Rhysem

A market anywhere is just another form of PvP. Nice to see devs who 'get it' though.

I find my memory for trend analysis ... could use some work. That's why I want to feed it to the PC.

You know, herostats could answer the inflation question at a basic level. It already tracks inf in. Add the market transactions in the chat window log tracking and that should give you what you need.

It just won't let you quantify what is getting more costly, only if the inf prices on the AH are rising over time.

I had to do some image editing (because it doesn't understand a full screenshot -- I did try!), but it seems to have no trouble pulling correct text out of a screenshot of the market I took. That's the <Sold> tab. Unfortunately it didn't pull out the IO name, just what I listed it for. Switching my font might help...

How much work that'd be to bundle a working OCR engine into a python app to screen-scrape is another matter. I at least did find other people doing screen grabs in python under windows, so that part shouldn't have to be manual (though if it was, watching the directory of screen captures would be easy enough).

That's pretty much your only option. The market is so fast during periods no bot could catch all the trades. Besides, the last-5 is already compressing out data, like bid a stack of 10 and have that show up only once for salvage or recipes. My hope would be to catch something resembling a representative sample.

For non-salvage stuff, the sales are slow enough that a few people checking it would give you passable data. The higher end sets people use aren't pushing that high of transaction volume -- you do see a day or two ago's sales on a lot of them.

Some market data is better than none, which is pretty much what "last 5" is for doing any sort of analysis or debates about the market, the state of it, etc. Also, even a weak snapshot like that would let you start trending for "are prices rising? by how much?" and similar.

I've written a bot that was within the EULA and ToS of a game. Also, what's being discussed isn't a bot. Try again.

Shrug. I'm a system administrator / programmer by trade. I automate away any boring task I can.

Doesn't make any sense until you're saving the things from going on the market in order to barter with someone else who has something you want where the market collapsed for that IO. At the point you're saving them and have a list and what not so you can trade... saving an extra 10 of 100 isn't a big chunk of extra work, you already did most of it saving the first dozen or so.

You can't make the logical leaps you are between the list of data collected by under 1% of the player base inputting what they list and sold for and an all-knowning database that can tell you the exact minimum you need to bid to get an IO down to the last 1 inf. Not that, as grouchy points out, it is clear we can get that data.

Put another way: those not participating in collecting data mean the market keeps its qualities despite more people having data.

All you can say about the data we're talking about collecting is it gives you a better average and standard deviation. Nothing exact by any means.

Linking to one of the other market threads...

http://coh103.gtm.cityofheroes.com/s...=226514&page=3

The screenshot about half way down. The "you put" + "5% listing fee" and the "you got" + "% final cut" should be sufficient to track it down.

No, you're right actually there's going to end up being too many assumptions you'd have to make there. Since you don't know that they priced things in the same order they put them on the AH in.

I suppose the other option is doing something akin to fraps, but instead of saving the image doing OCR against it for the sold tab being highlighted then recording what is seen there. Messy though.

Then again at that point you may as well just OCR for the last-5-sales prices and be done with it. Still will have to deal with redundant/duplicate data, especially on high-volume movers.

Edit: I tried a quick test with tesseract-ocr but failed to get it to do anything with the low-rez screenshots. I'll try tonight at home when I can make my own high-res screenshots and see if it has more luck there.

This had dawned on me as a huge issue. My plan had been to make the client open-source -- preferably a scripting language. Since I'd been thinking I might host on AppEngine, that'd lead to Python as the language of choice.

Then it is easy to inspect that it is in fact only sending the sold data up to the server. Of course, you'd end up having to give it your google account credentials to send the data to appengine (gotta have a way to ban people who feed bogus data -- because some folks get their lulz from that). Some folks may not like that, so it'd be easy enough to implement an "upload file" web page too on appengine, but then that has the ability to pull out posting prices too, which is less good.

Ultimately it comes down to trust of the folks writing/running it.

Nobody's suggesting the latter. That'd be against the EULA/ToS because you'd have to crack their network data stream, and then write a bot to sit there and scan every page of the market (aka WoW's Auctioneer).

I'm talking about looking for the correct pairs of "You receive XYZ from the market. You paid XYZ to the market." that are generated as you collect sold items in your chat logs.

This is different from me, a self-confirmed marketeer, vs the unwashed masses how? I've already got a lot better pricing picture in my head even if it isn't computer-perfect-digital-accuracy. They don't. Honestly, this sort of idea actually hurts marketeers more than the unwashed masses because it levels the playing field between us some.

But it gives us good data to argue about, and I think a lot of marketeers are actually probably more interested in that than any slight loss-of-advantage over the masses it'd incur us.

Market sale prices isn't private info. I'm not talking about trying to crack their network protocol and pull any and all data your client fetches on the market. I'm talking about my, yours, or anyone else voluntarily feeding their sale data into a (public) database.

To be honest, you should be assuming this is already being done. I'd bet the gold farming shops are already paying someone the pittance they do to sit there and collect this data on the market. We already know from others who have posted things like 1 to 1 billion inf in a week RL that you can make quite some cash at the market if you even casually try. Having someone paid to run the market for 10 hours a day at a gold farming company... I'm sure they could do a lot better than 1 billion per week -- maybe faster than actually farming maps.

I had a previous "shared data" plan which was a lot less data than the market and a lot more use, personally, to me, which I can tell you'd like even less.

Correct.

UberGuy, the problem is multi-fold. What I'm looking at when I worry we're getting too close to collapse is something like this: There is currently something of a stigma against off-market trades. Mostly because it is inconvenient compared to just using the market. I'd rather buy there where I can get it NAO! But as trades start moving off the market, if I can't get the PvP IO I want on market, then I have to move off market (or do a merits, but the math doesn't work there -- you're (probably) better off with an off-market trade and selling the other results of a merits). At some point here you hit the first tipping point and people start lists of what they have and what they want and start looking for trades. Some will take inf to start with. Over time they'll quit and do barter-only, or barter-plus-inf with most of the value in the barter half.

The problem is this effect is self-reinforcing. One guy on the forums with one PvP IO and some greed the AH can't represent as a number? No big deal. Five PvP IOs within a day? That's a little more worrysome -- that's a faster exchange of them than the market is managing. So there's over billion inf that isn't getting destroyed to fees. A billion compared to Fulmen's estimate of inf reserves? A pittance... but then again how many mobs do you have to kill to make a billion?

As that happens more and more, suddenly everyone starts doing it and the market dries up. Low-value stuff (salvage, trash/niche recipes) will still be there... but the big stuff (purples first, other >100 mil IOs shortly after) will quickly filter off. One piece of a set pulling out of the market is going to try to drag the rest with it -- because if I have a gambler6 and keep other gamblers in my inventory, I might be able to trade the whole package up to say, a miracle6 someone else has.

So in short: its a failure mode like an avalanche (or even our favorite: cascading defense failure). The first failure isn't ... so bad. The next one, okay it gets a fair bit worse. Let the failed points reach 5% of what people want and suddenly it gets worse very quickly. I couldn't tell you where the "halfway/median" point is, but I'd guess it is under 25%. Maybe less. Maybe under 15%.

You can say, "but that's why we have limited storage!" To which I respond, "have you seen the crazy 88s?" If you know what they did, then the concept of limited storage falls apart. Have your trade mule (yep, I said it!) with their own private SG. Buy up enough of a base to fill it with invention storage. How many IOs is that? Something around 1800? Yes, that's /very/ limiting storage!

With gleemail you don't even have to have that storage on the server you play on, so you're not taking up a precious character slot on your play server. In fact I suspect most storage bases of that form would end up on virtue or freedom, simply because they're high pop. Then we'll have a barter server(s), not an auction house.

Plus the private base lets you do nice things, like letting people browse your wares by pulling an alt (purchasing mule) of theirs into your SG with permissions to look-not-take. Now they can browse and you don't have to do anything but invite the mule temporarily (or maybe even just a team invite? Not sure there.). You might even find players trying to run bots to do the above. I've seen it in other games before... Not pretty.

At 3 slots, I'd be amazed to find anyone who isn't slotting gamblers instead. Def/global rech, def, def/end and you're not at ED cap, but you're well enough into it the last fractional percents you gain isn't worth another slot. Low on end redux, but that's what your alpha slot or other power picks are for.

I'd specifically not mentioned exploiting what is not WAI.

Given I find I'm generally slot starved, the bonuses to a purple set of res/def/etc would need to be HUGE to get me to devote slots to them. Not saying it couldn't be done, but it'd probably be overpowered if it was.

Shrug. If they're that good, they would have been better served reclassifying steadfast into being rarer than it is. The wildly differing dev valuation doesn't sit well with me. Maybe its my MMO history, but the assumption that "oh this will stay rare because <insert reason> so it is okay if it is a little overpowered" always, always, always breaks down sooner or later and then suddenly Captain Nerfbat comes out swinging. We're now at that point with people being able to buy the +3% via hero merits. Sooner or later it isn't going to be rare. Next thing you know we'll be seeing Captain Nerfbat out.

It probably wouldn't bother me so much if I didn't have a def-based character who's going to suck the fallout of that (a SD -- who given their fairly crappy resists ought to be as easy to softcap defense as SR is, but can't).

Though more general I think the design of the IO system is ... lacking. There is symmetry that either needs fixing there's good cold/neg/psi resist stacking... and nothing for other types. Or outright strange things -- nobody stacks for resists because you're targetting against 100% resist, not 50% defense, yet you can pick up higher +def bonuses than +resist bonuses (particularly evident with two +def global IOs and only one +res global IO). I'd rather see more quantity of smaller bonuses per set, and lose the trash bonuses. +1.5% reduction in mez duration -- is that even larger than a server tic so it might matter?

I'm not saying everything should just be a copy of each other so that nothing is different, but I do think a more ... fair? even? distribution could have been made.

False. Assumes market prices can go arbitrarily high as opposed to being capped at 2 billion per transaction. I'm not convinced that skimming off 200 million per transaction of the high-end IOs is going to be enough.

Also, DO NOT WANT. Any market that hits that point on recipe prices is going to suddenly collapse in on itself in a black hole. At that point you can't make any profit (in fact, you make a loss much above 1.79 billion) crafting recipes into IOs to resell on the market. So now the supply of IOs drys up and the casual-crafting folks who don't want to craft are now excluded from the market. That further reduces market transactions which drops the amount skimmed off the top by the 10% fee.

Not to mention that long before Mako's hits 2 billion you're going to see the PvP IO and other big hitters pushed off the market into a in-game/forums, barter-only market. We already see this with the PvP +3% -- currently for cash over 2 billion on the forums, but if the market (read: purchasing power of an arbitrary quantity of inf) starts to collapse its going to turn barter only. DO. NOT. WANT. I've played other games where the trading market was interactively-player-ran and barter-only and frankly I'd much, much, much rather have an auction house. This is City of Heroes not City of Day Traders Standing Around Hawking Their Wares.

While I'm sure Fulmen's escrow service would love that, he might get sick of running it after a few days. :-P

No, you missed my point. If Rylas point about "well the bennie is huge it should be hard to get" stands, then why does Steadfast Protection stand where it is? Sure, the PvP IO is "better" because it takes you to 6%, but the same argument can be made about Steadfast if you already had the PvP IO first (maybe you PvP and got lucky and decided to IO yourself out then?).

It doesn't really matter which you got first -- you need both to get to 6%. One is classified as rare. One is merely uncommon. But somehow I'm supposed to believe that because one is "rare" that its okay for it to be so hard to get, yet its okay for steadfast to be easy. I question this assumption.

...

That leads me to a side question: if I place a second 2 billion bid for the PvP recipe (I've had one sitting for 3 weeks, no dice), the market will "combine" them since they're identical. Can I cancel my bids? It'd try to hand me 4 billion inf ... which is a no-no. Will it be smart and cancel just one, or would that become sunk money I'd never see again? Similar with selling two of said recipe (or, maybe a bit more realistic, say 10 purple recipes)?