There are few movements that are completely violent or nonviolent. There isn’t really any scholarly consensus on whether a radical violent flank actually helps the “core” nonviolent group, or under which circumstances it helps the group. I recommend reading “Why Civil Resistance Works” by Erica Chenoweth for an overview of the factors that lead to the success (or failure) of resistance as well as more concrete examples.

The numbers speak for themselves - violent resistance can succeed, but nonviolent resistance is more likely to succeed. The key finding from Chenoweth’s review of resistance movements between 1900-2006 is that the strongest determining factor in whether a resistance is successful is the percentage of the populace actively involved in resisting. This seems like a “water is wet” finding, but consider the difficulties in recruiting members to violent resistance (training involved, physical ability barriers, moral barriers, informational barriers - hard to advertise for recruits without informing on yourself, etc.) vs nonviolent resistance (almost no training necessary, easy to inform about time/place, fewer physical ability barriers, etc.)

Here’s the full passage about flank effects:

The coercive capacity of nonviolent resistance is not based on violent disruption to the social order. Rather, it is based on the removal of the adversary’s key sources of power through sustained acts of protest and noncooperation. Some may argue that nonviolent resistance is powerful only because regimes fear that they will become violent, thereby posing even greater threats. Social movement scholars refer to this dynamic as a “positive radical flank effect.” This concept posits that violence may sometimes increase the leverage of challengers, which occurs when states offer selective rewards and opportunities to moderate competitor groups to isolate or thwart the more radical organizations. In other words, the presence of a radical element in the opposition may make the moderate oppositionists in the nonviolent campaign seem more palatable to the regime, thereby contributing to the success of the nonviolent campaign. In this way, some argue that violent and nonviolent campaigns can be symbiotic, in that the presence of both types improves their relative positions.¹⁸

But opposition violence is just as likely—if not more likely—to have the opposite result. A “negative radical flank effect,” or spoiler effect, occurs when another party’s violence decreases the leverage of a challenge group. In this case, the presence of an armed challenge group causes the regime’s supporters to unify against the threat without making a distinction between violent and nonviolent challenges, which are lumped together as the same threat deserving the same (violent) response

There is no consensus among social scientists about the conditions under which radical flanks either harm or help a social movement.¹⁹ In our estimation, however, many successful nonviolent campaigns have succeeded because they systematically eroded or removed entirely the regime’s sources of power, including the support of the economic and military elites, which may have hesitated to support the opposition if they had suspected that the campaign would turn violent. The more a regime’s supporters believe a campaign may become violent, or that their interests will be gutted if the status quo is changed, the more likely that those supporters and potential participants may perceive the conflict to be a zero-sum game (Stephan and Chenoweth 2008, 9–13). As a response, regime supporters are likely to unite to counter the perceived threat, while potential participants may eschew participation for the reasons just identified. A unified adversary is much harder to defeat for any resistance campaign. In conflicts perceived as zerosum, furthermore, it is difficult for erstwhile regime supporters to modify and adapt their ideologies and interests according to shifts in power. Instead, they will fight tooth and nail to keep their grip on power, relying on brutal force if necessary. There is less room for negotiation, compromise, and power sharing when regime members fear that even small losses of power will translate into rolling heads. On the other hand, our central point is that campaigns that divide the adversary from its key pillars of support are in a better position to succeed. Nonviolent campaigns have a strategic advantage in this regard.