The majority of our published algos go long most of the time. Occasionaly they can short. We do have one algo with signals posted on our website that has 2 long and 1 short position every day ( a little artificial but we tried to make it resistance to prolonged market bear periods). It is called Minotaur 3. We didn't use that algo consistently in the real market (I think just for few months) and I'm not sure if any of our customers use it (we offer a bundle where that give you access to all the algos so we don't keep track of what each member trades with).

When we used MINO 3 we experienced occasionaly the limitation that a stock we were supposed to short was not available on Alpaca but one could try other brokers.

But in general the strategies described in the article are very flexible. One could design a spectrum of algorithms with the same core philosophy that go long, short, have momentum strategy or take advantage of mean reversion. The possibilities are infinite. The core philosophy is what matters, you rank your assets in terms of some predictive measure of performance and then select the outliers in this distribution.

If you have a particular market or algo characteristics that fit the general strategy described contact me and we can work on a personal customization for possible collaborator-partners.