Doesn't look all that much of a trend to me.
The 'decline in war' is mostly a statistical sleight of hand imo.
1. It assumes violence should scale proportionally with world population, which is a very questionable assumption. Why should we say we are getting more peaceful if war deaths increase at even 0.1% less than population? This is technological progress in keeping people alive, not a reduction in war.
2. Historical data tends to massively overstate deaths in conflict for reasons of a) propaganda b) adding together disease, famine, etc deaths that occurred in the decades surrounding major conflicts and that may or may not have been connected. We also don't add Spanish Flu to WW1 deaths for example, or global COVID deaths to the War in Yemen. As such any analysis of the data will be suspect (even without the flagrant cherry picking Steven Pinker does in his
Better Angels by using high estimates for most historical wars, and mid to low for more modern wars. My personal favourite is him using a footnote to say his very high An Lushan Rebellion number of 30 million deaths is pretty dubious yet he still actually uses it rather than choosing a lower one. Not exactly the height of academic integrity
)
3. The 20th C was violent
even accounting for massive population growth, and the "decline" in violence since will require another 200 or so years before we can claim it as a trend rather than randomness. One single massive war would reverse the trend and show an increase in violence.
If you are interested in a more scholarly argument against it: