FHM (Franchise Hockey Manager)

[MizzouRah]

Originally Posted by johnnyshaka

Sounds like there are still a few rather annoying issues like getting fired over and over again during a single season even when "Can't Be Fired" is selected. Also, hiring head coaches was setup but apparently wasn't supposed to be implemented for this release but was enabled for whatever reason so that messes things up, too.

I don't think I'm going to invest any time in this patch either but I'll continue to scan the forums to see how they plan to deal with the latest issues.

Dammit.. thanks though..

[Honolulu_Blue]

Originally Posted by johnnyshaka

Sounds like there are still a few rather annoying issues like getting fired over and over again during a single season even when "Can't Be Fired" is selected. Also, hiring head coaches was setup but apparently wasn't supposed to be implemented for this release but was enabled for whatever reason so that messes things up, too.

I don't think I'm going to invest any time in this patch either but I'll continue to scan the forums to see how they plan to deal with the latest issues.

Keep scanning the forums and keep us posted!

[DaddyTorgo]

In the meantime we can blame all the hockey-loving Scandinavian pirates who pirated SI's game instead of paying for it.

Cheap-ass socialist motherfuckers.

[jbergey22]

Havent tested this but that is one impressive looking patch. I have been hard on them two and this release but I think they are doing everything possible to make things right. Just seeing the work they put into this new patch changed my attitude from a negative to a positive outlook on the future of this series.

[jbergey22]

Originally Posted by DaddyTorgo

In the meantime we can blame all the hockey-loving Scandinavian pirates who pirated SI's game instead of paying for it.

Cheap-ass socialist motherfuckers.

LOL, I wont get into the name calling of society's. But the people that did that, fucked it up for the rest of us.

[DaddyTorgo]

Originally Posted by jbergey22

LOL, I wont get into the name calling of society's. But the people that did that, fucked it up for the rest of us.

I was kidding around. I actually love socialist Europeans.

[Johnny93g]

The UI was the main reason i shelved the game and eventually asked for my money back. I felt it lacked a common sense touch to it. Click fest, a chore to do anything, no real direction on what you needed to do.

So i went over to OOTP today, and there's a long thread on UI suggestions. Got my hopes up.

Reading through the thread, seems like the developer, who to his credit, is reading and responding, doesn't seem that interested in making many changes to how the UI plays. Shame. I'm the game's target market. Can't believe I have no interest in this game as it stands....

[DaddyTorgo]

Originally Posted by Johnny93g

The UI was the main reason i shelved the game and eventually asked for my money back. I felt it lacked a common sense touch to it. Click fest, a chore to do anything, no real direction on what you needed to do.

So i went over to OOTP today, and there's a long thread on UI suggestions. Got my hopes up.

Reading through the thread, seems like the developer, who to his credit, is reading and responding, doesn't seem that interested in making many changes to how the UI plays. Shame. I'm the game's target market. Can't believe I have no interest in this game as it stands....

[Marc Vaughan]

Originally Posted by Johnny93g

Reading through the thread, seems like the developer, who to his credit, is reading and responding, doesn't seem that interested in making many changes to how the UI plays. Shame. I'm the game's target market. Can't believe I have no interest in this game as it stands....

(Disclaimer - I haven't read the thread you're talking about and have only casually played FHM)

One thing to bear in mind with games is that often UI rework is left to future versions in the main - this is because a 'coherant/familiar' UI is sometimes better than an enhanced but disjoined/unfamiliar one ... on my own products I'll only generally do minor UI enhancements in updates, leaving revamps to the next full version to avoid alienating people who are 'used' to the current version.

[johnnyshaka]

A new patch (1.6.8) is out:

Version 1.6.8 Game Update Now Available - OOTP Developments Forums

Our latest post-release game update, Version 1.6.8, is now available.

You can download the new version at:

(PC) FHM - Update 1.6.8
(Mac) FHM - Update 1.6.8

PC Installation instructions: After you download the file just install it over your existing game (you don't need to uninstall the old version first.) Make sure to select the correct path to your game if needed.
Mac Installation instructions: After you download the .dmg file, open it, move the fhm.app file to your applications folder and copy over the old version.

After installation, the game should display "Version 1.6.8" at the bottom of the statup screen; if it shows a different build number, you'll need to re-download and reinstall (and before doing so, make sure the update file you downloaded has been deleted and your browser cache cleared.)

Note that if you decide to uninstall and reinstall, you should first release one of your licenses via Preferences-License on the start screen.

Saves from earlier versions are compatible with this version, but, as always, to get the best results from the patch it's recommended you start a new game. If you continue an earlier save and have problems, please let us know that it's a pre-1.6.8 save when you report the issue.

Introduction:

This is the followup patch to 1.6.2, fixing various issues noted with that one and correcting some other long-standing problems.

Free Agency, particularly in the NHL, got a lot of attention. Restricted Free Agents should now be handled properly, and the AI teams will now attempt offer sheets (and match them) at a realistic rate. Contract demands have been lowered for unproven players, which should also help teams' ability to retain their good young players. And players should now be much more willing to resume negotiations after breaking them off - for example, players given a minimum qualifying offer on June 30 in order to retain their rights will still become upset and refuse to negotiate for a while, but that period will end within a month. I still don't think the AI is doing the best possible job of handling free agents, particularly its own RFA's, but now that we've got the mechanics of free agency functioning properly, we can make adjustments to its behaviour and know that the rules will be applied properly.

The option to hire a head coach has been removed - this was the source of a large number of issues people were having with the AI mysteriously making decisions for them even while it was turned off. While we want to eventually offer the option to hire a coach and play only as a GM, it needs to be implemented properly and not in the incomplete, confusing way it was working in 1.6.2.

Barring any emergency fixes, this will be the last update for a while, until the match engine patch is ready. That patch will include major changes to the in-game view, tactics, and other related systems, as well as a few other new additions and changes we'll announce when we're a little closer to the release time.

Bug Fixes and Other Changes

Transactions, Contracts, and Rosters:
-No longer possible to hire a head coach - for now, the user is always Coach/GM.
-March 5 quickstart cap set to $71.1m (next year's value) to minimize the amount of roster adjustments needed.
-Contract demands for younger players, particularly those on second contracts, and older ones with limited experience at the level of the team they're negotiating with should now be significantly reduced.
-Players will be willing to resume broken-off negotiations much sooner now.
-Coach and GM firing rates re-tuned.
-Multiple coach firings in a season for one team should now be extremely rare.
-Late-season coach and GM firings now less likely.
-Rights to qualified RFA's should now be retained properly.
-Offer sheets can now be made to RFA's normally.
-AI teams will now attempt to sign RFA's to offer sheets, and get the option to match offer sheets to their own players (they tend to match very aggressively, as they do in real life, so successful offer sheets won't happen often.)
-Sudden changes in contract value on July 1 in certain leagues fixed.
-AI adjusted to place a higher value on draft picks.
-AI will now attempt to hold a contract space in reserve to allow more flexibility while trading.
-Users should no longer get GM job offers from leagues they're currently employed in.
-Abandoned contract negotiations in major junior should no longer result in 1-year, $0 contracts.
-NHL teams should have a slightly-increased preference for signing North American players.
-Emergency tryout contracts should tend to go to players from the same region/state as the team.
-Contract offers should no longer occasionally reappear after being removed by the player.
-Players should no longer be willing to accept contracts longer than 3 years with salaries below the league average level.
-Transfer deadline restrictions should now reset correctly for nonplayable leagues with post-January 1 deadlines.
-"Minor option" checkbox should be functional in negotiation screens if the rule is enabled for that league.
-Should no longer be possible on draft day to trade picks that have already been used.
-Entry-level contracts in leagues that allow 2-way contracts must now be 2-way.
-User should now be able to sign staff from lower-level leagues during the offseason.

Player and Staff Ratings, Development, and Performance:
-Player ability should now update properly on user-controlled teams.
-Limited the effect of scouting error on displayed player attributes.
-Players should no longer change uniform numbers while on the injured list.
-Determination and Leadership should no longer increase to high levels for almost all veteran players.
-Altered ranges for hidden personality descriptions in scouting reports, too many guys with neutral numbers were having those described as weaknesses.

League/Mode-Specific Issues:
-Crash with certain custom mode league setups fixed.
-Scouts should now work properly in historical mode.
-Various errors in handling of overage players in major junior fixed.
-In leagues with salary caps, contracts of players assigned to that team by a parent club should have a cap hit (to the team they've been loaned to) equal to the league minimum salary (in that league), instead of the $0 cap hit they currently get (mainly affects the ECHL.)
-Finnish junior leagues should now have no roster limit on 20-year-old players.
-CHL teams should no longer sign players with the college preference.
-OHL league rules now specify the correct number of overagers.
-Historical games started in years where the schedule started after January 1 should now begin at the correct date.
-UK teams should no longer be able to make transfer offers within their own league.
-Current NHL playoff system fixed so the two divisional leaders are seeded correctly.
-Player salary demands in historical games should now take into account entry-level restrictions.

News, History, and Related Issues:
-Rookies in historic mode will no longer be draft-ineligible, leading to empty drafts after 2-3 years.
-Draft reports now appear a couple of additional times a year.
-Draft reports should be "written" by the team's scout with the highest draft potential rating.
-Sorting protected list while in the process of setting it should no longer copy players to the list.
-Correct arena names should be displayed in box scores and game reports for arena mode.
-Draft reports removed from historical mode (they wouldn't work properly because of the way the drafts are scheduled in that mode.)
-Certain team names should no longer display as blank in "Last 4" and Career Stats.
-Owner season expectations should now be more realistic in historical mode.
-Maximum minor league salary on 2-way entry-level NHL deals now set to $70,000.
-Goals Against leaders in league home screen should now display correctly.

Gameplay:
-Crash when attempting to view playoff game reports fixed.
-If a player with a No-movement clause somehow winds up on a farm team (e.g. via editing), he should no longer be prevented from being called up by the clause.
-Players acquired on waivers should get scouted to A-level as soon as they join the new team.
-Random name files updated and expanded.
-Crash when trying to view staff tab of inactive teams fixed.
-User's GM contract should be automatically renewed (if you want to leave the team, use the resign option on the user menu.)
-Maximum editable injury frequency increased to 3.0 from 1.4.
-A few situation-specific crashes fixed.
-Debugging-related dumping of CSV reports turned off.
-"Strengths" description removed from evaluation of trading partner (it wasn't that useful, and wasn't showing anything most of the time anyhow.)
-When scouting is delegated to the AI assistant, its choice of targets shouldn't become overly focused on the local region after the first assignments end.

Graphical/UI:
-Sort order for PP and PK% in league stats fixed.
-Various changes to the way the game retains sort orders in a number of different screens.
-Team/league/game names on the upper bar of the screen should now link to the appropriate screen.
-Ties should now be displayed properly on the autosim screen in leagues where they're used.
-Right-click option to offer contract should always appear for eligible players now (i.e., no longer necessary to open the player screen to check if it works there.)
-Incorrect total for shooting percentage on split section of player screen fixed.
-New Draft Info view available (and set as default) during draft with more draft-relevant categories (height, weight, etc.)
-Certain weaknesses should no longer be repeated in scouting report.
-Current role description in scouting report should now properly describe role of player in league he is currently in.
-Should now be possible, after changing views, to go directly from looking at one of a team's farm teams to another.
-Various small typos and grammatical errors fixed.
-Players put on reserve list of farm teams should no longer become invisible to the user.
-Error in formula for playmaking rating fixed.
-A few unused filter names ("Minor Awards", etc.) removed from the news filter list.
-Attribute sorting on Ratings views of roster screen should now work properly.
-After looking at a multiple-tab table (e.g. free agents) and then adding a filter that reduces the number of entries to a number lower than the current tab, the table should automatically revent to the 1-100 tab.

[DaddyTorgo]

That's a ton of fixes - curious to hear how it plays now.

[Solecismic]

Congratulations on making Steam.

[DaddyTorgo]

Any news on how it plays now?

Jim - how's the FOF Steam campaign??

[Solecismic]

We can talk about that in the Greenlight item. Don't want to t/j FHM, especially on its big day.

[DaddyTorgo]

I'm curious how FHM plays now - especially now that it's on Steam with all those fixes. Any impressions from those here that have purchased?

[vex]

Is there a demo?

[cubboyroy1826]

Just figured I would ask as well now that I see the game is just $19.99 on Steam. Worth it or still too many issues?

[EagleFan]

Move the decimal one more place to the left and you may have the worth.

[Young Drachma]

I mean, it's the first gen of the game. No big shock. I figure they can work the kinks out and get it better, but should you pay for it now? No, I'd wait until the next version. You probably won't get much enjoyment out of it now.

[PilotMan]

Gonna bump this:

Gamersgate has this for $5 for another couple days.

Franchise Hockey Manager - Buy and download on GamersGate

[EagleFan]

Still over priced...

[DaddyTorgo]

Originally Posted by EagleFan

Still over priced...

Yeah?

[PilotMan]

I picked it up, but my expectations are low and I haven't played any hockey game seriously since SOM 20 years ago.

[Grover]

Still over-priced and still broken as hell.

[MizzouRah]

I've been using NHL 15 last gen for my hockey simulator, it's actually pretty decent with sim stats and player progression. I don't play the games like I do with NHL 15 current gen.

There is a new version of FHM in the works though.. so I'll keep checking back!

[Honolulu_Blue]

I bought this months ago and I just now remembered that I own it. And I'm "Hockey Boy" for fuck's sake!

That's sad.

[Johnny93g]

FHM2 has come out, and i was taking a look over at the ootp forums.

Saw these trades posted.

Trade with Dallas
- To Detroit: Tyler Seguin
- To Dallas: Pavel Datsyuk, Zach Nastasiuk, 2nd and 7th (2016)

Trade with Tampa
- To Detroit: Victor Hedman
- To Tampa: Johan Franzen, 1st and 2nd (2018), 2nd and 3rd (2017)

Detroit was the human team. That's pretty bad i think. Trading older declining players for young franchise level players that easily.

A cpu trade has Edmonton trading the Hopkins to Florida for Pirri and prospect Mattheson.

How does that make any sense?

I haven't bought the game, but stuff like this makes it easy for me not to bother at this point.

I understand trade AI isn't easy, but to me, that's broken.

[Mike Lowe]

So join an online league! We start next week!

[jbergey22]

I received a couple of emails about this game and was thinking about purchasing. Problem is that I still have bitterness about how they handled the first one. A lot of broken promises. From the things I have read it is much improved but I did read that the trade AI isnt very good. I am sure I will end up getting this once I read a few more positive reviews.

[Johnny93g]

Originally Posted by Deepak25

The games great so far but what ive noticed is that on the second day of free agency teams are signing players and trading them right away

Saw that posted over on the FHM2 impressions thread.

If that's true, that's a game killer for me.

[NoSkillz]

I didn't get the original FHM but I think I'll bite the bullet and buy this tonight when I get home, if only because they've made it available for the Mac. I'll post thoughts if I get the chance to fiddle with it tonight.

I'm a die-hard supporter of Riz and EHM and bought the recent EHM-EA version but since it's not available for Mac at the moment, I can't play it.

[SlyBelle1]

I haven't played around with the new version much yet, but always seems trading logic in most games are criticized. I wonder if some of it is because we react to the actual player names, versus the ratings and game situation within the game itself, which may not mirror real life. Now that may mean the player's stats/ratings need to be adjusted, don't know. If this was a fictional universe and you really looked at all aspects within the game itself, maybe not as bad as it seems.

....but of course could also be bad depending on those things

[jbergey22]

The designers need to figure out a way to design a value for each player based on settings. Just a single number for each player such as in baseball Trout would be 100 based on age, contract, and production, where as certain people with bad contracts and past their prime(Joe Mauer) would have a negative value. Trade setting would determine to which level the computer will accept trades.

OOTP Baseball does a decent job of just not allowing you to throw a bunch of players in that have no value to get a player you want. Still could be improved a bit but its certainly improved over the years.

Throwing in a bunch of players that have no value should not make the computer accept a trade.

[Johnny93g]

Originally Posted by jbergey22

The designers need to figure out a way to design a value for each player based on settings. Just a single number for each player such as in baseball Trout would be 100 based on age, contract, and production, where as certain people with bad contracts and past their prime(Joe Mauer) would have a negative value. Trade setting would determine to which level the computer will accept trades.

OOTP Baseball does a decent job of just not allowing you to throw a bunch of players in that have no value to get a player you want. Still could be improved a bit but its certainly improved over the years.

Throwing in a bunch of players that have no value should not make the computer accept a trade.

Totally agree.

I don't care if the players have names we know or not. I don't want to play a game where Johan Franzen and a bunch of future draft picks get you Victor Hedman, or any equivalent of that. I acknowledge how difficult it is to get trade AI right, and how it will never be perfect, but an engine that SEEMS to allow this is not good enough in my books.

[Honolulu_Blue]

Originally Posted by Johnny93g

Totally agree.

I don't care if the players have names we know or not. I don't want to play a game where Johan Franzen and a bunch of future draft picks get you Victor Hedman, or any equivalent of that. I acknowledge how difficult it is to get trade AI right, and how it will never be perfect, but an engine that SEEMS to allow this is not good enough in my books.

I, for one, would not be opposed to such a game. Welcome to Detroit, Mr. Hedman!

[Sebastian Palkowski]

Thats exactly how the Trade AI in FHM works. Every player has a single number value, every draft pick gets a value based on the upcoming draft class and the expected pick number. Believe me, we try to tweak this as often as possible but getting the balancing right is really tough. Then you have scouting factored in as well. But the human just knows if e.g. a player is over/under-rated by the scouts whereas the AI does not know the name.

I know that the Trade AI is far from perfect and we use the feedback from users to further tweak it.

Originally Posted by jbergey22

The designers need to figure out a way to design a value for each player based on settings. Just a single number for each player such as in baseball Trout would be 100 based on age, contract, and production, where as certain people with bad contracts and past their prime(Joe Mauer) would have a negative value. Trade setting would determine to which level the computer will accept trades.

OOTP Baseball does a decent job of just not allowing you to throw a bunch of players in that have no value to get a player you want. Still could be improved a bit but its certainly improved over the years.

Throwing in a bunch of players that have no value should not make the computer accept a trade.

[JonInMiddleGA]

Originally Posted by Sebastian Palkowski

Thats exactly how the Trade AI in FHM works. Every player has a single number value, every draft pick gets a value based on the upcoming draft class and the expected pick number. Believe me, we try to tweak this as often as possible but getting the balancing right is really tough. Then you have scouting factored in as well. But the human just knows if e.g. a player is over/under-rated by the scouts whereas the AI does not know the name.

I know that the Trade AI is far from perfect and we use the feedback from users to further tweak it.

Seems as though the example that was given could be one worth looking into more specifically.

Trade with Tampa
- To Detroit: Victor Hedman
- To Tampa: Johan Franzen, 1st and 2nd (2018), 2nd and 3rd (2017)

I will admit that there could be too many save-specific variables -- mods? scouts quality? draft class? -- involved for an exhaustive examination of the formula but if the same deal is readily duplicated by numerous users at/near startup conditions (I have no idea whether that's the case) then it would
seems like a good opportunity to look at the formula, determine how & why something that smells rather off is happening & what tweaks could prevent it in a future build.

[djsatu]

I caved and bought the game, but the trade engine is beyond broken. Can't enjoy the game when you see stuff like this (All these were done at the 10 game mark, with trade frequency at 5)

Trade # 1

Florida sends Dave Bolland to Anaheim for Despres.

Again people say the game doesn't take names into account, but no way this trade happens in real life. No way.

Trade # 2 Minny sends Darcy Kuemper to Washington for Sean Collins.

Meh, nothing major here.

Trade # 3

Florida sends Willie Mitchell to Calgary for Derek Engellend.

2 aging players, but makes no sense for Calgary to make this trade as Engellend is 5 years younger.

Trade # 4

NJ sends Mackenzie Blackwood to Carolina for Jeff Skinner.

Unless Blackwood is an next up and coming top tier goalie prospect, there is no way Carolina trades one of there best young players for this prospect straight up.

Trade # 5

Carolina trades Eddie Lack and Brendan Collier to Ottawa for Wiercioch.

This trade doesn't make any sense on any level.

Trade # 6

Buffalo trades Foligno to Calgary for Smid.

Why would Buffalo do this?

Trade # 7

Nashville trades Jackman and Nystrom to Toronto for Reilly.

This is just as bad as the Couture - Engelland trade that was posted on here. Stud young d-man for 2 over the hill players that add nothing to Toronto.

Also Montreal put Tinordi on waivers, whom I happily picked up for free.
Picked up JT Brown from TB on waivers after he was passed on, then was offered a deal for him a week later from Nashville that I refused.

UPDATE: Ottawa waives Eddie Lack exactly 30 days after he was dealt

[JeffR]

Originally Posted by JonInMiddleGA

Seems as though the example that was given could be one worth looking into more specifically.

Trade with Tampa
- To Detroit: Victor Hedman
- To Tampa: Johan Franzen, 1st and 2nd (2018), 2nd and 3rd (2017)

I will admit that there could be too many save-specific variables -- mods? scouts quality? draft class? -- involved for an exhaustive examination of the formula but if the same deal is readily duplicated by numerous users at/near startup conditions (I have no idea whether that's the case) then it would
seems like a good opportunity to look at the formula, determine how & why something that smells rather off is happening & what tweaks could prevent it in a future build.

Well, let's break down what could be going on here:

Hedman may be underrated. Checking the database, doesn't look too bad. Top 10 in the league with potential to be Top 5.

Hedman's contract may be too expensive for his talent level. (Something that people really seem to miss when talking about the trade AI is that freeing cap space has tangible value in the NHL, and a deal that may not look like it makes sense on a 1-for-1 basis may be perfectly reasonable when you consider the cap implications.) But no, a $4m cap hit is actually cheap for what Hedman's worth.

Franzen may be overrated. Looks like a bit of a problem here, he's a little too good given his last two seasons. I'll tone him down some.

Franzen's contract may be particularly cheap for his talent level. Also around a $4m hit, which is in line with his talent level as currently stated in the game, but once I fix that, he'll be a little more expensive and less valuable.

Pick valuation could be off. The AI, when valuing picks, looks forward at the draft pool for that year and gets an approximate value for what a player picked in that particular spot will be worth. We're looking at the 2017 and 2018 drafts here, 2-3 years in advance, so, when researching, it's a little difficult for us to rate those kids at this point and they depend partially on randomization and some generated fictional players. If the systems that handle those things are a little too generous right now, that could be inflating the talent level in those rounds, and with it the value of those early-round picks that make up most of the deal. I'll have to dig through some test saves to evaluate that a little more accurately. Those first few drafts, between the first one where there's a ton of information available on the prospects but before the player generation system kicks in 100%, are always a pain to get balanced right.

There are the team motivations to look at, too. This is probably happening in the first place because Detroit's looking for a #1 defenceman. But normally Tampa wouldn't qualify as a target for them because they don't have a surplus of top defencemen - it's a long step from Hedman to Garrison and Coburn. We'll have to check the code there and make sure it's looking at the prerequisites properly. And Detroit arguably doesn't need another top defenceman in the first place with Kronwall and Green around. But it looks like Kronwall's ratings are a little too harsh on him, leaving them just below the threshold that makes the Wings think they need someone better. I'm going to bump him up a little.

Aside from the ratings fixes, the first attempt at improving the trade logic will probably tone down the salary/quality comparison a bit, so that doesn't have such a large effect on trade valuation. I'm also going to have the guys check the way the value reduction for age is working right now - that's a big problem with this deal, Franzen's age should hurt his value badly. I think him having 5 years remaining on his contract may be short-circuiting the logic normally used there, since the AI tends to assume (somewhat accurately) that a player won't retire before his contract is up. But I suspect it's not doing a good job discounting him based on his likely decline before then.

Unrelated to this deal, we also need to figure out how to put a roster hold on higher-quality recently-acquired players, so the AI isn't trading them again or dumping them so quickly. That'll be a little more involved, so it won't be in the first quick fixes.

Longer-term, the roster AI, including trading, is probably going to be looking at a fundamental overhaul. It's built on some of the oldest code in the game, and that's really starting to show. It needs to become a lot more efficient and flexible, both to speed up the game and produce better results, particularly with regard the different management styles necessary in different types of leagues. I definitely don't want to try doing college hockey before that happens, and the historical, junior and European team AI's would be considerably improved if they were designed from the ground up to operate in their unique environments.

On a more philosophical note: good trading AI is difficult to do even in very straightforward leagues. Markus has been working on it for a decade and a half in OOTP and still gets complaints. When you add the complications of drafts, free agency possibilities, cap implications, and everything else that affects the way NHL GM's value players, it gets a lot tougher. We'll always be doing what we can to improve it, but it's going to be a gradual process; making radical AI changes rapidly never turns out well. But it's never going to produce perfectly balanced trades 100% of the time, and it shouldn't - bad trades are part of sports, and not having any is as bad as having too many. We'll keep pushing towards the right balance between the two.

[JonInMiddleGA]

Originally Posted by JeffR

Well, let's break down what could be going on here:

Seems like a reasonable approach

Pick valuation could be off.

Your description of how the AI handles the value of picks seems like a solid enough one, with the caveat of assuming it gets the values right. Better than the more traditional (I think) gaming approach that might just say average-value-of-pick-X-is-Y.

But it looks like Kronwall's ratings are a little too harsh on him, leaving them just below the threshold that makes the Wings think they need someone better.

This is actually something that I see with a number of hockey games (I think). Chasing the magical one point improvement to get a "true #1" or making it a priority over other improvements that could be more meaningful overall. Not sure how far a game can go to make the logic "think" more broadly, it's not something I feel like I see done all that often tbh.

[Johnny93g]

Originally Posted by JeffR

Jeff talking about the Hedman Franzen trade

That's some really good stuff there, and i appreciate it. Everything you said makes sense, and it's the kind of thing that gives me some hope that this game will be good in the end. I really want it to be great. I'm a hockey guy who likes text sims.

I don't think ill be picking it up in the near future, but if this is where the development is going, then i do believe it's only a matter of time.

I would say that devaluing future picks significantly is probably the best way to handle that aspect. Because there's so much uncertainty surrounding them, picks 2-4 years in the future don't carry nearly as much value as current picks, and one of my pet peeves the complete sporting game genre is how the use of future picks can be somewhat easily exploited against CPU opponents.

[JeffR]

Originally Posted by JonInMiddleGA

Your description of how the AI handles the value of picks seems like a solid enough one, with the caveat of assuming it gets the values right. Better than the more traditional (I think) gaming approach that might just say average-value-of-pick-X-is-Y.

It's actually a little more complex than I described. There are certain ranges it'll be limited to, but the focus there is primarily on the down side - I wanted to preserve the chance of teams anticipating an occasional very deep draft class like the 2016 one is getting hyped to be.

[JeffR]

Originally Posted by Johnny93g

I would say that devaluing future picks significantly is probably the best way to handle that aspect. Because there's so much uncertainty surrounding them, picks 2-4 years in the future don't carry nearly as much value as current picks, and one of my pet peeves the complete sporting game genre is how the use of future picks can be somewhat easily exploited against CPU opponents.

Yup, the value is lowered a bit, it's not the literal potential of the prospect likely to go on that slot. The exact amount depends on the preferences of the team evaluating it - if they're in rebuild mode, it's less of a reduction, in win-now mode it's more.

[RomaGoth]

Originally Posted by Honolulu_Blue

I, for one, would not be opposed to such a game. Welcome to Detroit, Mr. Hedman!

I echo this sentiment. Now, if you will please sign here, initial here and here, we can have that 10 year contract done straight away.

[RomaGoth]

Originally Posted by JeffR

Well, let's break down what could be going on here:

Hedman may be underrated. Checking the database, doesn't look too bad. Top 10 in the league with potential to be Top 5.

Hedman's contract may be too expensive for his talent level. (Something that people really seem to miss when talking about the trade AI is that freeing cap space has tangible value in the NHL, and a deal that may not look like it makes sense on a 1-for-1 basis may be perfectly reasonable when you consider the cap implications.) But no, a $4m cap hit is actually cheap for what Hedman's worth.

Franzen may be overrated. Looks like a bit of a problem here, he's a little too good given his last two seasons. I'll tone him down some.

Franzen's contract may be particularly cheap for his talent level. Also around a $4m hit, which is in line with his talent level as currently stated in the game, but once I fix that, he'll be a little more expensive and less valuable.

Pick valuation could be off. The AI, when valuing picks, looks forward at the draft pool for that year and gets an approximate value for what a player picked in that particular spot will be worth. We're looking at the 2017 and 2018 drafts here, 2-3 years in advance, so, when researching, it's a little difficult for us to rate those kids at this point and they depend partially on randomization and some generated fictional players. If the systems that handle those things are a little too generous right now, that could be inflating the talent level in those rounds, and with it the value of those early-round picks that make up most of the deal. I'll have to dig through some test saves to evaluate that a little more accurately. Those first few drafts, between the first one where there's a ton of information available on the prospects but before the player generation system kicks in 100%, are always a pain to get balanced right.

There are the team motivations to look at, too. This is probably happening in the first place because Detroit's looking for a #1 defenceman. But normally Tampa wouldn't qualify as a target for them because they don't have a surplus of top defencemen - it's a long step from Hedman to Garrison and Coburn. We'll have to check the code there and make sure it's looking at the prerequisites properly. And Detroit arguably doesn't need another top defenceman in the first place with Kronwall and Green around. But it looks like Kronwall's ratings are a little too harsh on him, leaving them just below the threshold that makes the Wings think they need someone better. I'm going to bump him up a little.

Aside from the ratings fixes, the first attempt at improving the trade logic will probably tone down the salary/quality comparison a bit, so that doesn't have such a large effect on trade valuation. I'm also going to have the guys check the way the value reduction for age is working right now - that's a big problem with this deal, Franzen's age should hurt his value badly. I think him having 5 years remaining on his contract may be short-circuiting the logic normally used there, since the AI tends to assume (somewhat accurately) that a player won't retire before his contract is up. But I suspect it's not doing a good job discounting him based on his likely decline before then.

Unrelated to this deal, we also need to figure out how to put a roster hold on higher-quality recently-acquired players, so the AI isn't trading them again or dumping them so quickly. That'll be a little more involved, so it won't be in the first quick fixes.

Longer-term, the roster AI, including trading, is probably going to be looking at a fundamental overhaul. It's built on some of the oldest code in the game, and that's really starting to show. It needs to become a lot more efficient and flexible, both to speed up the game and produce better results, particularly with regard the different management styles necessary in different types of leagues. I definitely don't want to try doing college hockey before that happens, and the historical, junior and European team AI's would be considerably improved if they were designed from the ground up to operate in their unique environments.

On a more philosophical note: good trading AI is difficult to do even in very straightforward leagues. Markus has been working on it for a decade and a half in OOTP and still gets complaints. When you add the complications of drafts, free agency possibilities, cap implications, and everything else that affects the way NHL GM's value players, it gets a lot tougher. We'll always be doing what we can to improve it, but it's going to be a gradual process; making radical AI changes rapidly never turns out well. But it's never going to produce perfectly balanced trades 100% of the time, and it shouldn't - bad trades are part of sports, and not having any is as bad as having too many. We'll keep pushing towards the right balance between the two.

Originally Posted by JeffR

It's actually a little more complex than I described. There are certain ranges it'll be limited to, but the focus there is primarily on the down side - I wanted to preserve the chance of teams anticipating an occasional very deep draft class like the 2016 one is getting hyped to be.

Originally Posted by JeffR

Yup, the value is lowered a bit, it's not the literal potential of the prospect likely to go on that slot. The exact amount depends on the preferences of the team evaluating it - if they're in rebuild mode, it's less of a reduction, in win-now mode it's more.

Thanks for the all hard work and explaining here as to what your thoughts are. The first iteration of FHM was a dumpster fire, but you guys have really taken a lot of feedback and improved this version immensely. I bought it yesterday and will continue to support sports text sims as long as the developer shows they are listening and want to continuously improve their product.

[BishopMVP]

Originally Posted by JonInMiddleGA

Seems as though the example that was given could be one worth looking into more specifically.

Trade with Tampa
- To Detroit: Victor Hedman
- To Tampa: Johan Franzen, 1st and 2nd (2018), 2nd and 3rd (2017)

I will admit that there could be too many save-specific variables -- mods? scouts quality? draft class? -- involved for an exhaustive examination of the formula but if the same deal is readily duplicated by numerous users at/near startup conditions (I have no idea whether that's the case) then it would
seems like a good opportunity to look at the formula, determine how & why something that smells rather off is happening & what tweaks could prevent it in a future build.

I found some more trades along those lines... Dougie Hamilton (6 year contract, $5.75m/y deal) traded for a 1st & 2 2nd round picks. Then Tyler Seguin (again, 6 year contract at $5.75m/y) traded for Loui Erikson, Reilly Smith and Joe Morrow. No way a team would be dumb enough to trade one 22 y/o budding superstar for that low a return, let alone double down on the same choice a second time.

Snark aside (and assuming you don't have dynamic GM's where some are just bad at their job), good to see the developer looking to fix things. I guess my one contribution would be to ask if teams also project how good the other team is going to be - that's always been one area ripe for players ripping off the AI in most games. Even in FOF, if I want I can trade my future 1st and a little extra for a terrible team's 1st if I do it a year in advance.

[Ramzavail]

I was able to start a fictional league, rename all the team names, create minors and feeder leagues and sim 10 seasons within an hour. Sim speed is super impressive. Stats so far look pretty realistic. Save PCT looks a touch too low and GAA looks a touch too high, but I haven't fiddled with any sliders.

I am impressed with how many things they took from OOTP (although I wish they have implemented the ability to go back by hitting the right mouse button). Facegen, popup windows with stats when hovering over the player, icons for awards. Really impressive. I can see myself getting immersed in this a lot more than I ever did with (new) EHM.

[NoSkillz]

Agreed on the sim speed. It's VERY impressive indeed. It was taking me approximately one to two minutes to sim through an entire season in historical mode (during original six era - no farm teams, no other leagues running).

I have fiddled a bit with the historical mode and it's a lot of fun seeing those old names.

There was an "interesting" thing I saw regarding Bobby Orr as I was quick simming. I stopped simming in 1970 and noticed he was on Minnesota so I took a look at his history. He was drafted by Boston in 1966 in the game but one year later, he had signed with Minnesota as a free agent.

"Well, that's kind of strange", I thought.

He'd go on to win rookie of the year with the North Stars the next year and win the Conn Smythe for playoff MVP as well despite his North Stars losing the Cup final. He'd end up soon signing the highest contract in the NHL within two years.

Now, that *could* be a bug that Boston didn't get a 5 star prospect signed OR it could be a super cool feature if Orr (kind of like real life with Alan Eagleson) and his agent were greedy enough to hold out and not sign with Boston, becoming a free agent at the end of the year and eventually just signing a huge contract with Minnesota.

While the developer fixes the bugs with the new release, I think I'm going to write up a little dynasty about this historical mode here - it's lots of fun and the sim engine seems pretty spot on with stats thus far so it could be an enjoyable exercise. Would anyone be interested in reading that?

[RomaGoth]

Originally Posted by NoSkillz

...I think I'm going to write up a little dynasty about this historical mode here - it's lots of fun and the sim engine seems pretty spot on with stats thus far so it could be an enjoyable exercise. Would anyone be interested in reading that?

Sure would!

[JeffR]

Originally Posted by NoSkillz

Now, that *could* be a bug that Boston didn't get a 5 star prospect signed OR it could be a super cool feature if Orr (kind of like real life with Alan Eagleson) and his agent were greedy enough to hold out and not sign with Boston, becoming a free agent at the end of the year and eventually just signing a huge contract with Minnesota.

It's most likely a weird thing going on with pre-season roster handling, probably related to some combination of league rules settings common to the 60's-70's-80's, since that's when it's most apparent. We haven't been able to pin down exactly what's doing it yet.