I must apologise for not actually posting a response earlier - I played your game on the day you posted this thread...and to be honest have been somewhat in two minds about the entire experience. Having been to a number of sheepdog trials in my past and watching the fluidity of the dog with its handler, herding sheep through various barricades I did struggle to find that experience within your game -- The principle is straightforward enough and I experienced little difficulty in pushing the sheep through to the barricades...though I did find that herding tended to concentrate on indivdual sheep near the end part of each level to push them through - somewhat of an immersion breaker for myself personally.

I struggled however to think up ways in which to improve the basic design of the game you have created itself persay (mostly because the lack of real time threw me I think). Instead I kept on thinking of different versions of utilising the same map board and ways to make it more challenging. The best idea (imo) that I came up with was using an analogy of a hunter (animal) and it's (live) food -- The hunter has x amount of energy - each move expends energy -- catching food restores energy -- the goal of the game being to clean the board before you run out of energy. Introducing different types of food prey with different movement systems could provide an increasing level of difficulty not to mention the addition of physical barricades to navigate consuming energy.

Sorry that I cannot provide better feedback for you.

I think it's a bit silly to criticize this prototype on the grounds of immersion and realism in a sheep herding context. Unless your objective was a sheep herding simulator! But I'm assuming otherwise, and in terms of gameplay fun, although the prototype is pretty boring, I like the idea of turn based herding but it needs to be thought through..

Try to make different "pieces" that have different movement and different effect on the sheeps. It might be difficult to do that if you're not willing to loose the sheep and dog context (in this context, the herder is just a slow piece that serves no purpose). For example, pieces that are slower but have a larger influence, pieces with a smaller influence but are fast, pieces that attract, pieces that repel etc etc.. Maybe add some danger and, with it, a whole world of possibilities, pieces that interact with enemies in a way and with the sheep in other ways, pieces with attacking powers, defence powers, repelling powers, pieces with great attack but make no distinction between enemies and sheeps. phew! the sky is the limit

Whether to finish it or not really depends on how important the project is to you. Personally, I don't think it'd be a waste of time to get what you have to a state that you're comfortable calling finished. What that state is depends on how far you want to go. It could be just a few tweaks to the interface and creating a few more levels or it could be hiring an artist and looking into making it a mobile app.

Some quick thoughts... I had to read the instructions to know how to use the controls. Using the center button and wheel isn't as intuitive to me as it might be for others. I'm not sure how the current mechanics of move and then press a key would be received by average players. Moving the shepard one space at a time is... well he moves really slow and I don't really want to move him if he's not already in a helpful position. Still, I'm not so sure that removing him from the game would be a good idea as he adds a bit of... something to the scene. The ability to move many spaces isn't really that critical anyways as I find when I move the dogs, I don't usually use the full moves that are available to me. Often, I get the dogs & shepard into position and press space a few times to see if the sheep do what I expect. These short 'bursts' of movement allow for better control.

I found I had an easier time understanding what to do by playing the 4th level first. And that the 1st level was actually a more advanced level due to only having 1 dog.

As the game stands right now it reminds me of older games from the Win3.11 to Win95 days. Fun for awhile and maybe you get hooked for a bit but eventually you get distracted by something shiny somewhere else. I was hooked enough to want to finish the levels I started but not so much that I was looking to start new levels.

If you already have grand ideas for this game then it's difficult to say what effect additions might have. If you are planning additions you need to consider the near complete random movement of the sheep. I do think that adding in elements of danger that eliminate sheep would quickly get frustrating due to the very limited control over the sheep. It might be better to add elements that either affect the layout of the level or somehow affect the PCs. No suggestions come to my mind though.

One quick gameplay experiment I can think of would be to force the player to use all possible moves available to him. Maybe even require that you can't backtrack to a previous position. When all moves are completed, move the sheep automatically. I can imagine this having interesting potential when there are very few sheep but it may be frustrating when there's a larger heard. It may also have the potential to make obsticals more relevant to the game.