How do you value a stock? The market value is obvious — it’s whatever the shares trade for — but what about it’s real, intrinsic value?

Determining a stock’s true worth is a crucial part of value investing, the discipline made famous by Warren Buffett. Many investors also subscribe to it, but relatively few self-titled value investors actually know how to roll their sleeves up and value a stock.

[See: The 25 Best Blue-Chip Stocks to Buy for 2017.]

Which makes sense, because it’s not straightforward.

“The subject is a science, but a soft science,” says Angelo DeCandia, professor of business at Touro College in New York, of valuation.

That’s one reason it’s so hard to do. But if you can master stock price valuation, you can also become very rich. Below are four common ways to value stocks.

Peer comparisons. One of the most frequently used methods for figuring out what a company’s valuation should be is to use ratios, many of which individual investors will recognize already. A few common ratios area price-earnings, price-sales and price-book.

There are many more ratios that investors can use to impute an accurate price for a company’s shares, each with their own advantages and disadvantages.

“Regardless of the metric, all those in this category are suggesting that a company’s value is relative to the value of other, similar companies,” DeCandia says of comparative ratios.

If a company is growing sales and earnings faster than its peers, it probably deserves a premium valuation, and therefore a higher “multiple,” or ratio. Once you’ve determined what multiple you think the market should pay for a stock, you can use future projections to figure out how much a stock should be worth at, say, the end of next fiscal year.

Say you think Facebook (ticker: FB) should trade for 35 times earnings, and you expect it to earn $6.10 per share in 2018. By the beginning of 2019, then, Facebook should be trading for 35 x $6.10 = $213.50.

PEG ratio. The PEG ratio is shorthand for the price/earnings-growth ratio. It takes the price-earnings figure (the “multiple”) and divides that by the expected earnings growth rate. If Alphabet ( GOOG, GOOGL) trades at 30 times earnings and earnings are expected to grow 18 percent annually, the PEG ratio is 1.67. The lower the PEG ratio, the better.

Discount models. “The value of any business should be simply the present value (that is, value today) of all expected future cash flows from that business,” says Bob Johnson, president and CEO of the American College of Financial Services in Bryn Mawr, Pennsylvania.

That’s the academic description of how you determine the value of any financial asset, and as you might guess, it involves some calculation and assumptions.

Basically, to compute the value of a stock using this method you’ve got to project its future cash flows, then discount them to determine what their present value would be.

For dividend stocks, the dividend discount model is a popular formula. The most common version of the DDM looks like this: true price of stock = annual dividends per share next year/(discount rate – dividend growth rate)

Where the discount rate is the rate of return you need to earn on your investment.

Let’s say the McDonald’s Corp. ( MCD) dividend next year is expected to grow 5 percent — the same rate by which you assume the dividend will continue to grow. If last year the dividend was $3.66 and you want an 11 percent return, this is the equation: true stock price = $3.843/(.11 – 0.05), or $64.05.

The discounted cash flow model is a similar concept, except it takes the present value of all future free cash flows the company earns, not the dividends it pays. Calculating a stock’s value using the DCF model can be done with calculators (real or online) or spreadsheet programs.

The conceptual method. Yet another method is more conceptual and is the favored tool of Aswath Damodaran, a New York University finance professor who is famous for his valuation techniques. Damodaran believes that in order to value any company or determine any stock price, you must first articulate what you think its narrative will be.

[See: 7 of the Best Stocks to Buy for 2017.]

Since no one knows the future, you need to estimate what the probable narrative is, what some plausible narratives are, and what is a merely possible narrative.

Using discounted cash flows, you then value each scenario, making sure to heavily discount the less likely scenarios before adding them together.

But there are challenges. K.C. Ma, professor of finance at Stetson University, takes issue with the P/E ratio, even if it is the most visible and popular metric.

“It’s not forward-looking, and it fails to incorporate the different risk level of each stock,” Ma says. “The fair value of a stock should reflect the forward-looking fundamentals and be adjusted for the corresponding risk level.”

The P/E ratio is often used to judge how over- or under-priced market indices like the Standard & Poor’s 500 index are, too — it’s not just a yardstick for individual stocks. The P/E ratio also has flaws in that respect.

“Many people assume the market is oavervalued because P/E ratios are near the high end of their historical range,” Johnson says. “But merely looking at P/E ratios in isolation is problematic. P/E ratios must be looked at in the context of market interest rates.” In low-rate environments, companies have lower costs of capital, elevating stock prices.

Methods like the DCF model, too, fall far short of perfection.

In the past, Apple ( AAPL) and Amazon.com ( AMZN) have been two clear examples of why. If you tried to use a DCF model to value Apple before the iPhone launched, your calculations would have been wildly off, and could even have suggested AAPL was a dog.

As for Amazon, markets have consistently undervalued it as well — and not because fund managers weren’t working around the clock to figure out its true worth. Using the DCF model to calculate a fair price for a stock tends to fall apart when there are too many ways a company could evolve.

So while using standard multiples, discounted cash flows and other techniques can help you value a stock, none are perfect, and different methods work better for different companies.

[See: 10 Long-Term Investing Strategies That Work.]

What all methods have in common is that calculations done with a little dose of humility are always more realistic. The stock market is unpredictable because businesses and investor psychology are unpredictable, and those, ironically, are variables that will never change.

More from U.S. News

7 of the Best Cheap Stocks to Buy Under $10

The 10 Most Anticipated IPOs of 2017

5 Reasons Donald Trump’s Presidency Will Include a Recession

The Best Ways to Value a Stock originally appeared on usnews.com