Ladders in 1.0

I'm talking about released games in this genre that I've worked on.

Really there are three groups:
- single player only
- casual multiplayer with friends and with/against AI
- competitive multiplayer

The third group is the one that gets traction out of matchmaking and ladders. However, you need to have that group be large enough to actually keep it going consistently. This means you need a lot of players online to make those features work. It's not a coincidence that the games which people keep naming (SC2 for example) have huge numbers of players online. These guys sell many millions of units and therefore have the resources and the player base to cater to.

It's really not obvious to us that emphasizing competition is the way to go for initial release at all. Especially since we are trading off against features that everyone will use (more UI polish for one).

Then we get a bunch of people who are very very noisy about the feature and derail our plans and make us rethink. Sometimes we come to a different conclusion but honestly I'm not seeing really great arguments on this one. The loudness has made us reconsider but the arguments aren't all that compelling so far.

So far the arguments seem to boil down to:
1) The game will fail and have no users without a ladder - ok, we disagree. Our data shows the opposite. Where to next on this? Why not let us worry about the sales and marketing of the game? We respect your opinions but you guys really don't have good data on this stuff.

2) I personally want this feature - ok, great. It will be coming in a future update but isn't prioritized for 1.0.

What are the other arguments? I really need solid stuff here.

I've been trying to get something going that caters to you casual folks (I like to call myself casual, but who am I kidding, right?). It just hasn't hit the finish line yet.

I love this idea. Put the AI on the ladder

Pointing to PA stats is certainly something we would consider.

Just a short interruption here. I think your math is flawed.
1. 5/50 times 5/50 is not correct because I want to know the probability for me when I click the button. So I have only 5/50 where I have success.
2. I don't need the 3/4 chance of success because the formula already counts the success probability.
3. 2/168 times 2/168 isn't correct either. This one is hard to get and I'm not sure what is right here. I think your number is correct if the group has 2 people. For more people it is wrong. But I can't calculate the right number because it is way too complicated for me. However I think this is a variation of the birthday problem (the variation where you look for a specific day). But I don't know if I'm right here, it seems more complex than the birthday problem because you don't need people to be online in the exact same hour. I think if you want that and have an average playtime of 1 hour per week, you have a 99,7% chance of someone being online in the same hour as you, if there is 1000 people. The formula is P = 1 - q^n with P being the probability, n being the number of people and q being the probability for one person not being online in a specific hour q = 1 - 1/168.
n = (ln(1-P))/ln(1-(1/168)) is the number you need for a probability P (you can never reach 100%!). So if you want 99,99% it is: n = 1542,72. With this number of people someone will be online in the same hour as you in 99,99% of all cases. If the people go online randomly. I hope this is correct. There are some simplifications behind this calculation like: 1 person will come online 1 hour per week and stay online for this 1 hour.
4. Math is exciting but too complicated for me
5. I think the number required for matchmaking is way lower than your number because of the things I mentioned above.

Can someone do the full calculation please?

Sorry I know this is off-topic but I got excited about finding an answer to this problem. Still looking and questioning my approach though...

At least the newbs will have something to strive for, right? Especially with how well Sorian has made the AI....kinda.....

Ok....Solid stuff....Pizza sounds good, but Bacon is better...hmm....

Exquisite coffee? no, already got that.....

Here's my answer to your request of 'Do me a solid':

An in-game ladder, by nature, will grab more folks than a modded ladder will. Sure, you've got those folks who go LOOKING for it, but then you have the guys who kinda just run around playing, and happen across it. That's the group I'm predicting will help grow the competitive scene from day one. The more of those people we have around, the better, because they are the ones viewing streams, watching videos on Youtube, and sharing that on Facebook. They discuss it lightly at work. The name gets passed around in younger, less mature circles (high school and college, for example). As Matiz mentioned, it snowballs. What I'm getting at here is that except for the problems involved with delaying the release, building a matchmaking and ladder system is a win-win situation.

As I said in an earlier thread, you all decided through (hopefully) extensive research and number-crunching that your short-term gains by releasing earlier will outmatch your long term retention by releasing later, but with the MM and ladder. I may not agree with those results, but i cannot find fault if you followed them. I'm truly grateful you decided not to.

Also, question: Is it going to be Trueskill or ELO or what? Can we please NOT HAVE ELO?? Thanks a TON!

For PA Stats to work you really need to open up your authentication system to modders. PA Stats has already been seriously hacked at least once because it is forced to blindly trust client reporting. System Sharing could really benefit from oAuth or some other way to verify users too.

PA Stats so far has no skill based matchmaking though. That's because as a mod it has only a fraction of the player base and it's even harder to get enough players. That's why I think a build in ladder is that important. It reaches ALL players. It can be so much more active than what PA Stats has.

Though if you do "point to pa stats" in a meaningful way I'd surely look into improving it.

@cptconundrum is right, having some official auth thing would be very helpful, but not required. Using the ingame chat and a bit of trickery would in theory allow me to implement my own auth already.

With all of this being said. if 1.0 comes in lets say September. I'm totally thinking of a version of Dota 2's "the tournament" in the future for PA. Maybe with 3vs3 or something and like the prize is like 1k or 10k. for each player on the winning side. We can make big sponsers and I'm such something like a mac mini or a Planetary annhilation themed PC in the future for the grand prize would be nice.

To sum it up I want to host a Tournament 2014 type thing for PA in the future.

I'm just glad to see people taking a mathematical approach to it. I'm not sure I completely followed the original math but our math shows you do need quite large groups of players to make matching work effectively. Far larger than the amount of concurrency we've ever had or anticipate anytime soon.

Yup, love this idea. We just need a player base big enough to make it matter.

We have to build up to this. Starting out like this might lead to it falling on its face (*cough* planetside 2 *cough*). Once we start getting the demand for such a tourney, we will be able to supply it.

Yeah there are some issues here. Probably best to talk about it offline.

Would you mind involving me in this conversation? There are improvements I would be making to the System Sharing mod if I knew that I could trust user input.

Well i just wanna trow my 2 cents in this discussion:

I'm a relatively new player, i bought the game when it came out on Gamma. And i never played a single game against the AI, until the galactic wars patch came out even then i only played for a day. I like to test the strategy against other players. They always surprise me, Usually when I play against AI once I figure a recipe to win I just use it again and again, and the game grows stale. I don't have many friends that like this type of game. So i just trowed myself in the random matches with teams, and only shared, cause I thought i can learn by watching others and help a little bit.

And now i feel like i know what I'm doing, I still like play in teams cause i like the cooperation. But now I see a real problem for experienced players and new ones. The new players get shredded in this game, and when I enter a lobby I don't know the level of players I have on my team or that I'm going up against. Thats why i stoped playing 2v2 cause if the player in your team is a newbie and in the other team there are 2 vets u get shredded. In 3v3 you have more chance for compensating for a player who still don't know what he is doing. Is better to fight if the teams are at least a bit balanced, but right now we don't have a way of doing that.

My goal in this game is not to compete in a e-sports scene, or pro level. I just wanna have fun. And it's no fun to win against a opponent that never stood a chance, as its also not fun to get stomped before u have a chance to actually learn to play.

I'm not a game designer so i don't know how to implement this, but i believe a system for people to know what they are going up against is really necessary at day one. Cause people will want to dip their toes in the multiplayer scene, and probably early, since the game doesn't have a typical campaign. And if they find that they can't even play without being eviscerated people will put the game away after a few weeks.

A tool for this purpose is IMO crucial to make the multiplayer fun for EVERYONE. I just wanted to give a bit of feedback from one of the "average players". Hope this help you guys at Uber in making the decision.

I'm following what you are saying and am sympathetic. We'll certainly be talking about ways to do this but it doesn't have a very obvious solution.

I don't really like being the guy who sees the obvious but.....

Matchmaking? and displayed ladder ranks? Not sure if I missed something here.

Thanks for the feedback neutrino, I just wanted for you guys at Uber to know the we the average players want this too.
I know it is a tough problem to solve, we the community just want u know that there is problem. And everyone is just trowing ideas at you, trying to help you solve this cause everyone loves this game that you guys are making. But i think I can say that most of us have faith in the work that u guys do. And that we believe that you guys will find a solution.

This is exactly what my concern is. I just don't want to see people leaving because there being destroyed. That's no fun for anybody.

I thought of that as soon as I saw AI games in PA Stats. Trouble is you have really have to record separate scores for each personality and economy multiplier (the multipliers could be bucketed a bit) The AI also changes from version to version - but people change as well ;^)