The only cure for lag is to eradicate it.  That's the point... you can't POLL when the gap of time in between is too great.  There's no way to resolve it, aside from removing time as a component of the auction.  Playstation network players have been suffering it for years.  If there was a software fix to this, it would have been in place a long time ago.  The reality is, only hardware, and removal of latency can actually fix this.  The good news is, it applies as a problem to everyone.
-S

I strongly disagree with this.  The problem is that the host computer is taking the last update that IT got when IT's time ran out, and then awarding the bid to the immediately to this last update. 

What it should be doing is getting a final bid price from each client before determining a winner.  Yes, in theory this could allow for easier client cheating if people hacked software, but that is not the problem.

To be clear, we can't fix the problem of not being up to date on someone else's bid where you think you are ahead but are not.  But if you finish the auction with a bid of $800, your client should submit this final bid, the server should wait until all final bids are received, and then select the winner.   It is silly to see the land awarded to the player who bid $760 when you are staring at your screen with an $800 bid.  This is just an implementation issue.  Put another way, the total time of the auction should be the same for all clients, instead of (host time - ping time).

Happened to me yesterday.  I was in 1st with 3 other human players and there were two plots left; I got the second one.  The 4th placed player got the other plot.  I was not hosting.
http://www.planetmule.com/hi-score-game/game?game_id=17504

But to be fair, this can be considered a bug.  It is annoying to be staring at the screen and then seen that someone who was bidding less than you won the auction.  It doesn't poll the final bid of players properly.  If I see $800 as my final bid, it should not be won by someone who bid $780 just because they have less lag.  This seems fixable to me (compared to other ping-advantaged actions like plot selection which are probably harder to fix).

In v1.1 I believe I once saw the AI settle either 4 or 5 plots in one turn (had been abusing the smithore supply & no mules, plus he bought at auction).  But that was some time ago.

OK, here's an interesting problem with that.

Let's say 3 of our players get involved in the 'run the price up' game.  Or worse, three people bid... but then two run it way up to North Craziland.

Then they all run back down to the 'buy' line, including the poor sucker who stayed down there, as prices went way, way past their ability to buy.

... who buys it? At what price?

The person lowest in the rankings? What if they don't WANT it at this crazy price they couldn't control... and can't even afford it?

The person highest in the rankings? Sure would be easy for a lower-ranked person to force a sale at an insane price, that would screw everyone.

The person who was the last to come down?  What if one of the others got the 'jump' on them, on the way down?

The one with the highest bid, ever during the process?  I suppose that's the least prone to awful.

I think we can design around that.  If the price is overdriven, throughout the auction we keep track of the $HIGHERBID that a player bid compared everyone else.  If everyone attempts to withdraw, that player remains committed to that (minimal) price. If players initially push it up, then withdraw, a player will be stuck at the bid line.  It's a real bid.  Like ebay's autobid feature where you set the max you will pay, but actually it's just one step more than everyone else.  Nothing stops a player from paying more then the $HIGHERBID if they move it up to prevent a last minute drive from someone else. It still allows players to jockey about like they do in early auctions.  I think the AI tends to underbid, but I like the jockeying that they do and we must retain that (and the emotional part of bidding).  If the price is not overdriven, $HIGHERBID is not necessary.

I also really think this should be fixed.

I also have a hard time scheduling this but often I'm looking for a game around 04:00 irata time.

Fortunately, the top 10 is less and less important for the game quality as more players get games under their belts.  I've had success recently beating the top players, then then I think I fell to the #67 or something.  I admit that I was also in some games where a ~#130 or something did some bizarre, useless things throughout the game such that you wondered if they were serious, and it took away from the game.  BigHeadZack, I think understand where you are going but sometimes it is nice to see how you sit compared to the other players in a current system where there is a strong reward just for playing, and playing newbies is equivalent to beating 3 other knowledgeable players.  I personally oscillate between taking solace in a good colony score and just playing FTW.

I have seen this several times as well.  Definitely a bug.

This happened to me when I first tried to host (version 1.0 or 1.1) but I had only opened by TCP port, but not my UDP port.   I'm guessing that you connect with a persistent TCP connection, but game joins and other events like plot selections are done with UDP.  Once opened my UDP port also, I haven't seen that again, or on any other server with someone who has previously successfully hosted a full game.

I think these improvements rock!

The one other improvement I hoped for -- and maybe it won't matter without smithore abuse-- was overdriving land auctions such that no other player could affordably buy them; since you could now pull the price down once the overdriving player withdrew.

The crystite spoilage may be a bit harsh in combination with pirates, but we'll see.   Without the smithore abuse, colony scores will be much more realistic...

The deserts I might have left in but whatever.

Thanks for the awesome recreation.

Camtasia might do what you want.  There is a free trial version.