A lot of people are staring at the chips, and they think the cross strips are the "major base" and they're full of blood at the bottom. It's really not that original. The distribution of chips is essentially oneStatistical estimation modelEvery dayExchange rateAnd the deal distribution, putting the historical chips in different price bands, then decomposition over time. It's not the real data that the exchange gives you, it's the probability distribution calculated by the software, and it understands that all the formulas that follow。
What's the distribution of chips
The simplest model is actually a rule. Each trading day has a hand change rate of h, and the new amount of chips generated on that day is equal to the total value v, while the old chips are diluted - the stock at each price is multiplied by (1-h). Rolling down day after day, you get a list of all the shares left in each price slot. The larger h, the faster the old, the more concentrated the distribution is in the near future; the smaller h, the thicker the old, the longer the peak。
But it's not so rough in the software. One day running from the opening to the closing price, falling between the top and the lowest, the k-line will be broken down into a dozen or 20 segmentsTrade weightV is assigned to each paragraph. One more time decay factor, the lower the weight of the old chips, the common sense of "the early generations have changed." different software decomposition factors vary and peaks drift。
What's the chip distribution formula like
S i(t) = s i(t-1) x (1 - h t) + v t xf (p i). S i is the available leverage on price i, h t switch rate on day, v t total on day, f(p i) is the probability density on price i. The core logic is six words --Old minus, new plusI don't know。
F this function is different from one household to another. Some assume that the deal is made on the same day near the average price, either directly with the lowest even distribution, or directly with a logarithmic normal. So you change the software to look at the same ticket, and you're gonna be 78 percent off the top. It's not who's wrong, it's the bottom assumption. If you knew this floor, you wouldn't have taken one picture against another。
The next time you see the colored strips on the board, you can count them: and it tells you, "as a model, what percentage of the cost is concentrated there," not iron. Knowing where the algorithm boundaries are is not fooled by a probability map. All in。
