I'm going to take the contrarian point here:
If you're still designing elements for the game, a playtest is the wrong choice
See a much better written explanation here:
You need to stop playtesting. It's not just that you're doing it wrong, it's that you're doing it wrong in ways that hurt your games. Furthermore, by promoting a culture of design in which playtest is held up as the be all and end all of the design process, to which all other elements of design must bow down, you are causing at best cause other, newer designers to feel inferior and inadequate and at worse cause them to mutilate their own designs in ways similar to which you have mutilated your own.
Now, as he explains, there are excellent reasons to playtest, mainly to test the comprehensibility of your rules.
However, when working out the game mechanics, you should have a complete statistical model of your game. You must not rely on other people to figure out that the mechanics of your game do not match the intent.
Instead, using graph theory and statistics and game theory, you should be able to completely model the mechanical-theoretical (and mechanical-practical) domains of your game. You, as game designer, need to prove that every given combination of elements fits your intent and the expected behaviour of players.
This restriction of "complete mathematical model" also has the nice side effect of reducing reliance on silly systems that obscure and obfuscate the stats behind the scene in some vain hope that it will make them go away.
Regarding:
I currently have a Dice roll mechanic to cover combat/skill checks, character stats and skills. I have something similar to nature from mouseguard. Eventually I want to introduce feat like abilities but I don't think I'm able to make enough of them in time for the playtest.
This means that you aren't ready for a playtest. At all.
Instead, you should be using statistical models to look at all possible outcomes of the combat/skill check system and the modes of modification that feats introduce. Only when you believe that your statistical model is correct and complete (which by definition requires making all of the mechanical elements for that subsystem of your game) and rendered into some textual form, is it time to test.
You will be testing two things:
- Did you explain your system correctly?
You should be able to predict the outcomes of the mechanical interactions that your playtesters attempt. (See: making your beliefs pay rent.) If the outcomes do not coincide with your models, try to investigate where your model was incorrect or your explanation of system was incorrect.
- Did your models capture the complete range of action?
It is quite possible that you did not anticipate ranges of action being taken, and playtesting can show those.
The danger is Do not design your game for your playtesters. Instead, use them to identify flaws in your writing and your models of system.
A single PBP doesn't need all that many players, so "advertising" may not be the best approach. First consider all the techniques already listed on this site at Where can I find other RPG players? Even if the game is online, it might be nice to have friends, acquaintances, or even just people from your area in it. Here's an example of someone using meetup.com to advertise for a Demon: The Fallen PBP in the Denver area. This has the additional benefit of maybe turning into real tabletop at some point.
Also consider whether you really need to host your own PBP. There are a lot of PBP sites (and forums) out there already, like rpgcrossing.com and rpgm.com, and so PBPs on them are somewhat more self-advertising. Most are free and are simply forum-driven.
Speaking of forums - forums are still the #1 kind of site used by gamers. As a result pretty much every large RPG forum - rpg.net, enworld.org, etc. have places to advertise for games both in person and pbp, and often just run pbp there on the forums as well. Most every RPG publisher has a forum as well, and most will have dedicated subforums to these as well or at least be friendly to them in a general announcements subforum.
RPOL.net is for online forum gaming, but they do have an Offsite Gaming forum where they allow posts for other PBPs.
Best Answer
I would say take a look at online gaming pages. Most of them have the ability for you to mark it as a custom system, and then put that it's a playtest either in the title or the body. Or they have a forum where you can say that you are seeking playtesters for your custom RPG system and ask if they are interested, and then give them the link to that game.
Also most of these have the ability to designate co-GMs so that you could see it all AND also still observe instead of actively GMing.
For personal experience there, I'm active on roll20.net. There I've done a few playtests time and again by just titling the game "Playtest-Fantasy Dragons war" as example. In the description, I describe a few details about the setting and also the game system; for example:
Additionally I make a "Check here if you're interested in playtesting" forum post and link to the "group up" page for that game there.
The mentioning of how the system basically is laid out is exceedingly important in addition to the setting used, as players use BOTH of these to decide they are interested or not... Alongside the info that it is for a playtest, you should also make clear what they should expect.
Thus, if you expect to tweak rules every 2 sessions and also to get a report by the players who they like that, then you should include that information in the initial description of the system as that is also important information for them to know.
The above should hold true regardless of which online gaming site you would use (thus play-by-post and chat-based or "phone"-based gaming).