How does the Brave Browser "-tor" option work differently than the Tor Browser Bundle?

How does the Chromium-based cross-platform Brave browser "-tor" option work
differently than the Firefox-based cross-platform Tor Browser Bundle?
https://brave.com/download/

Linux: brave-browser --incognito --tor
Windows: C:\Program Files (x86)\BraveSoftware\Brave-Browser\Application\brave.exe --incognito --tor

I know only the basics of how the Tor Browser Bundle works, at least
superficially, where you first get a directory server and then you have an
entrance, middle, and exit node, where each node-to-node traffic is
encrypted separately.

I can "presume" that the Brave browser "--ingognito --tor" option set works
similarly, but, maybe it's completely different in how it works.

I see the question on the net of how the "Brave browser with tor" differs
from the "Tor Browser with tor", where the answers imply "something" is
different, as shown below.

o Brave adds Tor to reinvent anonymous browsing
https://nakedsecurity.sophos.com/2018/07/02/brave-adds-tor-to-reinvent-anonymous-browsing/
"For users who currently require leakproof privacy, we recommend
using the Tor Browser, which provides much stronger and well-tested
protection against websites or eavesdroppers using advanced techniques
to uncover a true IP address."

o Brave Browser Integrates Tor Into New Private Tab Feature
https://www.pcmag.com/news/362191/brave-browser-integrates-tor-into-new-private-tab-feature
"What's different with Brave's implementation? Mainly convenience.
You can open normal tabs in the browser, and then Tor-powered
tabs alongside them ¡X all within one window. ...
Brave's implementation of Private Tabs using Tor is experimental
and has some known leaks which we intend to fix in future versions"

Do you have further knowledge or information on how the two browsers differ
in terms of tor-based anonymity?

On Tue, 26 Mar 2019 04:36:56 -0000 (UTC), arlen holder wrote:

How does the Chromium-based cross-platform Brave browser "-tor" option work
differently than the Firefox-based cross-platform Tor Browser Bundle?

I ran a quick test of http://panopticlick.eff.org/ fingerprinting.

*TBB*: Within our dataset of several hundred thousand visitors tested in
the past 45 days, only one in 752.26 browsers have the same fingerprint as
yours.

*Brave*: Your browser fingerprint appears to be unique among the 297,725
tested in the past 45 days.

The Tor browser conveyed around 9 bits of identifying information; while
the Brave browser conveyed _twice_ that much identifying information!

The big differences appeared to be that Brave was worse in the following:
o Hash of WebGL fingerprint
o Hash of canvas fingerprint
o Time Zone
o HTTP_ACCEPT Headers
o User Agent

Most of the rest were either exactly the same, or close enough.
o Interestingly, TBB was slightly worse than Brave on "System Fonts".
o And TBB was twice as bad on "Screen Size and Color Depth".

As a control, I ran Firefox, which came up with the same number of
identifying bits of information as did the Brave browser in tor mode.
*Firefox*: Your browser fingerprint appears to be unique among the 297,748
tested in the past 45 days.

Some of the offending items we
o Hash of WebGL fingerprint (it was about double Tor but half of Brave)
o Hash of canvas fingerprint (it was about double Tor but half of Brave)

Interestingly, the identifying bits from "System Fonts" was lowest on
Firefox, but not appreciably so as they were all within one or two bits of
identifying information.

It's hard to summarize, but clearly the Tor Browser Bundle had, overall,
about half the number of bits of identifying information leaked than did
either Firefox or Brave in anonymity mode.