Just over a month ago, the was setting record highs.
Now, more than 70 percent of the index's components is in a correction or worse with some high-profile names like Twitter, Caterpillar, Ford, and AMD deep in a bear market.
It hasn't found a bottom yet, according to Craig Johnson, chief market technician at Piper Jaffray.
"We had this kind of fake-out/breakout scenario. It's ended up becoming a fake-out similar to what we had seen in 2000 and 2007 and now we're coming back down to retest support at 2,700," Johnson said Tuesday on CNBC's "Trading Nation."
Before Tuesday's session, the S&P 500 had not traded below 2,700 since briefly touching and bouncing off that support level in late June. It is also the level it plunged below during its February sell-off, when it sank to 2,532.
It ended Tuesday roughly 1.5 percent above it.
"Usually when we get to this kind of weak internal readings we end up seeing some sort of flush-out. I think that flush-out is still ahead. I think it's going to be another 5 to 10 percent lower from here and it's probably going to take about 14 to 16 weeks to work out itself out," said Johnson.
A 5 percent decline would take the S&P 500 to around 2,603, its lowest level since early May. A 10 percent drop from current levels would push it negative for the year and 16 percent from record highs.
"The long-term uptrend support line comes in around 2,500 on the S&P," said Johnson. "That's where I think you'll find an investable bottom to get made and it probably is going to be a little bit longer before we get there."
Stacey Gilbert, market strategist at Susquehanna, says while options investors are betting on a recovery, any shock to the market could produce an even more dizzying decline.
"The majority of the flow that we see is investors selling out of their protection, buying some upside," Gilbert said on "Trading Nation" on Tuesday. "My biggest concern for the market right now is not what's known, it's the potential gap risk that's out there. … If we do go down, I am worried that the down drop is significantly larger than obviously ... what we would see on an upturn."
