There is alot of inflation on this site, and I mean an enormous amount, as you can see.

I personally find that a serious issue, but others seem not to care, and kestas actively likes it to continue. I feel it reduces the effectiveness of the points system something considerable, because if you play more games then you get a higher rating... simple as that.

I take issue with that statement, Ghostmaker. :)

When lot's of people try hard to achieve something, many of them will actually achieve it... The points you need to get into the top 100 will reflect new people such as myself joining and doing well.

Don't complain... Embrace the competition and reflect in an Olympian style that your competitors make you better :)

You can get competition just as fierce in a non-inflating system. The fact is that it is hard to gauge what a "good" rating is. This is because of inflation and inherent inaccuracy. Thewonderllama, for example, is hideously underrated.

Requiring double the number of points isn't so bad, the user base has at least doubled since the points system was introduced by just about any metric you choose. It'd be unfair if new players had a higher mountain to climb just because they were new

Also if Rait hadn't played a single game since the points system was first introduced he wouldn't have fallen far from 1st place; it's hardly at Zimbabwean levels, so I'm pretty pleased with how it has performed

id like ghostmaker to put his ELO ranking alongside the help/top 100 based on dip points ranking.

Using the % system as the way of ranking would be a non-inflationary system and would also allow consistency in the ranking. At the moment people invest their points in new games and drop down the ranking, while they may have a very high % score which would have them near the top of that ranking system.

hmm, i've done that twice and gotten 6% and 11% (for different months). does it vary a lot? i didn't post because i assumed i has a mathematical error.

Inflation of the point system is a consequence that new points are created all the time (by new players, or by players going down 100).

On the other hand, using % will be a little difficult... How do you measure "Survive" against "Win", or between them? It is not the same to survive with 1 sc than to survive with 16. And it is better a player that played two games and won both against one that has played 100 and 'only' survived all of them?

Finally, regarding this stuff of going down on the ranking system when you bet, I think that as you do not lose your points until you are defeated in a game, they should not be discounted on the ranking system. They have to be discounted on your personal account (because if not, you would be able to use them several times...), but the ranking should sum up both the points in your account plus those you have at a stake.

Couldn't the points be broken down into Available and Total categories; with the Total being used for rankings?

Ed: that's unrelated.

It's related to the last paragraph of Fidel's post...and who are you, the forum police?

No, I am.

No, Chrispminis is.

If you two are the forum police, is Sicarus the Godfather and mapleleaf a vigilante?

Requiring double the number of points isn't so bad, the user base has at least doubled since the points system was introduced by just about any metric you choose.

>Good point

It'd be unfair if new players had a higher mountain to climb just because they were new

>You've just contradicted your point...

Also if Rait hadn't played a single game since the points system was first introduced he wouldn't have fallen far from 1st place; it's hardly at Zimbabwean levels, so I'm pretty pleased with how it has performed

>Well, Kestas, inflation in the rest of the developed world is bad if it gets to 4%!

>Not to mention that Rait is an outlier


Basically, I'm all for points, we need to keep n00bs out of our games. I just don't feel that having a ranking sytem by points actually ranks anyone well. It Rait was to join a 7200 bet game he would be below me in the 'rankings'. I think you should keep the points, but implement ELO to 'rank; players and make the top 100 list.

that's a common suggestion.

>>"inflation in the rest of the developed world is bad if it gets to 4%!"
so if someone set up another phpDip server and no one played there, it could theoretically be called "Switzerlandesque"?

Requiring double the number of points isn't so bad, the user base has at least doubled since the points system was introduced by just about any metric you choose.

Ah, but it is not a flat thing you see. There is not an equal number of players from 100-200 as from 800-900. If there was an equal spread of players accross every point level, that would be fine, but we have a pyramid style system (if it weren't for rebates, that would be a negative skew) which means that to double the point number to be in the top 100 you should have to multiply the total populance by three or four fold.

no it doesn't. as the new players join some old ones move upward.

No you don't....

Since the ranking goes from 1 to X (where X is the newest player), the new player get stuck onto the end.

i meant point-wise, not by rank. obviously the average RANK is always getting lower.

Well, I don't have any issue with the points inflation that's happening really. In my mind, that's not a real problem with the points system. The real problem, which is inherent in all systems with are based highly on blunt win percentage, is that you will get more points for playing weaker players, whereas you will lose points for playing stronger players. Despite that one player may be more skilled than another, the points will not reflect this because one might hang with the top guns and mostly lose, and the other will tackle CD-ridden newbie games they will likely win.

I resent being the forum police. I prefer being the forum "Smoky the Bear". Be careful with matches children!

2 observations here, quite separate so I 'll put them in different posts:

1: The fact that your points go down when you invest them in games.

This is easy to fix: simply calculate the expectation for all the games you are playing and add this to your score. Expectation is a probability term - in a winner takes all game, it would mean the % chance of you winning multiplied by the pot. Calculating the probability of you winning is of course not simple, but for the purposes of your points being accurate when you have points invested in ongoing games we only need a 1st approximation:

For any ongoing game,

your Expectation = Current Pot x Num of SCs you have / Num of SCs owned by all active (non CD) players (22 at beginning, 34 in a mature game with no CDs).

So, to stop everyones points (and thus ranking) fluctuating as they bet on games, Simply calculate the expectation for everyone's games and add them to their points.

Obviously they can't bet points that are invested in game, so you are shown 2 points balances: Your total points, and your available balance (just like your bank account)

The more astute among you will have realized that being allocated Russia when a game starts makes you points go up slightly whereas any other country makes your points go down, but that's life.

2. Inflation.

The main reason that there is inflation is that more "money" is being created (added to the phpDip economy) all the time, as new players join and players with no points are topped up by phpDip "Social Security". Where does this money come from? Answer: from devaluation of the currency, causing inflation.

The only way to stop inflation is to raise taxes, to pay for new players to join.

Do you tax the rich (eg rait) just because they have money? This mean if they don't play any games their ranking would drop as new players are added.

Or do you tax money made (earnt) by winning games?

Personally if the inflation bothered me, I would favour taxing the pots won. I don't think I would tax at a straight % of the win, rather I would bias the tax to the winner, at the expense of the mere survivors.

Example: if the chancellor sets the target inflation rate to 0%, then no new money can enter the game. If 5000 points are dished out to new players and "dole spongers" in a particular month, we need to raise 5000 in taxes to pay for this. If games worth 50,000 finish in a particular month, then we need to tax the winnings at 10% (5k/50k).

This would mean a straight 10% off the win of a winner takes all game - simple. For PPSC games, you could tax the same way, but personally I would bias it towards the winner thus:

Each SC is worth 2.94%. Starting with the player with fewest SCs, remove one SC's worth of taxes from each until you have payed the required amount. (the last player to pay pays whatever fraction of an SC is required. Unless there are lots of survivors and the tax rate is low, the winner will pay his share of the taxes (with a 10% tax rate in a 2 player game won 18-16, the winner would pay 4.12%, 2nd place 5.88%).

If you were to introduce this system, it would be sensible to have a one-off tax on everyone's points and put this in the treasury, to subsidize the the system for a year or 2. Otherwise it is unfair on new players because the existing players earnt their points with no taxes.

and finally: If you hate the idea of tax, don't moan about inflation, moan about immigration (new players) and dole spongers (crap players) - it is the supply of points to them that causes the inflation.

Have I bored everyone to death?

No, I have actually had that idea before.

i actually think rankings shouldn't be based on points, but by using the win rate percentages and then if tied on that using survival rate and number of games to break it.

Thank you...

Neville, I like your idea. I have not programmed in php, and I do not have time to learn right now, but it is very interesting to experiment that. (Are you an economist? Because it is an economist's solution... ;-)).
FF

I actually joked about adding tax to the system to make a point about how things can get overly complicated. The scary thing even when joking about it you can see the logic in it, but there is a worry about it getting too complex

Hmm... seems to me that the only worry would be in the time required to implement.

I, myself, had the thought of a tax system that would tax the top X players 1 point each to reach the amount of points needed. The max will be 100 points (ie when there's a new player), and the top 100 will all lose a point (the cost of 'fame').

Personally, I think that an 'accounts' page would be neat, which will produce a statement of where a player's points went (and their tax rate).

Of course, as we are under the sole rule of the benevolent kestas, we're not going to do much talking about it.

Then again, perhaps a system of government would be an idea to implement, with Rait as the pseudo-monarch, with the top 12 forming the Elder Council, the top 25 the Elite Chamber, and the top 100 the General assembly.

Or maybe I just need some sleep...

Hmm... regardless, I've been inspired by my notion of government, I might develop something to try out PHP...

@kestas: Yes I was joking when I started the inflation post (hence the "if I was bothered"), but by the end I was beginning to persuade myself of the logic! :-0

Having thought about it a little, if it were to be implemented, I favour ongoing tax on people's balances slightly as well as taxing the pots. The reason being that with no new money coming into the system, players with high points scores could just sit on their winnings and keep the top spots, despite never playing any of the new hot shot players. Taxing the balances would force all players to keep winning games in order to keep their spot. (and encourage people to play more, which is what we want).

I also quite like the idea of not giving players who go CD the full 100 when they run out of money. Out of interest, how many new players / point resets are there per week, on average?

@Churchill. I think if we're going to discuss government we should start a new thread.
I do like your idea of the system publishing accounts each month (Inlation target, Pots won, Points handed out to new players / to zero pointers, tax rate (determined from the previous accounting period), tax raised, budget deficit, balance (if we raised one off tax at start of system) at start and end of period, tax rate for next accounting period. You could also display summaries of pots won / tax paid for the players, so you can see who is up-and-coming.
Personally I don't think inflation is an entirely bad thing. As the number of players increase, the average standard of play increases, so should the number of points needed for top spot.

But raising enough taxes to keep the inflation down to the rate of increase of active players would be a good idea. (active players = number of unique players in an active game).

Oh dear, this has turned into another long post

@Churchill - if you want to collaberate on a php government game/simulation, I'd enjoy that. If you like the economics of it, look up Monte-Carlo engine/simulator.
Personally I'm thinking of implementing an online version of Avalon Hill's Republic of Rome, which I've never succesfully played face-2-face but would be great fun for the sort of player we get here on phpDip.

@fidel: No I'm hang-gliding instructor / PHP developer by trade
No offense taken (I share opinion about economists with Naseem Taleb - author, Fooled By Randomness)

Oh and one final refinement - make sure that the points of inactive players are recycled back into the system (taxing balances would do this, but too slowly for sub-100 point players). I suggest a higher balance tax rate for players with no current games (after a 2 week holiday period)

HAHAHAHA!!!!
No offense intented. It was just a joke. I will try to read about the work you mentioned.
But, going back to taxes, I like the idea.
ELO may be also good, but it is definitely much more complex (and perhaps good for a parallel ranking system...)

Wow, Neville... It'd be great. I've never actually done anything in PHP, but it would be awesome to work with you and learn (I've picked up other languages fairly quickly on my own).

Where can I learn "Republic of Rome" (I know, I should google it, but I'm busy right now... ;-). I do not know php -- well, too much to be confortable -- but I would also like to be in touch with the team developing it...
(Perhaps games are a good motivation for learning a new language...)

http://www.boardgamegeek.com/game/1513
I have a mint condition copy that I never played, because I can't find players for it, but when I have played it before, every time the empire was destroyed by barbarians because the players backstabbed each other too much.

Thanks, Neville!
Regarding the topic of this thread, inflation, there is a game it was shown in other thread that is an example of why there is so much inflation.

http://phpdiplomacy.net/profile.php?userID=4393

Here one single moron (I assume he is a multi) has the ability to create 1545 points with no effort. That is, our points go down to pay for these points. And also his "main" account has more points than me, that played hard for them...
Silly.

Taxing just means rather than having the same points but are worth less, you have fewer points that are worth the same. No difference, just a division.

Sure, but at least one may know where the points are going...

Is there no way to exclude this kind of jerks from the system? For example, if there is a division between untrusted and trusted accounts. You start as untrusted, and only promoted to trusted if certain requirements are met (like maximum number of CDs, some proofs that the account corresponds to a real player and not just to another personality of a multi, etc.)
In that way you start playing to prove yourself valuable member, and then promote to the *real* games... ;-)
The "untrusted" part would be similar to the present one, but all the people that really plays for fun can have a separate arena. And in that part there would be no multis or other jerks like this "jesus".

I'm kind of a socialist, so taxing the rich to give to the poor is something I am, in theory, very in favor of.

However, in this case, I believe there is just no need.

To answer to the last first, taxing does not necesarily mean "fewer points that are worth more", because it depends on whose points you take.

I think that keeping track of a player's CDs, and lower their minimum points (FE: 5 points less for every time you went on CD) would be a good idea wich might help a bit with at least one of your troubles.

And, if I were to tax something it would be the points of those who have many but do not play. A ratio of points betted/points total should at every time be, say, 20%, and if you don't comply, you are taxed, say, 10% of the difference (the difference being "the extra points you should have betted into a game to reach the ratio". This could be run twice a week, without notice.

Penalizing those going CD is a good idea. It is annoying to see how a game is totally unbalance because the neighbour of your opponent is letting him take his sc's with no effort...

One way to get rid of inflation is not to issue any points to new players. Joining a game would still cost (players would be allowed to go into the negative) and the points in a game are always positive. Parameters could be established by players to define the minimum (and presumably the maximum) strength permitted to join a game. The universal average of all players across all games, regardless of how many players join over time, will continue to be zero. That would rid the problem of inflation. It would not though necessarily do a very good job of ranking individuals. And someone with a tally of 100 would not necessarily be twice as good as someone with a tally of 50. And how would one compare someone with a tally of 50 to someone with a tally of -50? You would know off the bat that the individual with 50 points had won a number of games while the individual with -50 had lost a number of games, but it would give you now indication of how many.
It would be possible to integrate the games played, won, survived, and lost into the ranking by displaying both at the same time.
For example: 50 - 75 (40;30;20;10)
Total points accumulated 50
Total games played 75
Total games won 40%
Total games survivied 30%
Total games lost 20%
Total games abandoned 10%

I would also suggest that there be a penalty imposed for abandoning a game, and that this penalty grow progressively if games are repeatedly abandoned.

Whether this system has merit, I can hardly say, but it will definitely exorcise inflation from the scores.