Mistake:
Wrong:     left: 50%; translateX(-50%);
Correct:    left: 50%; transform:translateX(-50%);

Able to reproduce the issue on reported chrome version #70.0.3538.102 and latest chrome #72.0.3615.0 using Windows 10 by following steps as per comment#0.The following is bisect information.
Note: Issue is not seen on Mac OS and Linux.

Bisect Information:
==============
Good Build: 62.0.3176.0
Bad Build: 62.0.3177.0

Change Log: https://chromium.googlesource.com/chromium/src/+log/1ca1d303d8c55d0b079cd8f3504fbb9835fc4b54..6756bd6774abc6149e65f5fb55854ab4973e3bbe
Skia-deps-roller Change Log :https://skia.googlesource.com/skia.git/+log/5910ed347a63..121ad19a7ece

Unable to find the correct suspect from the above change log hence marking it as untriaged. Requesting someone from Blink>Scroll to help us in assigning the issue
Thanks.!

Looks like pixel snapping issue so the bisect makes sense. I suspect it's:

Update SkClipStack's bound computation

But +cc both CL authors.

I've found that centering has nothing to do with this issue, and it is always caused by "translateX(0.5px)".
Simpler sample:
http://sinst.html.xdomain.jp/zengaku/jkkn/chrome_bug10_.html

Here is a Skia fiddle that more directly demonstrates the issue:

https://fiddle.skia.org/c/e0062438cca5c8379dca827d186ea2a8

In short, non-AA rects are being drawn through non-AA rect clips with a non-integer translation.

The following revision refers to this bug:
  https://skia.googlesource.com/skia/+/9548c3b42ffaefa27541c987ef34ba8491abb584

commit 9548c3b42ffaefa27541c987ef34ba8491abb584
Author: Robert Phillips <robertphillips@google.com>
Date: Wed Jan 09 21:35:30 2019

No longer round the non-AA clip bounds

By no longer rounding the non-AA clip bounds we increase the probability that a non-AA clip will be dropped when applied to a non-AA rect draw (particularly at half pixel translations).

To make this work in practice we must also make the clip stack check if the non-AA clip is relevant before we preemptively employ a scissor clip.

Note that I have verified that this removes the gross clipping errors from the Chrome repo case but the rounding to scissor clip behavior could still happen in other cases.

Bug:  906496 
Change-Id: Iba51b9061fb434144e3a9b3fd91479109fcf67f4
Reviewed-on: https://skia-review.googlesource.com/c/182141
Reviewed-by: Michael Ludwig <michaelludwig@google.com>

[modify] https://crrev.com/9548c3b42ffaefa27541c987ef34ba8491abb584/src/core/SkClipStack.cpp
[modify] https://crrev.com/9548c3b42ffaefa27541c987ef34ba8491abb584/src/gpu/GrRenderTargetOpList.cpp
[modify] https://crrev.com/9548c3b42ffaefa27541c987ef34ba8491abb584/src/gpu/GrReducedClip.cpp

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/eb4ceeddc3172e444b1b0e2c9c72389db8ca6378

commit eb4ceeddc3172e444b1b0e2c9c72389db8ca6378
Author: chromium-autoroll <chromium-autoroll@skia-public.iam.gserviceaccount.com>
Date: Thu Jan 10 03:43:46 2019

Roll src/third_party/skia a7b2874b86e0..d2fa7e33798c (7 commits)

https://skia.googlesource.com/skia.git/+log/a7b2874b86e0..d2fa7e33798c


git log a7b2874b86e0..d2fa7e33798c --date=short --no-merges --format='%ad %ae %s'
2019-01-10 recipe-roller@chromium.org Roll recipe dependencies (trivial).
2019-01-09 herb@google.com Make pointers to glyphs in the SkGlyphCache stable
2019-01-09 robertphillips@google.com No longer round the non-AA clip bounds
2019-01-09 halcanary@google.com Docs: drawText -> drawSimpleText
2019-01-09 halcanary@google.com tools: Remove sk_tool_utils::set_portable_typeface()
2019-01-09 halcanary@google.com drawText Cleanup, part 6
2019-01-09 skia-autoroll@skia-public.iam.gserviceaccount.com Roll third_party/externals/angle2 4f3b207d049a..b5ba549abf2f (3 commits)


Created with:
  gclient setdep -r src/third_party/skia@d2fa7e33798c

The AutoRoll server is located here: https://autoroll.skia.org/r/skia-autoroll

Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+/master/autoroll/README.md

If the roll is causing failures, please contact the current sheriff, who should
be CC'd on the roll, and stop the roller if necessary.

CQ_INCLUDE_TRYBOTS=luci.chromium.try:android_optional_gpu_tests_rel;luci.chromium.try:linux-blink-rel;luci.chromium.try:linux-chromeos-compile-dbg;luci.chromium.try:linux_optional_gpu_tests_rel;luci.chromium.try:mac_optional_gpu_tests_rel;luci.chromium.try:win_optional_gpu_tests_rel

BUG= chromium:906496 
TBR=bsalomon@chromium.org

Change-Id: I3a10167e440551c784561dd9c09eceac1ec14a93
Reviewed-on: https://chromium-review.googlesource.com/c/1403964
Reviewed-by: chromium-autoroll <chromium-autoroll@skia-public.iam.gserviceaccount.com>
Commit-Queue: chromium-autoroll <chromium-autoroll@skia-public.iam.gserviceaccount.com>
Cr-Commit-Position: refs/heads/master@{#621442}
[modify] https://crrev.com/eb4ceeddc3172e444b1b0e2c9c72389db8ca6378/DEPS