basis? How?

This I can answer - it is akin to gambling. You lose on a long enough streak because probabilities are not in your favor, but can get lucky for short periods. Like with gambling, you just have to know when to stop.

Great stuff, thanks.

[lonely@thetop]

So many numbers.. this is supposed to be speculation, please keep it that way. Can you just tell me when price is going to reach one thousand american dollars? Thank you.

[MoneyGod]

So many numbers.. this is supposed to be speculation, please keep it that way. Can you just tell me when price is going to reach one thousand american dollars? Thank you.

For this figure you have to wait for long time now because currently I don't thing its going to touch this one in near future

[hellscabane]

This is a fairly solid analysis. Although the way I look at this, there is an assumption of a somewhat static (at least in terms of a moving trendline) growth and decline period that has to oscillate in some fashion.

As for short-term trading, you can compound the given parameters and work the numbers in the opposite direction; these will be done during periods of significant deviation from the “baseline” trend. Given any growing or declining trend, there is a period of “relative stability” that acts as a sort of counterbalance. With that, any round-trip transaction that exceeds the ask-bid spread with the addition of fees and taxes, ends up being profitable in terms of fiat. The trick is determining the appropriate “baseline” trend.

[TERA]

It's not all about the percentage of the drop: It's also
-Volume
-Volatility
-How confident you are that this is in fact the top
-How many trading opportunities you will see during the volatility. You don't just have to trade once. You can trade multiple times with all that volatility. During the crashes of April and November I daytraded rapidly, multiplying my coins by 400-1000% within a week.

[Torque]

I agree, good analysis.  Some folks in the November rally incorrectly called the top in the 400's and prematurely sold.  Then they had to basically wait 6 months to buy back their coins for close to what they sold them for.  In the end they really gained nothing by selling.

So from that, I gather that it is far better to wait until after the bubble has popped (provided the price will rally back to a double top or something) to assess if it is a good time to sell or not.  After December, folks had the chance to sell at 995, 800, 700, 650, etc., etc.

Also, I think bubbles in the future may fall less and less from the top each time.  This one is currently hovering around 39% of the peak.

[wachtwoord]

Never.

[UglyTroll]

Don't ever say "Never" 

[Wandererfromthenorth]

Always.

[wachtwoord]

Don't ever say "Never" 

Trading is a losing game. You sell because you need liquidity to by things you need. You sell because you wish to re-balance your portfolio. You don't sell "to rebuy lower".

[TERA]

Don't ever say "Never" 

Trading is a losing game. You sell because you need liquidity to by things you need. You sell because you wish to re-balance your portfolio. You don't sell "to rebuy lower".

Please do not say this as a global assertion just because it applies to you personally and to inexperienced traders. There are actually people who are very effective at trading. Not trading may be what is right.. for you, but not for everyone.

[wachtwoord]

Don't ever say "Never" 

Trading is a losing game. You sell because you need liquidity to by things you need. You sell because you wish to re-balance your portfolio. You don't sell "to rebuy lower".

Please do not say this as a global assertion just because it applies to you personally and to inexperienced traders. There are actually people who are very effective at trading. Not trading may be what is right.. for you, but not for everyone.

It is different when you are a market maker (only big institutions can be) but other than that it's like people at the casino claiming they have a system. And some people get lucky. That's what keeps the casino going.

[Wandererfromthenorth]

that's exactly why while trading you should think yourself as a casino instead of a gambler that throws money away at slot machines.

-Your (even basic) technical analysis, entry and exit strategies make it more probable to engage in a winning trade than in a losing one.
-losing trades must be closed early compared to winning ones.


With multiple trades in the end you have a profit.
You can make a very good profit even if you win "only" 60-70% of your total trades

[oda.krell]

What an awfully complicated way to set up a very basic question: how often do you trade, and do you make a consistent profit doing so.

I don't see how the taxes factor into the frequency of trading. In the tax system that applies to me, any selling of btc before a year will incur a tax depending on my personal tax rate, so any profit of an individual trade will need to be lowered by that tax rate. If the trade itself was profitable though, it will be so even after tax.

Here's a list of the trades cycles I took part in during this bear market. I'm sure you can easily find the corresponding dates on the chart.

sell at 1200, buy at 400: ((1200*0.75)/(400*1.25))*0.95 ~= 1.7 times btc
sell at 1000, buy at 400: 1.4 times btc
sell at 700, buy at 340: 1.17 times btc

The assumption here is that you missed the optimal selling point after a 25% decline, and the optimal buying back point after a 25% increase. Then take off 5% for slippage and fees. Taxes not included in this calculation yet. The overall result is about 2.8 times the amount of btc than initially held. The same could have been achieved of course with a single perfect trade (from 1200 to 400), but it's not a very realistic assumption.

The numbers above are not exactly what I personally traded, but not completely unrelated to my trades either.

[TERA]

Trading requires a lot of discipline, and experience. You have to know how to properly guage momentum, read indicators, calculate targets, follow indicators for your closing trade, set stoplosses, improvise, make quick judgements, and accept defeat. Most people don't have any background in trading and do not use one of these components, and so they fail. Therefore they think trading is 'rigged' or 'futile'.

One guidance that professional traders use is that they only have to be right about the direction of the trade about 30% of the time and they will still come out ahead as long as they are cutting their losses and letting their winners run.

[chessnut]

Blind people can be trained to process a signal from a digital camera through the sensitive nerves of their tongue and in a controlled environment make their way through a room full of obstacles without collision with any of them.

the only indicator/oscillator one needs is a chart, and eyes will always serve one better than a macd will, if one trains them. the question is only discipline.

[Wary]

Trading is a losing game. You sell because you need liquidity to by things you need. You sell because you wish to re-balance your portfolio. You don't sell "to rebuy lower".

Trading is a sero-sum game, i.e. it is a losing game for the less smart half of players and winning game for a more smart half of players. So the answer seems to be an easy one: if you're smarter than average, do trade, if not - don't.

But:

1. If you are asking if you should trade, you belong to the less smart/knowledgeagle half. So it's kind of catch-22: answer "yes, do trade" should be given only to those who don't ask the question.
2. Most of people consider themselves to be smarter than average. Trading would be very costly method of self-deception.
3. Winners to losers ratio in trading is probably 1:10 rather than 1:1. So even if you are really smarter than 8 out of 10 bitcoin traders, you still may belong to the loser's camp.
 

[Wary]

The assumption here is that you missed the optimal selling point after a 25% decline, and the optimal buying back point after a 25% increase.

So your strategy is not to try to guess the top/bottom beforehand, but detect them after the fact?

[hellscabane]

Trading requires a lot of discipline, and experience. You have to know how to properly guage momentum, read indicators, calculate targets, follow indicators for your closing trade, set stoplosses, improvise, make quick judgements, and accept defeat. Most people don't have any background in trading and do not use one of these components, and so they fail. Therefore they think trading is 'rigged' or 'futile'.

One guidance that professional traders use is that they only have to be right about the direction of the trade about 30% of the time and they will still come out ahead as long as they are cutting their losses and letting their winners run.

Yeah, this is what a lot of people tend to do in this realm. A lot of the trading depth in the crypto-currency world comes from those who trade through emotion and expectations of winning. The mindset that they have to win, they should win, or that they will win because they are "better" ends up beating a lot of people.

Of course though, one has to also recognize that in this scene, a lot of the traditional fundamentals do not apply because of the lack of regulation, constant dilution and exceptionally high risk that exists with Bitcoin (and all of the many alternate crypto-currencies).